Preliminary Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 1 Features ■ System-on-a-chip integrated circuit supports lowspeed ATM access for next-generation wireless base transmission station (BTS), base station controller (BSC), node-B, radio network controller (RNC), and remote access concentrator (RAC) applications. ■ IC provides an integrated octal framer that supports T1/E1/J1 formats. ■ Supports inverse multiplexing for ATM (IMA) over selected group and link mappings ranging from four two-link groups up to one eight-link group per ATM Forum AF-PHY-0086.001. ■ Integrates an ATM adaptation layer 2 (AAL2) segmentation and reassembly (SAR) function for support of low-speed data or voice traffic per ITU I.363.2. ■ Provides AAL5 SAR functionality per ITU I.363.5. ■ Provides quality of service (QoS) connection identifier (CID) multiplexing per ITU I.366.1. ■ Enables ATM layer user network interface (UNI) or IMA mode, selectable on a per-link basis for flexible transport of delay critical voice and data traffic. ■ Guarantees QoS for a variety of traffic types (including delay-sensitive voice, real-time data, non-real-time data, and signaling information) through an advanced hierarchical three-level priority scheduler and per-VC queueing. ■ Supports 2032 bidirectional AAL2 CIDs. ■ Supports 2032 bidirectional high-speed data connections or virtual circuits (VCs) via embedded context memory; filters control cells and accepts control cells via a host microprocessor interface. ■ On-board memory is used for connection management and queue data storage. No external memory is needed. ■ Software package includes the following: — Device manager source code (C-based device manager ready-to-use with host RTOS). — Setup file utility to provision TAAD08JU2. — Firmware for embedded controller (executable binary). — API reference manual available for device manager software. ■ Designed in 0.16 µm, low-power CMOS technology. 2 Physical ■ 3.3 V digital I/O compatibility; 1.5 V core power ■ 520 enhanced ball-grid array (EBGA) package ■ –40 oC to +85 oC temperature range 3 Standards ITU I.363.2, ITU I.363.5, ITU I.366.1, ITU I.366.2, ITU I.432, ITU I.361, ITU I.371, ITU G.703, ITU G.704, ITU G.804, ITU G.732, ITU G.706, ITU I.610, ITU G.775, ITU G.733, ITU G.735, ITU G.965, ITU O.162, ANSI® T1.403, ANSI T1.231, ATM Forum AF-PHY-0086.001 ATM Forum AF-PHY-0039.000 ATM Forum AF-TM-0121.000 ETS 300.417-1-1 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Preliminary Data Sheet August 18, 2003 Table of Contents Contents 1 2 3 4 5 6 7 8 9 10 11 12 13 2 Page Features ............................................................................................................................................................. 1 Physical .............................................................................................................................................................. 1 Standards ........................................................................................................................................................... 1 Description ....................................................................................................................................................... 11 Pin Definitions ..................................................................................................................................................12 Pin Description .................................................................................................................................................12 Package Pin Layout .........................................................................................................................................21 Block Diagram ..................................................................................................................................................27 Software Components ......................................................................................................................................28 9.1 Firmware.................................................................................................................................................29 9.2 Device Manager......................................................................................................................................29 9.3 Setup File Utility (SFU) ...........................................................................................................................30 9.4 TAAD08JU2 Application Code................................................................................................................31 9.5 System Software.....................................................................................................................................32 9.6 Software Development Environment ......................................................................................................32 9.7 Notes ......................................................................................................................................................33 Functional Overview.........................................................................................................................................34 10.1 Receive Direction Data Flow ..................................................................................................................34 10.1.1 PHY Layer ................................................................................................................................34 10.1.2 Low-Speed PHY Links ..............................................................................................................34 10.1.3 High-Speed PHY Links .............................................................................................................35 10.1.4 TC and IMA Layers...................................................................................................................35 10.1.5 ATM Layer ................................................................................................................................36 10.1.6 AAL Engine...............................................................................................................................36 10.1.7 Embedded Device Controller....................................................................................................37 10.2 Transmit Direction Data Flow..................................................................................................................37 10.2.1 SSCS/AAL Layer Interaction ....................................................................................................37 10.2.2 ATM Layer ................................................................................................................................37 10.2.3 IMA/TC Layer............................................................................................................................38 10.2.4 PHY Layer ................................................................................................................................38 Modes of Operation..........................................................................................................................................39 11.1 Interface Modes ......................................................................................................................................39 11.1.1 UTOPIA-2 Expansion Port Multiplexing Modes ........................................................................39 11.1.2 System Interface Port Multiplexing Modes ...............................................................................39 11.1.3 Line-Interface Modes ................................................................................................................40 11.2 Device Operating Modes ........................................................................................................................40 11.2.1 Operating Mode 1: Internal PHY Mode.....................................................................................40 11.2.2 Operating Mode 2: External PHY Mode ...................................................................................42 11.2.3 Operating Mode 3: SAR-Only Mode .........................................................................................43 11.2.4 Operating Mode Summary........................................................................................................43 Applications ......................................................................................................................................................44 12.1 BTS Network Interface Termination ........................................................................................................44 12.2 VToA Trunking Application......................................................................................................................46 12.3 Low-Speed ATM Access.........................................................................................................................47 12.4 AAL2 Cross Connect ..............................................................................................................................47 Embedded Device Controller (EDC) ................................................................................................................48 13.1 Introduction .............................................................................................................................................48 13.2 Features..................................................................................................................................................48 13.3 EDC Functional Description....................................................................................................................48 13.4 Host Interface .........................................................................................................................................48 13.5 Host Interface Signals and Timing ..........................................................................................................49 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Table of Contents (continued) Contents Page 13.6 Host Interactions.....................................................................................................................................49 14 Framer Block ....................................................................................................................................................53 14.1 Introduction .............................................................................................................................................53 14.2 Features..................................................................................................................................................53 14.3 Framer-to-Line Interface Unit Physical Interface ....................................................................................54 14.3.1 Line Interface References/Standards .......................................................................................54 14.3.2 Clocking Modes ........................................................................................................................54 14.3.3 Frame Formats .........................................................................................................................55 14.3.4 Transmit Framer Functions.......................................................................................................55 14.4 DS1 Transparent Framing Format ..........................................................................................................55 14.5 CEPT 2.048 Basic Frame Structure Transparent Framing Format.........................................................56 14.6 Receive Framer Nonalignment Mode (DS1/E1) .....................................................................................57 14.6.1 Loss of Frame Alignment Criteria .............................................................................................57 14.6.1.1 Frame Bit Errors .........................................................................................................57 14.6.1.2 CRC Errors.................................................................................................................57 14.7 Frame Alignment Criteria........................................................................................................................57 14.8 Performance-Monitoring Functional Integration Into Framer ..................................................................58 14.9 Performance Report Message................................................................................................................61 14.10 ESF Data Link.........................................................................................................................................62 14.11 Facility Data Link ....................................................................................................................................62 14.11.1 Facility Data Link References/Standards ..................................................................................62 14.11.2 Receive Data Link Functional Description ................................................................................63 14.11.3 SLC-96 Superframe Receive Data Link....................................................................................63 14.11.4 DDS Receive Data Link Stack ..................................................................................................63 14.11.5 CEPT; CEPT CRC-4 (100 ms); CEPT CRC-4 (400 ms) Multiframe Sa Bits Receive Stack .....63 14.11.6 Receive Data Link Stack Idle Modes ........................................................................................64 14.11.7 Transmit Facility Data Link Functional Description ...................................................................64 14.11.8 SLC-96 Superframe Transmit Data Link...................................................................................64 14.11.9 DDS Transmit Data Link Stack .................................................................................................64 14.11.10 Transmit ESF Data Link Bit-Oriented Messages ......................................................................64 14.11.11 CEPT, CEPT Multiframe Transmit Data Link Sa Bits Stack ......................................................65 14.11.12 Transmit Data Link Stack Idle Modes .......................................................................................66 14.11.13 SLC-96, DDS, or CEPT ESF Frame Alignment ........................................................................66 14.12 Concentration Highway Interface (CHI) ..................................................................................................66 14.12.1 CHI References/Standards .......................................................................................................66 14.12.2 Transmit/Receive CHI Features................................................................................................66 14.12.3 Double NOTFAS System Time-Slot Mode................................................................................67 14.12.4 Transparent Mode.....................................................................................................................67 14.12.5 Loopbacks ................................................................................................................................67 14.12.6 Nominal CHI Timing..................................................................................................................68 14.12.7 CHI Timing with CHI Double Time-Slot Timing (CHIDTS) Mode Enabled ................................69 14.12.8 Clocking Scheme......................................................................................................................69 15 Transmission Convergence (TC) Block ............................................................................................................70 15.1 Introduction .............................................................................................................................................70 15.2 Features..................................................................................................................................................70 15.3 TC—Receive Direction ...........................................................................................................................71 15.4 TC—Transmit Direction ..........................................................................................................................71 15.4.1 HEC Generation/Checking .......................................................................................................72 15.5 Cell Delineation.......................................................................................................................................72 15.6 Cell Payload Scrambling/Descrambling..................................................................................................72 15.7 Cell Mapping...........................................................................................................................................72 15.8 Facility Maintenance ...............................................................................................................................72 Agere Systems Inc. Agere Systems - Proprietary Use pursuant to Company instructions 3 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 Table of Contents (continued) Contents Page 15.9 Cell Rate Decoupling ..............................................................................................................................73 15.10 Functionality............................................................................................................................................73 16 Inverse Multiplexing for ATM (IMA) Block.........................................................................................................74 16.1 Introduction .............................................................................................................................................74 16.2 Features..................................................................................................................................................75 16.3 Multi-PHY UTOPIA Slave Interface ........................................................................................................76 16.4 Link Processor ........................................................................................................................................76 16.5 Group Processor.....................................................................................................................................77 16.6 Delay Compensation Buffer (DCB).........................................................................................................78 16.7 Programming the DCB............................................................................................................................85 16.7.1 Link Start-Up Guardband Field .................................................................................................85 16.7.2 Link Maximum Operational Delay.............................................................................................85 16.8 Features Not Supported in IMA ..............................................................................................................85 17 ATM Port Controller (APC) Block .....................................................................................................................87 17.1 Introduction .............................................................................................................................................87 17.2 Architecture.............................................................................................................................................88 17.3 Features..................................................................................................................................................89 17.4 Summary of Commands .........................................................................................................................90 17.5 Buffer Management ................................................................................................................................90 17.6 Scheduling ..............................................................................................................................................92 17.6.1 Ingress Scheduling ...................................................................................................................92 17.6.2 Fabric Backpressure.................................................................................................................93 17.6.3 Egress Scheduling....................................................................................................................93 17.7 ABR Flow Control ...................................................................................................................................93 17.8 Control Plane Functions .........................................................................................................................94 17.8.1 APC Support for Control Plane Functions ................................................................................94 17.9 Management Plane Functions ................................................................................................................94 17.9.1 Operation Administration and Maintenance (OAM) ..................................................................94 17.10 Statistics Counters ..................................................................................................................................95 17.11 Ingress Enqueue Operations ..................................................................................................................95 17.11.1 Connection Look-Up .................................................................................................................96 17.11.2 OAM Processing.......................................................................................................................97 17.11.3 Policing .....................................................................................................................................98 17.11.4 Buffer Thresholding ..................................................................................................................98 17.11.5 Egress—APC VC Queueing Structure......................................................................................98 17.12 Connection Management........................................................................................................................99 17.12.1 Connection Admission Control .................................................................................................99 17.12.1.1 CBR............................................................................................................................99 17.12.1.2 rt-VBR ........................................................................................................................99 17.12.1.3 nrt-VBR ......................................................................................................................99 17.12.1.4 ABR............................................................................................................................99 17.12.1.5 UBR............................................................................................................................99 18 ATM Adaptation Layer (AAL) Block ................................................................................................................100 18.1 Introduction ...........................................................................................................................................100 18.2 Features................................................................................................................................................100 18.3 Definitions .............................................................................................................................................101 18.4 Architecture...........................................................................................................................................102 18.4.1 Datapath Flows .......................................................................................................................102 18.4.2 Subblock Architecture.............................................................................................................105 18.4.3 Subblock Definition .................................................................................................................106 18.4.4 Subblock Flows.......................................................................................................................107 18.4.5 Address Translation ................................................................................................................108 4 Agere Systems - Proprietary Use pursuant to Company instructions Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Table of Contents (continued) Contents Page 18.4.6 18.4.7 18.4.8 18.4.9 18.4.10 18.4.11 18.4.12 18.4.13 18.4.14 18.4.15 18.4.16 18.4.17 18.4.18 Queueing and Scheduling ......................................................................................................109 Modes .....................................................................................................................................109 User Data Types (UDT) and AAL Types ................................................................................. 110 UDT: ATM Cell ........................................................................................................................ 111 AAL Type: AAL0 ..................................................................................................................... 111 AAL Type: AAL2 ..................................................................................................................... 111 AAL2 Subtype: SPAAL2 (Single-Packet AAL2) ...................................................................... 112 CPS-AAL0 .............................................................................................................................. 113 AAL Type: AAL5 ..................................................................................................................... 113 UDT: Packet ATM (PATM)....................................................................................................... 114 UDT: HPF................................................................................................................................ 115 AAL Type: NPAAL (No Particular AAL)................................................................................... 116 Nonuser Data Types: ESI Messages ...................................................................................... 116 18.4.18.1 ESI Message Format ............................................................................................... 116 18.4.18.2 ESI Violation Code ................................................................................................... 117 18.4.18.3 ESI Packet Length ................................................................................................... 117 18.4.19 Service Types ......................................................................................................................... 117 18.4.20 CPS_SERVICE....................................................................................................................... 118 18.4.21 SEG_AAL2_SSSAR_SERVICE ............................................................................................. 119 18.4.22 SEG_AAL2_SSTED_SERVICE.............................................................................................. 119 18.4.23 SEG_AAL5_SERVICE............................................................................................................ 119 18.4.24 TRANSPARENT_SERVICE....................................................................................................120 18.4.25 REASS_AAL2_SSSAR_SERVICE .........................................................................................120 18.4.26 REASS_AAL2_SSTED_SERVICE .........................................................................................120 18.4.27 REASS_AAL5_SERVICE .......................................................................................................120 18.5 Provisioning ..........................................................................................................................................122 18.5.1 Some Notes on Terminology and Command Referencing......................................................122 18.5.2 System Interface.....................................................................................................................122 18.5.3 Port Table................................................................................................................................123 18.5.4 MEMI Shared Memory............................................................................................................124 18.5.4.1 MEMI-SM Provisioning Constraints..........................................................................125 18.5.4.2 VC Table...................................................................................................................125 18.5.4.3 AAL2 VC Table .........................................................................................................126 18.5.4.4 Connection Table .....................................................................................................127 18.5.4.5 Level 0 Queue Descriptor ........................................................................................129 18.5.4.6 ICID Table ................................................................................................................129 18.5.5 SQASE Shared Memory .........................................................................................................129 18.6 Configuration ........................................................................................................................................130 18.6.1 Connection and Channel Setup..............................................................................................130 18.6.1.1 AAL2 Data Flow (CPS/SSSAR/SSTED) ..................................................................133 18.6.1.2 CPS-AAL0 Data Flow...............................................................................................133 18.6.1.3 AAL0/AAL5 Data Flow..............................................................................................133 18.6.1.4 HPF Data Flow .........................................................................................................133 18.6.2 Configuration for QoS .............................................................................................................134 18.6.2.1 Packet Scheduling ...................................................................................................134 18.6.2.2 IL1Q Scheduler Algorithm ........................................................................................134 18.6.2.3 IL2Q Scheduler Algorithm ........................................................................................134 18.6.2.4 Latency Policing .......................................................................................................136 18.6.2.5 Latency-Sensitive Data Discard ...............................................................................136 18.6.2.6 Internal Queue Housekeeping .................................................................................136 18.6.2.7 Reference Clock Generation ....................................................................................136 18.6.2.8 Latency Timer Enable/Disable Functions.................................................................137 Agere Systems Inc. Agere Systems - Proprietary Use pursuant to Company instructions 5 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 Table of Contents (continued) Contents Page 18.6.2.9 Queue Length Policing.............................................................................................137 18.6.2.10 Connection Queue Length Policing..........................................................................138 18.6.2.11 IL1Q and L1Q Length Policing .................................................................................139 18.6.2.12 IL2Q Length Policing................................................................................................140 18.6.3 Configuration for Exceptions...................................................................................................140 18.7 Interface Timing Diagrams....................................................................................................................143 18.7.1 SIF UT2/UT2+ Interface .........................................................................................................143 18.7.2 Polling Algorithms for UTOPIA 2 and UT2+ Modes................................................................152 18.7.2.1 Receive Interface Polling .........................................................................................152 18.7.2.2 Transmit Interface Polling.........................................................................................153 18.7.3 NIF ..........................................................................................................................................153 18.7.4 ESI ..........................................................................................................................................153 19 Absolute Maximum Ratings............................................................................................................................154 20 Power Requirements ......................................................................................................................................154 21 Handling Precautions .....................................................................................................................................155 22 Electrical Characteristics ................................................................................................................................155 22.1 Logical Interface Electrical Characteristics, Version 2.1 .......................................................................155 22.2 Logical Interface Electrical Characteristics, Version 3.1 .......................................................................156 23 Timing Characteristics ....................................................................................................................................157 23.1 Input Clocks, Versions 2.1 and 3.1 .......................................................................................................157 23.2 Host Interface Timing............................................................................................................................158 23.3 Reset Timing.........................................................................................................................................160 23.4 Concentration Highway (CHI) Timing, Versions 2.1 and 3.1 ................................................................161 23.5 Fabric Interface—Ports A and B, Versions 2.1 and 3.1 ........................................................................162 23.6 Expansion UTOPIA2 Interface..............................................................................................................164 23.6.1 Receive Interface Timing ........................................................................................................164 23.6.2 Transmit Interface Timing .......................................................................................................164 23.7 Enhanced Services Interface (ESI), Versions 2.1 and 3.1....................................................................164 23.8 JTAG.....................................................................................................................................................165 23.9 System Interface, Version 2.1...............................................................................................................165 23.9.1 Receive Interface Timing, Version 2.1 ....................................................................................165 23.9.2 Transmit Interface Timing, Version 2.1 ...................................................................................165 23.10 System Interface, Version 3.1...............................................................................................................166 23.10.1 Receive Interface Timing, Version 3.1 (SUCLK Is an Input.) ..................................................166 23.10.2 Transmit Interface Timing, Version 3.1 (SUCLK Is an Input.) .................................................166 24 Referenced Documents..................................................................................................................................167 25 Glossary .........................................................................................................................................................169 Appendix A. Revision History ................................................................................................................................172 6 Agere Systems - Proprietary Use pursuant to Company instructions Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 List of Figures Figure Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure 11. Figure 12. Figure 13. Figure 14. Figure 15. Figure 16. Figure 17. Figure 18. Figure 19. Figure 20. Figure 21. Figure 22. Figure 23. Figure 24. Figure 25. Figure 26. Figure 27. Figure 28. Figure 29. Figure 30. Figure 31. Figure 32. Figure 33. Figure 34. Figure 35. Figure 36. Figure 37. Figure 38. Figure 39. Figure 40. Figure 41. Figure 42. Figure 43. Figure 44. Figure 45. Figure 46. Figure 47. Figure 48. Figure 49. Figure 50. Figure 51. Figure 52. Page Pin Configuration Diagram ................................................................................................................21 Architecture of the TAAD08JU2 Device ............................................................................................27 Software Components .......................................................................................................................28 Device Manager APIs and TAAD08JU2 Communications ................................................................30 TAAD08JU2 Software Development Environment ............................................................................32 TAAD08JU2 Interfaces......................................................................................................................39 Mode 1: Internal PHY Mode Operation .............................................................................................40 Example of Sharing Span Line with TDM and ATM Data..................................................................41 Mode 2: External PHY Mode .............................................................................................................42 Mode 3: SAR-Only Mode...................................................................................................................43 BTS Application .................................................................................................................................44 BTSs Require ADM Functions...........................................................................................................45 Gateway Controller............................................................................................................................45 Remote Access Concentrator Application .........................................................................................46 Edge/Access Switch Application .......................................................................................................46 AAL2 Cross Connect .........................................................................................................................47 Stand-Alone AAL2 Cross Connect ....................................................................................................47 Standard Host Interface Timing.........................................................................................................49 DS1 Transparent Frame Structure ....................................................................................................56 CEPT Transparent Frame Structure..................................................................................................56 System Loopbacks ............................................................................................................................67 Nominal Concentration Highway Interface Timing ............................................................................68 CHIDTS Mode Concentration Highway Interface Timing ..................................................................69 IMA Application..................................................................................................................................74 IMA High-Level Interconnect Block Diagram.....................................................................................75 Logical View of Three-Link Group’s DCB Shortly After Starting to Receive Data from the Line .......78 Logical View of Three-Link Group’s DCB When It Starts Reading DCB ...........................................79 Logical View of Three-Link Group’s DCB After It Starts Reading DCB .............................................80 DCB During Normal Operation ..........................................................................................................81 Starting to Add a Link to a Group ......................................................................................................82 Link Now Being Read ........................................................................................................................83 Effects of Link #3 and Link #4 Faults.................................................................................................84 APC Block Integrated Memory Configuration....................................................................................87 Switch Fabric Connections for Dual TAAD08JU2 Switch Mode........................................................88 AAL Engine Block Diagram .............................................................................................................100 SIF-to-NIF, NIF-to-SIF .....................................................................................................................102 SIF Loopback, NIF Loopback ..........................................................................................................103 NIF Adaptation Loopback ................................................................................................................103 Host-to-SIF, SIF-to-Host..................................................................................................................104 Host-to-NIF, NIF-to-Host .................................................................................................................104 SAR Subblock Diagram...................................................................................................................105 Logical View of the Enqueue (Left) and Dequeue (Right) Address Translation Procedure.............108 Simplified Diagram of SQASE Queueing Structure.........................................................................109 User Data Types and AAL Types at the Interfaces .........................................................................110 User Data Type (UDT) vs. AAL Type Mapping ...............................................................................111 SPAAL2 Data Format ......................................................................................................................112 CPS-AAL0 Data Format ..................................................................................................................113 PATM Format ..................................................................................................................................114 HPF Format .....................................................................................................................................115 Transferring an HPF Packet over the Host Interface Example........................................................118 Port Table ........................................................................................................................................123 VC Table..........................................................................................................................................125 Agere Systems Inc. Agere Systems - Proprietary Use pursuant to Company instructions 7 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 List of Figures (continued) Figures Figure 53. Figure 54. Figure 55. Figure 56. Figure 57. Figure 58. Figure 59. Figure 60. Figure 61. Figure 62. Figure 63. Figure 64. Figure 65. Figure 66. Figure 67. Figure 68. Figure 69. Figure 70. Figure 71. Figure 72. Figure 73. Figure 74. 8 Page AAL2 VC Table................................................................................................................................126 Connection Table ............................................................................................................................127 SQASE Queueing Structure ............................................................................................................131 Connection Queue Length Policing .................................................................................................138 IL1Q/L1Q Length Policing ...............................................................................................................139 IL2Q Length Policing .......................................................................................................................140 UT2/UT2+ Header at the SIF Interface ...........................................................................................143 Cell Transmission on the SIF Interface ...........................................................................................144 Packet Transmission on the SIF Interface with No Stalls................................................................145 Packet Transmission on the SIF Interface with the PHY Stalling ....................................................146 Packet Transmission on the SIF Interface with the Master Stalling ................................................147 Reception of a Cell on the SIF Interface..........................................................................................148 Reception of a Packet on the SIF Interface with No Stalls ..............................................................149 Reception of a Packet on the SIF Interface with the PHY Stalling ..................................................150 Reception of a Packet on the SIF Interface with the Master Stalling...............................................151 ESI Functional Timing Diagram.......................................................................................................153 Data Read from TAAD08JU21 ........................................................................................................158 Data Written to TAAD08JU2............................................................................................................159 Power-On Reset ..............................................................................................................................160 Stable Reset ....................................................................................................................................160 CHI Transmit I/O Timing..................................................................................................................161 CHI Receive I/O Timing...................................................................................................................162 Agere Systems - Proprietary Use pursuant to Company instructions Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 List of Tables Table Table 1. Table 2. Table 3. Table 4. Table 5. Table 6. Table 7. Table 8. Table 9. Table 10. Table 11. Table 12. Table 13. Table 14. Table 15. Table 16. Table 17. Table 18. Table 19. Table 20. Table 21. Table 22. Table 23. Table 24. Table 25. Table 26. Table 27. Table 28. Table 29. Table 30. Table 31. Table 32. Table 33. Table 34. Table 35. Table 36. Table 37. Table 38. Table 39. Table 40. Table 41. Table 42. Table 43. Table 44. Table 45. Table 46. Table 47. Table 48. Table 49. Table 50. Table 51. Table 52. Page Pin Definitions................................................................................................................................... 12 Transmission Line Interface Signals (48 Signals) ............................................................................ 13 CHI Interface Signals (20 Signals) ................................................................................................... 14 UTOPIA 2 Expansion Interface Signals (52 Signals) ....................................................................... 14 System Interface Signals (62 Signals).............................................................................................. 15 Switch Fabric Interface Signals (50 Pins)......................................................................................... 17 APC External Statistics Interface Signals (18 Signals) .................................................................... 18 SAR External Statistics Interface Signals (18 Signals)..................................................................... 18 Host Interface Signals (49 Signals) .................................................................................................. 19 JTAG Interface Pins (6 Signals) ...................................................................................................... 19 Global/Miscellaneous Signal Pins (10 Signals) ................................................................................ 19 Power Supply Pins (4 Analog Power Pins, 120 Digital Power Pins) ............................................... 20 Signal-to-Ball Mapping ..................................................................................................................... 22 Host Registers .................................................................................................................................. 50 Frame Alignment Criteria.................................................................................................................. 58 Performance Monitor Functional Descriptions.................................................................................. 59 Performance Report Message Format ............................................................................................. 61 Performance Report Message Field Definition................................................................................. 62 Shared Tx Stack Format for CEPT Frame ....................................................................................... 65 Cell Headers of Idle, Unassigned, and Invalid Cells......................................................................... 73 TC Functionality................................................................................................................................ 73 TAAD08JU2 Exceptions to the IMA PICS Proforma ........................................................................ 85 PATM Fields ................................................................................................................................... 114 HPF Fields...................................................................................................................................... 115 ESI Message Format (AALXDATA[15:0])....................................................................................... 116 ESI Violation Codings..................................................................................................................... 117 AAL Type vs. Service Type Compatibility....................................................................................... 118 Transport of Congestion Indication and Loss Priority..................................................................... 121 PortIndex to Enqueue Block Port Mapping..................................................................................... 123 MEMI-SM Resources ..................................................................................................................... 124 SQASE-SM Resources .................................................................................................................. 130 L1Q and IL2Q Scheduling .............................................................................................................. 135 Example Stage-Two Divider Settings ............................................................................................. 137 Exceptions ...................................................................................................................................... 141 Absolute Maximum Ratings............................................................................................................ 154 Power Requirements ...................................................................................................................... 154 Operating Conditions...................................................................................................................... 154 Handling Precautions ..................................................................................................................... 155 Version 2.1 Logic Interface Characteristics .................................................................................... 155 Version 3.1 Logic Interface Characteristics .................................................................................... 156 Versions 2.1 and 3.1 Main System Clock (GCLK) Timing Specifications....................................... 157 Version 2.1 UTOPIA Input Clocks (UCLK) Timing Specifications .................................................. 157 Version 3.1 UTOPIA Input Clocks (UCLK_A[B]) Timing Specifications ......................................... 157 Host Read Timing Characteristics .................................................................................................. 158 Host Write Timing Characteristics .................................................................................................. 159 Version 2.1 CHI Transmit Timing Characteristics........................................................................... 161 Version 3.1 CHI Transmit Timing Characteristics........................................................................... 161 CHI Receive Timing Characteristics............................................................................................... 162 Version 2.1 Fabric Interface Timing Specifications (Transmit Interface) ........................................ 162 Version 3.1 Fabric Interface Timing Specifications (Transmit Interface) ........................................ 163 Version 2.1 Fabric Interface Timing Specifications (Receive Interface) ......................................... 163 Version 3.1 Fabric Interface Timing Specifications (Receive Interface) ......................................... 163 Agere Systems Inc. Agere Systems - Proprietary Use pursuant to Company instructions 9 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 List of Tables (continued) Table Table 53. Table 54. Table 55. Table 56. Table 57. Table 58. Table 59. Table 60. Table 61. Table A-1. 10 Page Expansion UTOPIA2 Receive Interface Timing Specifications: 50 MHz ........................................ UTOPIA2 Transmit Interface Timing Specifications: 50 MHz......................................................... Version 2.1 ESI Interface Timing Specifications............................................................................. Version 3.1 ESI Interface Timing Specifications (SAR and APC) .................................................. JTAG Timing Specifications ........................................................................................................... Version 2.1 Receive Interface Timing............................................................................................. Version 2.1 Transmit Interface Timing............................................................................................ Version 3.1 Receive Interface Timing (SUCLK Input) .................................................................... Version 3.1 Transmit Interface Timing (SUCLK Input) ................................................................... Revision History.............................................................................................................................. Agere Systems - Proprietary Use pursuant to Company instructions 164 164 164 164 165 165 165 166 166 172 Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 4 Description TAAD08JU2 provides a flexible network-interface solution for next-generation applications in which efficient transport of narrowband voice and broadband data information is critical to guaranteeing network QoS for the user and transmission efficiency for the network operator. Constructed using Agere’s 0.16 µm CMOS technology, the chip has an integrated octal framer, IMA processor, cell scheduler and router, and AAL2/5 SAR functions. TAAD08JU2 operates in either UNI or IMA mode (selectable on a per-span line basis). The complete AF-PHY0086.001 management information base (MIB) is supported. Flexible provisioning of link and group combinations enables a mix of IMA and UNI mappings to various AAL services. Support for AAL2 is provided via an AAL/CPS function that maps/demaps variable-sized CPS packets to/from ATM-SDU. A total of 2032 bidirectional CIDs are supported. These CIDs can be transported within a programmable number of VCs per direction. TAAD08JU2 supports up to 124 AAL2 VCs, which may be allocated between ingress and egress traffic. Support for high-speed data switching is provided whereby AAL5 VCs are routed through to the system interface toward their destinations. TAAD08JU2 provides support for up to 2032 bidirectional AAL5 VCs via an internal context memory. TAAD08JU2 provides the following: ! Integrated policing ! F4/F5 operations, administration, and maintenance (OAM) ! Cell processing ! Statistics collection for performance monitoring Communication with TAAD08JU2 is accomplished through a 32-bit microprocessor interface. The system interface is through two choices: a UTOPIA 2 interface with support for both 8-bit and 16-bit data bus width and a UTOPIAderived packet interface with support for both 8-bit and 16-bit data bus widths. TAAD08JU2 provides a complete ATM access function from AAL/CPS mapping functions (for AAL2 and 5) through ATM/TC/PHY layers. The highly integrated, flexible architecture results in unified OAM features, simpler operation, and best-in-class operation with respect to area, power, and function. Agere Systems Inc. Agere Systems - Proprietary 11 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 5 Pin Definitions Table 1. Pin Definitions Type I Iu Id O O-6 I/O P Description Input only. All 3.3 V inputs are designed to be TTL compatible. Input with high-value pull-up resistor internal to TAAD08JU2. Input with high-value pull-down resistor internal to TAAD08JU2. Output only. These outputs have IOL/IOH = 10 mA. Output only. These outputs have IOL/IOH = 6 mA. Bidirectional input and output. Power or ground. 6 Pin Description Many of the pins of the TAAD08JU2 device are multiplexed for different functions. In these cases, both functions are shown in the same row of Table 2. 12 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 6 Pin Description (continued) Table 2. Transmission Line Interface Signals (48 Signals) Signal LRXCLK(0:7) LRXPDATA(0:7)/ LRXDATA(0:7) Type Id Id LRXNDATA(0:7)/ LRXBPV(0:7) Id LTXCLK(0:7) Id/O Description Line Interface Receive Clocks. Receive path clock from the LIU. Line Interface Receive Positive Rail Data/Line Interface Receive Data. When the TAAD08JU2 device is configured to operate in a dual-rail line interface mode, this pin is the positive receive data from the external LIU. When the TAAD08JU2 device is configured to operate in a single-rail line interface mode, this pin is the receive data from the external LIU. Line Interface Receive Negative Rail Data/Line Interface Receive Bipolar Violations. When the TAAD08JU2 device is configured to operate in a dual-rail line interface mode, this pin is the negative receive data from the external LIU. When the TAAD08JU2 device is configured to operate in a single-rail line interface mode, this pin is the receive bipolar violations signal from the external LIU. Line Interface Transmit Clock. These pins can be individually programmed as either a clock input or output in one of three modes: 1. Global clock mode. All the LTXCLK signals are outputs derived from a global clock input signal (see CHI interface CRXCLK in Table 3). 2. All of the LTXCLK signals are outputs of the corresponding LRXCLK signals looped back internally. 3. Independent transmit clock mode. Each LTXCLK pin is an input from the line interface. LTXPDATA(0:7)/ LTXDATA(0:7) O LTXNDATA(0:7) O Agere Systems Inc. The clock rates, when used as either inputs or outputs, are 1.544 MHz or 2.048 MHz. When the CHI interface is active, these pins must be configured in the global clock mode. Line Interface Transmit Positive Rail Data/Line Interface Transmit Data. When the TAAD08JU2 device operates in a dual-rail line interface mode, this pin is the positive transmit data sent to the external LIU. When the TAAD08JU2 device operates in a single-rail line interface mode, this pin is the transmit data sent to the external LIU. Line Interface Transmit Negative Rail Data. When the TAAD08JU2 device operates in a dual-rail line interface mode, this pin is the negative transmit data sent to external LIU. When the TAAD08JU2 device operates in a single-rail line interface mode, this pin outputs an 8 kHz frame sync pulse. Agere Systems - Proprietary 13 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 6 Pin Description (continued) Table 3. CHI Interface Signals (20 Signals) Signal CRXCLK Type I Description CHI Receive Clock. This signal is used to perform two basic functions: 1) this pin is used to clock the CHI receive interface; 2) depending on transmit line clock mode, the clock on this pin can be used to drive the Tx line clock (of which there are several suboptions). These suboptions are as follows: ! When the receive CHI interface is used, then 2.048 MHz, 4.096 MHz, 8.192 MHz, or 16.384 MHz can be supplied. Internally, TAAD08JU2 will derive 1.544 MHz and 2.048 MHz to support T1, E1, or a mix of T1 and E1 lines. ! If the CHI is not used, this pin can be used to drive the line clocks in two submodes, as follows: — This pin would directly drive the Tx interface. In this mode, 1.544 MHz (T1) or 2.048 MHz (E1) is applied to CRXCLK, and all eights links run at this line rate. — A 2.048 MHz reference is applied to CRXCLK and TAAD08JU2 internally derives 1.544 MHz and 2.048 MHz to support either T1, E1, or a mix of T1/E1 transmit lines. This mode is essentially a subset of option 1 above, except the CHI is not used. ! This pin is not used when the TAAD08JU2 is programmed into independent transmit clock mode or receive loop timing mode. The CRXCLK pin does not require a clock to be connected if that pin is not being used for any framer or CHI clocking modes. However, if a clock is required, this can be easily accommodated by connecting GCLK to this pin. CHI Receive Data. These are the received CHI data inputs at 2.048 Mbits/s, 4.096 Mbits/s, or 8.192 Mbits/s. CHI Receive Frame Sync. Global 8 kHz frame sync for the receive CHI ports. CHI Transmit Clock. Global system clock for transmit defined as a 2.048 MHz, 4.096 MHz, 8.192 MHz, or 16.384 MHz global input clock. CHI Transmit Data. These are the transmitted CHI data outputs clocked by the CTXCLK at 2.048 Mbits/s, 4.096 Mbits/s, or 8.192 Mbits/s. CHI Transmit Frame Sync. Input global 8 kHz frame sync for the transmit system. ! CRXDATA(0:7) I CRXFS CTXCLK I I CTXDATA(0:7) O CTXFS I Table 4. UTOPIA 2 Expansion Interface Signals (52 Signals) Signal UMODE UCLK 14 Type I I Description UTOPIA Expansion Interface Mode. This pin sets the mode of operation for this interface. The modes are described below: 0: Master Mode: This signal is set low when TAAD08JU2 is programmed to operate in either internal or external PHY mode. In this case, the internal APC block controls the expansion interface pins as a UTOPIA master. 1: Slave Mode: This is set high when the TAAD08JU2 is programmed into SAR-only mode. In this mode, this interface connects to the SAR block as a UTOPIA slave. UTOPIA Expansion Clock. This is the UTOPIA clock input for both the transmit and receive UTOPIA. The clock frequency applied to this pin should be less than or equal to GCLK. A clock must be supplied on this pin at all times. Even if this interface is not otherwise used, a clock must still be provided. Typically, this can be easily accommodated by connecting GCLK to this pin. Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 6 Pin Description (continued) Table 4. UTOPIA 2 Expansion Interface Signals (52 Signals) (continued) Signal URXDATA[15:0] URXPRTY URXSOC URXENB URXADDR[4:0] URXCLAV UTXDATA[15:0] UTXPRTY UTXSOC UTXENB UTXADDR[4:0] UTXCLAV Type Description d Master: I UTOPIA Expansion Receive Data. In master mode, these signals are the parSlave: O allel 16-bit data input bus from an external PHY device. In slave mode, these signals are a data output bus to an external PHY device. These signals are clocked in/out on the rising edge of UCLK. Bit URXDATA[15] is the MSB. Master: Id UTOPIA Expansion Receive Data Parity. This signal either receives (master Slave: O mode) or sends (slave mode) the receive data parity signal. When a master, TAAD08JU2’s APC block can be configured to check for odd parity or can be disabled. When a slave, this can be configured to odd, even, or no parity. d Master: I UTOPIA Expansion Receive Start of Cell. Active-high signal asserted when Slave: O URXDATA contains the first word of a cell. Master: O UTOPIA Expansion Receive Enable. Active-low signal asserted by the ATM Slave: Iu layer to signal that a transfer will occur at the next rising edge of UCLK. Master: O UTOPIA Expansion Receive Address. 5-bit address used by the UTOPIA Slave: Id master to select the UTOPIA slave for the receive signal path. Bit 4 is the MSB. Master: I UTOPIA Expansion Receive Cell Available. Active-high signal asserted when Slave: O a complete cell is available in the FIFO of the device selected by URXADDR. Master: O UTOPIA Expansion Transmit Data. In master mode, these signals are a paralSlave: Id lel 16-bit data output bus to an external PHY device. In slave mode, these signals are a parallel 16-bit data input bus from an external PHY device. Data is clocked out/in on the rising edge of UCLK. Bit UTXDATA[15] is the MSB. Master: O UTOPIA Expansion Transmit Data Parity. This signal either sends (master Slave: Id mode) or receives (slave mode) the transmit data parity signal. In slave mode, this can be configured to odd, even, or no parity on UTXDATA bus. In master mode, transmit parity can be either odd or disabled. The default is odd. Master: O UTOPIA Expansion Transmit Start of Cell. Active-high signal asserted when Slave: Id UTXDATA contains the first word of a cell. Master: O UTOPIA Expansion Transmit Enable. Active-low signal asserted by the ATM Slave: Iu layer to signal that UTXDATA and UTXSOC contain valid data. Master: O UTOPIA Expansion Transmit Address. 5-bit address used by the master to Slave: Id select the UTOPIA slave for the transmit signal path. Bit 4 is the MSB. Master: Id UTOPIA Expansion Transmit Cell Available. Active-high signal asserted Slave: O when the polled slave is ready to receive complete cell can be stored in the FIFO of the device selected by UTXADDR. Table 5. System Interface Signals (62 Signals) Signal SMODE[2:0] Type Id Description System Interface Mode. The two LSBs (SMODE[1:0]) determine the operating mode of the interface, while SMODE[2] determines the clock mode in UTOPIA cell and packet modes. SMODE[1:0] Description 00: UTOPIA mode 01: Packet (UT2+) mode 10: Unused 11: Unused SMODE[2] sets the clock mode for the system interface. A low on this pin causes TAAD08JU2 to input SUCLK, and a high sets TAAD08JU2 to generate SUCLK. Agere Systems Inc. Agere Systems - Proprietary 15 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 6 Pin Description (continued) Table 5. System Interface Signals (62 Signals) (continued) Signal SUCLK Type Id/O STXDATA[15:0] O STXADDR[4:0] O STXSOC/STXSOP O STXPRTY O STXENB O STXEOP O STXSIZ O STXERR Id STXCLAV/STXPA Id STXSPA Id SRXDATA[15:0] Id SRXADDR[4:0] O SRXSOC/SRXSOP Id SRXPRTY Id SRXENB O SRXEOP Id 16 Description System Interface Clock. Pin programmable to be an input or output. The clock frequency applied to this pin should be less than or equal to GCLK. System Interface Transmit Data. Parallel data bus to the ATM layer clocked out on the rising edge of SUCLK. Bit 15 is the MSB. System Transmit Address. 5-bit address used to select the external UTOPIA slave for the transmit signal path. System Transmit Start of Cell/Packet. In cell or packet mode, when STXSOC is high, the first word of the packet is present on the STXDATA bus. STXSOC is considered valid only when STXENB is asserted and is updated on the rising edge of SUCLK. System Transmit Data Parity. This signal sends the parity bits for the STXDATA bus. System Transmit Enable. Active-low signal asserted by the ATM layer to signal that STXDATA contains valid data. System UT2+ Transmit End of Packet. This signal is high when the last word of a packet is on the STXDATA bus. STXEOP is valid only when STXENB is asserted and is updated on the rising edge of SUCLK (UT2+ mode only). System UT2+ Transmit Size. This signal indicates the size of the current word on STXDATA. STXSIZ is valid only when STXEOP is asserted. If the last word contains two valid bytes, STXSIZ is high while that word is on the STXDATA bus (16-bit UT2+ mode only). System UT2+ Transmit Error. SRXERR is an active-high signal that indicates when the current packet is to be aborted and discarded, if possible. STXERR is valid only when STXEOP and STXENB are asserted and is sampled on the rising edge of SUCLK (UT2+ mode only). System Transmit Cell/Packet Available. Active-high signal asserted when a complete cell/packet can be stored in the FIFO of the external device selected by STXADDR. System UT2+ Transmit Selected Multi-PHY Packet Available. While STXCLAV shows the polled status of the external UTOPIA slave, this signal indicates the status of the current selected external slave. When asserted, this signal indicates that the current selected slave has more space than the predefined space in its FIFO (UT2+ mode only). System Receive Data. This signal is the parallel 16-bit data bus to the ATM layer clocked out on the rising edge of SUCLK. Bit SRXDATA[15] is the MSB. System Receive Address. 5-bit address used to select the external UTOPIA slave for the receive signal path. SRXADDR[4] is the MSB. System Receive Start of Cell/Packet. Active-high signal asserted when SRXDATA contains the first word of a cell or packet. System Receive Data Parity. Programmable for odd, even, or no parity over SRXDATA. System Receive Enable. Active-low signal asserted by the ATM layer to signal that a transfer will occur at the next rising edge of SUCLK. System UT2+ Receive End of Packet. This signal is active-high, and it indicates that the last word of a packet is on the SRXDATA bus. SRXEOP is valid when SRXENB is asserted and is sampled on the rising edge of SUCLK. (UT2+ mode only.) Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 6 Pin Description (continued) Table 5. System Interface Signals (62 Signals) (continued) Signal SRXSIZ Type Id SRXERR Id SRXCLAV/SRXPA Id SRXVAL Id Description System UT2+ Receive Size. This signal indicates the size of the current word on SRXDATA. SRXSIZ is valid when SRXEOP is asserted. A logic one indicates that SRXDATA[15:0] are valid, and a logic zero indicates that SRXDATA[15:8] are valid. (UT2+ mode only.) System UT2+ Receive Error. This is an active-high signal that indicates that the current packet is to be aborted and discarded, if possible. SRXERR is only valid when SRXEOP and SRXENB are asserted and is sampled on the rising edge of SUCLK. (UT2+ mode only.) System Receive Cell/Packet Available. In cell mode, when asserted, this signal indicates that a subsequent cell is available after the current transfer. In packet mode, when asserted, it indicates that more data than the predefined amount is available. System UT2+ Receive Data Valid. This is an active-high signal asserted by a slave device when in UT2+ mode to indicate that data is valid on the current clock cycle. This signal allows for the slave device to control data flow by deasserting this signal, thus pausing the current packet transmission. When the slave has valid data to put on the data bus, it will resume transmission of the current packet by asserting SRXVAL. For every clock cycle that there is valid data on the data bus, SRXVAL must be asserted. SRXVAL is a shared 3-state signal between all active MPHYs and only the currently selected MPHY may drive this signal. Table 6. Switch Fabric Interface Signals (50 Pins) Signal Type Switch Fabric A Port AATXDATA[7:0] O-6 AATXPRTY O-6 AATXSOC O-6 AATXCLKP O-6 AATXCLKN O-6 AARXDATA[7:0] Id AARXPRTY Id AARXSOC Id AARXCLKP Id Agere Systems Inc. Description APC Port A Transmit Data. Parallel data bus used to transfer cells from TAAD08JU2 to the switch fabric. APC Port A Transmit Parity. Odd parity calculated over AATXDATA. Odd parity means an odd number of ones including the parity bit. APC Port A Transmit Start of Cell. Active-high signal asserted when AATXDATA contains the first word of a cell. APC Port A Transmit Differential Clock Positive. This clock is the reference that is sent with the AATXDATA and is used by the receiving APC or switch fabric to clock in the data. This clock is derived from GCLK and is twice the GCLK frequency. APC Port A Transmit Differential Clock Negative. This clock is the reference that is sent with the AATXDATA and is used by the receiving APC or switch fabric to clock in the data. This clock is derived from GCLK and is twice the GCLK frequency. APC Port A Receive Data. Parallel data bus used to transfer cells from the switch fabric to TAAD08JU2. APC Port A Receive Parity. Odd parity calculated over AARXDATA. Odd parity means an odd number of ones including the parity bit. APC Port A Receive Start of Cell. Active-high signal asserted when AARXDATA contains the first word of a cell. APC Port A Receive Differential Clock Positive. This clock is used by TAAD08JU2 to clock AARXDATA into the device. This clock is typically twice GCLK. Agere Systems - Proprietary 17 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 6 Pin Description (continued) Table 6. Switch Fabric Interface Signals (50 Pins) (continued) Signal AARXCLKN Switch Fabric B Port ABTXDATA[7:0] Type Id Description APC Port A Receive Differential Clock Negative. This clock is used by TAAD08JU2 to clock AARXDATA into the device. This clock is typically twice GCLK. O-6 APC Port B Transmit Data. Parallel data bus used to transfer cells from TAAD08JU2 to the switch fabric. ABTXPRTY O-6 APC Port B Transmit Parity. Odd parity calculated over ABTXDATA. Odd parity means an odd number of ones including the parity bit. ABTXSOC O-6 APC Port B Transmit Start of Cell. Active-high signal asserted when ABTXDATA contains the first word of a cell. ABTXCLKP O-6 APC Port B Transmit Differential Clock Positive. The frequency is derived from GCLK. Maximum frequency is 100 MHz. ABTXCLKN O-6 APC Port B Transmit Differential Clock Negative. The frequency is derived from GCLK. Maximum frequency is 100 MHz. d ABRXDATA[7:0] I APC Port B Receive, bits 7:0. Parallel data bus used to transfer cells from the switch fabric to TAAD08JU2. ABRXPRTY Id APC Port B Receive Parity. Odd parity calculated over ABRXDATA. Odd parity means an odd number of ones, including the parity bit. d ABRXSOC I APC Port B Receive Start of Cell. Active-high signal asserted when ABRXDATA contains the first word of a cell. ABRXCLKP Id APC Port B Receive Clock Positive. This clock is used by TAAD08JU2 to clock the ABRXDATA into the device. This clock is typically twice GCLK. d ABRXCLKN I APC Port B Receive Clock Negative. This clock is used by TAAD08JU2 to clock the ABRXDATA into the device. This clock is typically twice GCLK. Switch Fabric Miscellaneous Signals AGTSYNC O Global Time-Slot (Cell Time) Synchronization Pulse. Asserted high once every 68 cycles of internal APC clock (which is 2 x GCLK). AHPSWF Id Agere Test Mode Pin. Should be tied to ground. Table 7. APC External Statistics Interface Signals (18 Signals) Signal AEDATA[15:0] Type O AECLK O AESYNC O Description APC External Statistics Data. A 16-bit data bus used to transfer data between the APC and an optional external adjunct device. APC External Statistics Clock. Used as a reference to transfer data between the APC and an external adjunct. APC External Statistics Sync. A single-cycle pulse signaling the beginning of the 34 clock cycle (AECLK) external statistics interface time slot. The absence of this synchronization pulse indicates that cell processing is disabled. Table 8. SAR External Statistics Interface Signals (18 Signals) Signal REDATA[15:0] RECLK 18 Type O O Description SAR External Statistics Data. A 16-bit data bus used to transfer data between the SAR and an optional external adjunct device. SAR External Statistics Clock. Used as reference to transfer data between the SAR and an external adjunct. Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 6 Pin Description (continued) Table 8. SAR External Statistics Interface Signals (18 Signals) Signal RESYNC Type O Description SAR External Statistics Sync. A single-cycle pulse signaling the beginning of the (RECLK) SAR statistics interface time slot. The absence of this synchronization pulse indicates that cell processing is disabled. Table 9. Host Interface Signals (49 Signals) Signal HMODE[1:0] HCLK Type Id Description Host Interface Mode Select. Selects the mode of operation of the microprocessor interface. These pins must be connected to ground. Host Interface Clock. This interface is rising edge clocked. Data written to TAAD08JU2 is latched on the rising edge of clock, and address information and data outputs on the rising edge of clock. The maximum speed for this clock is 66 MHz. Host Data Bus. This is a bidirectional 32-bit data bus used to transfer data to/from TAAD08JU2. Host Address Inputs. 10-bit address for register read or write operations. This addressing is by 32-bit word. Host Chip Select. The active-low signal validates HA[9:0] for read and write transfers. Host Write Enable. 0 causes an active-low write, and 1 causes a read. Host Advance. A high signal on this pin causes TAAD08JU2 to increment a previous host address by one. Host Slave Mode Interrupt. Active-low interrupt request signal from TAAD08JU2. I HD[31:0] I/O HA[9:0] Id HCEN HWEN HADV Iu Iu Id HIRQ O Table 10. JTAG Interface Pins (6 Signals) Signal TMODE TMS TDI TRSTN TDO TCK Type Id Description JTAG Test Mode. This pin is an input with an internal pull-down. TMODE = 1 is reserved for Agere testing. Test Mode Select. This pin enables JTAG test mode. Pin has an internal pull-up. Test Data Input. Serial test input during JTAG testing. Test Reset. This signal must be asserted low on powerup. Test Data Output. Serial test output during JTAG testing. Test Clock. An internal pull-up exists on this pin. This pin is used to clock state and test data into and out of the TAAD08JU2 during JTAG testing. Iu Iu Iu O Iu Table 11. Global/Miscellaneous Signal Pins (10 Signals) Signal GOE Type Iu GCLK I Agere Systems Inc. Description Global Output Enable. When GOE is 0, all TAAD08JU2 outputs assume a highimpedance state except TDO. When GOE is 1, all outputs operate normally. An internal pull-up is provided on this pin. Global Clock. Maximum clock frequency is 52 MHz. All output clocks are derived from this clock. GCLK should have a minimum 60/40 duty cycle and a maximum frequency tolerance of ±0.05%. GCLKs frequency can range from 25 MHz to 50 MHz. When running slower than 50 MHz, the maximum throughput of TAAD08JU2 (which is 155 Mbits/s at 50 MHz) is degraded proportionately. Agere Systems - Proprietary 19 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 6 Pin Description (continued) Table 11. Global/Miscellaneous Signal Pins (10 Signals) (continued) Signal GRESET Type Iu GPLLBYP Id GPLLOUT O FPLLOUT O SCANMODE SCANCLK1 Id Id SCANCLK2 Id IDDQ I Description Global Reset. Active-low reset signal. On initial powerup, GRESET must be asserted for at least 250 µs after stable clocks are provided to TAAD08JU2, in order to allow the internal PLLs to stabilize. GCLK and HCLK must be continuously applied during reset. When asserted, all internal circuity is reset to its default condition. If TAAD08JU2 has already been powered on and operating, the device can be reset by asserting GRESET for at least 8 clock cycles of GCLK and 8 clock cycles of HCLK. Note that both of these clock signals must be applied for TAAD08JU2 to be properly reset. GRESET must be inactive for 2 ms before boot sequence can commence. Global PLL Bypass. This pin is used to bypass the operation of the global clock synthesizer PLL. This pin is intended for Agere manufacturing testing, and must be tied low for normal operation. Global PLL Output. A reference clock output of the PLL used only for Agere test purposes. Framer PLL Output. A reference clock output of the PLL used only for Agere test purposes. Scan Mode. This pin is used by Agere to scan test this device. Scan Clock 1. This pin enables Agere internal scan testing, and should be tied low for normal operation. Scan Clock 2. This pin enables Agere internal scan testing, and should be tied low for normal operation. IDDQ Test Mode Enable. This pin should be tied high for normal operation. When tied low, this pin is used for IDDQ testing. Table 12. Power Supply Pins (4 Analog Power Pins, 120 Digital Power Pins) Signal GVDDA Type P GVSSA P FVDDA P FVSSA P VDD33 VDD15 VSS P P P 20 Description Global PLL Power Supply. Separate dedicated 3.3 V power supply to power the global clock synthesizer PLL. This power supply should be a low-noise supply. Global PLL Ground. Separate 3.3 V ground to power the global clock synthesizer PLL. This should be a low-noise ground. Framer PLL Power Supply. Separate dedicated 3.3 V power supply to power the framer PLL. This power supply should be a low-noise supply. Framer PLL Ground. Separate dedicated 3.3 V ground to power the framer PLL. This should be a low-noise ground. 3.3 V Digital Power Pins. 1.5 V Digital Power Pins. Digital Ground Pins. Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 7 Package Pin Layout TAAD08JU2 uses the SuperBGATM EBGA-520 package (1.27 mm. pitch). ! Body size: 40 x 40 x 0.78 mm ! Maximum height off-board: 1.67 mm ! Solder ball pitch: 1.27 mm ! Thermal, ΘJA: 11 oC/W 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 A 1.5 VSS VSS VSS 394 VSS 386 381 376 374 VSS 366 361 356 351 VSS 350 345 340 335 VSS 327 325 320 315 VSS 307 VSS VSS VSS 1.5 B VSS 1.5 VSS 399 396 391 388 383 378 3.3 371 367 362 357 352 VSS 349 344 339 334 330 3.3 323 318 313 310 305 302 VSS 1.5 VSS C VSS VSS 1.5 3.3 398 393 389 384 379 3.3 372 368 363 358 353 1.5 348 343 338 333 329 3.3 322 317 312 308 303 3.3 1.5 VSS VSS D VSS 2 E 7 5 3.3 1.5 400 395 390 385 380 375 373 369 364 359 354 1.5 347 342 337 332 328 326 321 316 311 306 301 1.5 3.3 299 VSS 3 1 8 1.5 397 392 387 382 377 3.3 370 365 360 355 1.5 346 341 336 331 3.3 324 319 314 309 304 1.5 300 298 296 294 6 4 297 295 293 291 VSS G 15 13 12 11 9 292 290 289 288 286 H 20 18 17 16 14 287 285 284 283 281 J 25 23 22 21 19 282 280 279 278 276 K 27 3.3 3.3 26 24 F VSS 10 TOP VIEW 277 275 3.3 3.3 274 L VSS 30 29 28 3.3 M 35 34 33 32 31 3.3 273 272 271 VSS N 40 39 38 37 36 P 45 44 43 42 41 R 50 49 48 47 46 T VSS VSS 1.5 1.5 1.5 U 51 52 53 54 55 V 56 57 58 59 60 W 61 62 63 64 65 236 237 238 239 240 Y 66 67 68 69 70 231 232 233 234 235 AA VSS 71 72 73 3.3 3.3 228 229 230 VSS VSS 270 269 268 267 266 265 264 263 262 261 VDD15 (1.5 V) 260 259 258 257 256 255 254 253 252 251 VDD33 (3.3 V) 1.5 1.5 1.5 VSS VSS 246 247 248 249 250 SIGNAL PIN/SPECIAL VDD-VSS 241 242 243 244 245 AB 74 3.3 3.3 75 77 224 226 3.3 3.3 227 AC 76 78 79 80 82 219 221 222 223 225 AD 81 83 84 85 87 214 216 217 218 220 AE 86 88 89 90 92 209 211 212 213 215 AF VSS 91 93 95 97 204 206 208 210 VSS AG 94 96 98 100 1.5 104 109 114 119 124 3.3 131 136 141 146 1.5 155 160 165 170 3.3 177 182 187 192 197 1.5 201 203 205 207 AH VSS 99 3.3 1.5 101 106 111 116 121 126 128 132 137 142 147 1.5 154 159 164 169 173 175 180 185 190 195 200 1.5 3.3 202 VSS AJ VSS VSS 1.5 3.3 103 108 112 117 122 3.3 129 133 138 143 148 1.5 153 158 163 168 172 3.3 179 184 189 193 198 3.3 1.5 VSS VSS AK VSS 1.5 VSS 102 105 110 113 118 123 3.3 130 134 139 144 149 VSS 152 157 162 167 171 3.3 178 183 188 191 196 199 VSS 1.5 VSS AL 1.5 VSS VSS VSS 107 VSS 115 120 125 127 VSS 135 140 145 150 VSS 151 156 161 166 VSS 174 176 181 186 VSS 194 VSS VSS VSS 1.5 5-9957(F) Figure 1. Pin Configuration Diagram Agere Systems Inc. Agere Systems - Proprietary 21 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Preliminary Data Sheet August 18, 2003 7 Package Pin Layout (continued) (Pins shown as Reserved must be left unconnected). Table 13. Signal-to-Ball Mapping 22 Ball Pad Number Data Sheet Ball Name E4 1 AEDATA[15] D2 2 AEDATA[14] E3 3 AEDATA[13] F5 4 AEDATA[12] E2 5 AEDATA[11] F4 6 AEDATA[10] E1 7 AEDATA[9] F3 8 AEDATA[8] G5 9 AEDATA[7] F2 10 AEDATA[6] G4 11 AEDATA[5] G3 12 AEDATA[4] G2 13 AEDATA[3] H5 14 AEDATA[2] G1 15 AEDATA[1] H4 16 AEDATA[0] H3 17 URXDATA[15] H2 18 URXDATA[14] J5 19 URXDATA[13] H1 20 URXDATA[12] J4 21 URXDATA[11] J3 22 URXDATA[10] J2 23 URXDATA[9] K5 24 URXDATA[8] J1 25 URXDATA[7] K4 26 URXDATA[6] K1 27 URXDATA[5] L4 28 URXDATA[4] L3 29 URXDATA[3] L2 30 URXDATA[2] M5 31 URXDATA[1] M4 32 URXDATA[0] M3 33 URXENB M2 34 URXPRTY M1 35 URXSOC N5 36 URXCLAV N4 37 URXADDR[0] N3 38 URXADDR[1] Table 13. Signal-to-Ball Mapping (continued) Ball Pad Number Data Sheet Ball Name N2 39 URXADDR[2] N1 40 URXADDR[3] P5 41 URXADDR[4] P4 42 UCLK P3 43 UTXADDR[4] P2 44 UTXADDR[3] P1 45 UTXADDR[2] R5 46 UTXADDR[1] R4 47 UTXADDR[0] R3 48 UTXCLAV R2 49 UTXSOC R1 50 UTXPRTY U1 51 UTXENB U2 52 UTXDATA[15] U3 53 UTXDATA[14] U4 54 UTXDATA[13] U5 55 UTXDATA[12] V1 56 UTXDATA[11] V2 57 UTXDATA[10] V3 58 UTXDATA[9] V4 59 UTXDATA[8] V5 60 UTXDATA[7] W1 61 UTXDATA[6] W2 62 UTXDATA[5] W3 63 UTXDATA[4] W4 64 UTXDATA[3] W5 65 UTXDATA[2] Y1 66 UTXDATA[1] Y2 67 UTXDATA[0] Y3 68 UMODE Y4 69 STXDATA[15] Y5 70 STXDATA[14] AA2 71 STXDATA[13] AA3 72 STXDATA[12] AA4 73 STXDATA[11] AB1 74 STXDATA[10] AB4 75 STXDATA[9] AC1 76 STXDATA[8] AB5 77 STXDATA[7] AC2 78 STXDATA[6] Agere Systems Inc. Preliminary Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 7 Package Pin Layout (continued) Table 13. Signal-to-Ball Mapping (continued) Table 13. Signal-to-Ball Mapping (continued) Ball Pad Number Data Sheet Ball Name Ball Pad Number Data Sheet Ball Name AC3 79 STXDATA[5] AG9 119 SRXDATA[9] AC4 80 STXDATA[4] AL8 120 SRXDATA[8] AD1 81 STXDATA[3] AH9 121 SRXDATA[7]7 AC5 82 STXDATA[2] AJ9 122 SRXDATA[6] AD2 83 STXDATA[1] AK9 123 SRXDATA[5] AD3 84 STXDATA[0] AG10 124 SRXDATA[4] AD4 85 STXCLAV/STXPA/ AL9 125 SRXDATA[3] AE1 86 STXSPA AH10 126 SRXDATA[2] AD5 87 STXSOC/STXSOP AL10 127 SRXDATA[1] AE2 88 STXPRTY AH11 128 SRXDATA[0] AE3 89 STXSIZ AJ11 129 SMODE2 AE4 90 STXENB AK11 130 SMODE1 AF2 91 STXERR AG12 131 SMODE0 AE5 92 STXEOP AH12 132 REDATA[15] AF3 93 RESERVED AJ12 133 REDATA[14] AG1 94 STXADDR[0] AK12 134 REDATA[13] AF4 95 STXADDR[1] AL12 135 REDATA[12] AG2 96 STXADDR[2] AG13 136 REDATA[11] AF5 97 STXADDR[3] AH13 137 REDATA[10] AG3 98 STXADDR[4] AJ13 138 REDATA[9] AH2 99 SRXADDR[4] AK13 139 REDATA[8] AG4 100 SRXADDR[3] AL13 140 REDATA[7] AH5 101 SRXADDR[2] AG14 141 REDATA[6] AK4 102 SRXADDR[1] AH14 142 REDATA[5] AJ5 103 SRXADDR[0] AJ14 143 REDATA[4] AG6 104 SUCLK AK14 144 REDATA[3] AK5 105 SRXVAL AL14 145 REDATA[2] AH6 106 SRXCLAV AG15 146 REDATA[1] AL5 107 SRXSOC AH15 147 REDATA[0] AJ6 108 SRXPRTY AJ15 148 RESYNC AG7 109 SRXEOP AK15 149 RECLK AK6 110 SRXERR AL15 150 HIRQ AH7 111 SRXENB AL17 151 HADV AJ7 112 SRXSIZ AK17 152 HWEN AK7 113 SRXDATA[15] AJ17 153 HCEN AG8 114 SRXDATA[14] AH17 154 HCLK AL7 115 SRXDATA[13] AG17 155 HA[9] AH8 116 SRXDATA[12] AL18 156 HA[8] AJ8 117 SRXDATA[11] AK18 157 HA[7] AK8 118 SRXDATA[10] AJ18 158 HA[6] <10pt>Agere Systems - Proprietary 23 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Preliminary Data Sheet August 18, 2003 7 Package Pin Layout (continued) Table 13. Signal-to-Ball Mapping (continued) 24 Table 13. Signal-to-Ball Mapping (continued) Ball Pad Number Data Sheet Ball Name Ball Pad Number Data Sheet Ball Name AH18 AG18 AL19 159 160 161 HA[5] HA[4] HA[3] AK28 199 RESERVED AK19 162 HA[2] AH27 AG28 AH30 AG29 AF27 AG30 AF28 AG31 AF29 AE27 AF30 AE28 AE29 AE30 AD27 AE31 AD28 AD29 AD30 AC27 AD31 AC28 AC29 AC30 AB27 AC31 AB28 AB31 AA28 AA29 AA30 Y27 Y28 Y29 Y30 Y31 W27 W28 W29 W30 W31 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED RESERVED TCK TDO TRSTN TDI TMS TMODE GOE IDDQ SCANMODE SCANCLK2 SCANCLK1 GPLLOUT GRESET GCLK GPLLBYP GVSSA GVDDA FPLLOUT FVSSA FVDDA RESERVED AJ19 163 HA[1] AH19 164 HA[0] AG19 165 HMODE1 AL20 166 HMODE0 AK20 167 HD[31] AJ20 168 HD[30] AH20 169 HD[29] AG20 170 HD[28] AK21 171 HD[27] AJ21 172 HD[26] AH21 173 HD[25] AL22 174 HD[24] AH22 175 HD[23] AL23 176 HD[22] AG22 177 HD[21] AK23 178 HD[20] AJ23 179 HD[19] AH23 180 HD[18] AL24 181 HD[17] AG23 182 HD[16] AK24 183 HD[15] AJ24 184 HD[14] AH24 185 HD[13] AL25 186 HD[12] AG24 187 HD[11] AK25 188 HD[10] AJ25 189 HD[9] AH25 190 HD[8] AK26 191 HD[7] AG25 192 HD[6] AJ26 193 HD[5] AL27 194 HD[4] AH26 195 HD[3] AK27 196 HD[2] AG26 197 HD[1] AJ27 198 HD[0] Agere Systems Inc. Preliminary Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 7 Package Pin Layout (continued) Table 13. Signal-to-Ball Mapping (continued) Table 13. Signal-to-Ball Mapping (continued) Ball Pad Number Data Sheet Ball Name Ball Pad Number Data Sheet Ball Name V27 241 RESERVED H31 281 LRXNDATA7/LRXBPV7 V28 242 RESERVED J27 282 LRXNDATA6/LRXBPV6 V29 243 RESERVED H30 283 LRXNDATA5/LRXBPV5 V30 244 RESERVED H29 284 LRXNDATA4/LRXBPV4 V31 245 RESERVED H28 285 LRXNDATA3/LRXBPV3 U27 246 RESERVED G31 286 LRXNDATA2/LRXBPV2 U28 247 RESERVED H27 287 LRXNDATA1/LRXBPV1 U29 248 RESERVED G30 288 LRXNDATA0/LRXBPV0 U30 249 RESERVED G29 289 LRXPDATA7/LRXDATA7 U31 250 RESERVED G28 290 LRXPDATA6/LRXDATA6 R31 251 RESERVED F30 291 LRXPDATA5/LRXDATA5 R30 252 RESERVED G27 292 LRXPDATA4/LRXDATA4 R29 253 RESERVED F29 293 LRXPDATA3/LRXDATA3 R28 254 RESERVED E31 294 LRXPDATA2/LRXDATA2 R27 255 RESERVED F28 295 LRXPDATA1/LRXDATA1 P31 256 RESERVED E30 296 LRXPDATA0/LRXDATA0 P30 257 RESERVED F27 297 LRXCLK7 P29 258 RESERVED E29 298 LRXCLK6 P28 259 RESERVED D30 299 LRXCLK5 P27 260 RESERVED E28 300 LRXCLK4 N31 261 RESERVED D27 301 LRXCLK3 N30 262 RESERVED B28 302 LRXCLK2 N29 263 RESERVED C27 303 LRXCLK1 N28 264 RESERVED E26 304 LRXCLK0 N27 265 RESERVED B27 305 LTXNDATA7 M31 266 RESERVED D26 306 LTXNDATA6 M30 267 RESERVED A27 307 LTXNDATA5 M29 268 RESERVED C26 308 LTXNDATA4 M28 269 RESERVED E25 309 LTXNDATA3 M27 270 RESERVED B26 310 LTXNDATA2 L30 271 RESERVED D25 311 LTXNDATA1 L29 272 RESERVED C25 312 LTXNDATA0 L28 273 RESERVED B25 313 LTXPDATA7/LTXDATA7 K31 274 RESERVED E24 314 LTXPDATA6/LTXDATA6 K28 275 RESERVED A25 315 LTXPDATA5/LTXDATA5 J31 276 RESERVED D24 316 LTXPDATA4/LTXDATA4 K27 277 RESERVED C24 317 LTXPDATA3/LTXDATA3 J30 278 RESERVED B24 318 LTXPDATA2/LTXDATA2 J29 279 RESERVED E23 319 LTXPDATA1/LTXDATA1 J28 280 RESERVED A24 320 LTXPDATA0/LTXDATA0 <10pt>Agere Systems - Proprietary 25 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Preliminary Data Sheet August 18, 2003 7 Package Pin Layout (continued) Table 13. Signal-to-Ball Mapping (continued) 26 Table 13. Signal-to-Ball Mapping (continued) Ball Pad Number Data Sheet Ball Name Ball Pad Number Data Sheet Ball Name D23 321 LTXCLK7 A13 361 ABTXDATA[7] C23 322 LTXCLK6 B13 362 ABTXDATA[6] B23 323 LTXCLK5 C13 363 ABTXDATA[5] E22 324 LTXCLK4 D13 364 ABTXDATA[4] A23 325 LTXCLK3 E13 365 ABTXDATA[3] D22 326 LTXCLK2 A12 366 ABTXDATA[2] A22 327 LTXCLK1 B12 367 ABTXDATA[1] D21 328 LTXCLK0 C12 368 ABTXDATA[0] C21 329 CTXDATA7 D12 369 ABTXCLKN B21 330 CTXDATA6 E12 370 ABTXCLKP E20 331 CTXDATA5 B11 371 ABTXSOC D20 332 CTXDATA4 C11 372 ABTXPRTY C20 333 CTXDATA3 D11 373 AARXDATA[7] B20 334 CTXDATA2 A10 374 AARXDATA[6] A20 335 CTXDATA1 D10 375 AARXDATA[5] E19 336 CTXDATA0 A9 376 AARXDATA[4] D19 337 CTXFS E10 377 AARXDATA[3] C19 338 CTXCLK B9 378 AARXDATA[2] B19 339 CRXDATA7 C9 379 AARXDATA[1] A19 340 CRXDATA6 D9 380 AARXDATA[0] E18 341 CRXDATA5 A8 381 AARXCLKN D18 342 CRXDATA4 E9 382 AARXCLKP C18 343 CRXDATA3 B8 383 AARXSOC B18 344 CRXDATA2 C8 384 AARXPRTY A18 345 CRXDATA1 D8 385 AATXDATA[7] E17 346 CRXDATA0 A7 386 AATXDATA[6] D17 347 CRXFS E8 387 AATXDATA[5] C17 348 CRXCLK B7 388 AATXDATA[4] B17 349 ABRXDATA[7] C7 389 AATXDATA[3] A17 350 ABRXDATA[6] D7 390 AATXDATA[2] A15 351 ABRXDATA[5] B6 391 AATXDATA[1] B15 352 ABRXDATA[4] E7 392 AATXDATA[0] C15 353 ABRXDATA[3] C6 393 AATXCLKN D15 354 ABRXDATA[2] A5 394 AATXCLKP E15 355 ABRXDATA[1] D6 395 AATXSOC A14 356 ABRXDATA[0] B5 396 AATXPRTY B14 357 ABRXCLKN E6 397 AGTSYNC C14 358 ABRXCLKP C5 398 AHPSWF D14 359 ABRXSOC B4 399 AESYNC E14 360 ABRXPRTY D5 400 AECLK Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 8 Block Diagram SWITCH FABRIC BUS TAAD08JU2 B TC IMA APC FRAMER A AAL ENGINE SYSTEM INTERFACE CHI TDM INTERFACE LINE INTERFACE UTOPIA EXPANSION INTERFACE EDC HOST PROCESSOR INTERFACE 0145(F) Figure 2. Architecture of the TAAD08JU2 Device As seen in Figure 2, TAAD08JU2 provides a complete ATM low-speed access function. In comparison to current alternative devices, TAAD08JU2 provides framing, transmission convergence, inverse multiplexing for ATM, ATM port management, and AAL SARing functions in a single, highly integrated device. Furthermore, TAAD08JU2 is architected to be flexible and scalable to effectively handle alternative higher-rate physical interfaces. TAAD08JU2 provides the following features as a highly integrated system on a chip (SOC): ! A complete, integrated, low-speed ATM access device solution. ! Flexible solution for transporting mixed traffic classes with QoS guarantees. ! System-on-a-chip performance with a simpler OAMP API to enhance time-to-market. TAAD08JU2 terminates a variety of low-speed physical link protocols (T1/E1/J1) via an integrated framer. Each link can carry ATM cell streams corresponding to multiple connections. Transmission convergence (TC) provides cell delineation through HEC generation and checking. TC also provides cell rate decoupling between the ATM and PHY layers through insertion/discard of idle ATM cells. The IMA block provides for inverse multiplexing over ATM using one to four groups with two to eight links per group. The ATM port controller (APC) block provides all of the functionality of the Agere APC device, such as switching, traffic shaping, and policing. The AAL engine provides a number of segmentation and reassembly options based on AAL2 and AAL5 standards while maintaining multiple traffic classes and qualities of service. Agere Systems Inc. Agere Systems - Proprietary 27 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 9 Software Components This section discusses the software that is used to initialize, configure, and operate TAAD08JU2. The purpose of this section is to introduce these important software components, to familiarize the user with them, and to provide links to the appropriate documentation. TAAD08JU2 is a highly integrated SOC. The numerous functional blocks (see chapter 8) that the system comprises are connected together and controlled by an internal 32 bit ARM® processor. This processor runs software (known as firmware) that controls all the internal workings of TAAD08JU2 and all the external communications with the user's system. Control of the TAAD08JU2 system occurs via a communications protocol between the firmware running on the internal ARM and software running on an external microprocessor (host processor). Commands are sent to the firmware and responses are returned using this protocol. A library of C functions translates high-level API function calls into the command format expected by firmware and implements the communications protocol. This library of software is known as device manager (DevMan for short). A separate utility is also used to create a binary setup file that configures TAAD08JU2's operating modes and initializes tables. This utility is the TAAD08JU2 setup file utility (formerly known as Newport Setup File Utility, NSFU for short). End users of TAAD08JU2 write application code that, through the device manager, controls the TAAD08JU2 in their systems. Many levels of software are used to abstract the user from the internal details of the hardware in TAAD08JU2. (There is no direct access to the internal registers in TAAD08JU2. Control is strictly through firmware commands and indications.) The user's view of TAAD08JU2 is through the high-level device manager API calls. Figure 3 shows the scope of software components surrounding the TAAD08JU2 device. SYSTEM SOFTWARE APPLICATION CODE SFU DEVICE MANAGER FIRMWARE NP Figure 3. Software Components The rectangular block at the center of the figure refers to the physical TAAD08JU2 (formerly known as Newport, NP for short) hardware device. The physical device has a set of control registers, an input buffer, and an output buffer. These registers and buffers are known as the embedded device controller (EDC) registers. All control and data transfer is done through these registers and buffers. From the software point of view, TAAD08JU2 is a memory-mapped device with interface registers used to send commands and exchange data with the internal devices. See chapter 13 for details on the TAAD08JU2 registers that are visible to the host processor. 28 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 9 Software Components (continued) 9.1 Firmware The TAAD08JU2 chip contains an embedded ARM processor to manage the internal devices and communicate with the host processor. The ARM processor has no software permanently stored in TAAD08JU2. The user's host processor must load the ARM’s instructions into RAM located inside the TAAD08JU2 before the TAAD08JU2 can operate. These instructions are known as firmware. The ARM firmware files are supplied with the TAAD08JU2 software release. The firmware can be viewed as the assembly language of TAAD08JU2. All user-controllable functions of TAAD08JU2 are implemented as firmware commands (i.e., assembly-language statements in a microprocessor). To execute a command, the binary parameters of the command are loaded into the TAAD08JU2 EDC input buffer by the host processor. The command word is then written to the EDC command register. The ARM receives an interrupt, and the firmware then reads the command word from the command register and interprets the command parameters. The firmware executes the command. The firmware places the results of executing the command into the EDC output buffer and writes an indication word into the EDC indication register. TAAD08JU2 hardware then generates an external interrupt to the host processor notifying it that firmware has made data available. This is the handshaking that occurs to exchange commands and data between the external processor and TAAD08JU2 firmware. In summary, the key points of the TAAD08JU2 ARM firmware are: ! TAAD08JU2 contains an embedded ARM processor—a true SOC. ! Two binary files contain the ARM instruction code and data. ! The instruction code for the ARM processor is not permanently stored in TAAD08JU2, and thus must be downloaded into TAAD08JU2 every time the device is powered on or reset. ! Application software never talks directly to the ARM. Only the device manager sends firmware commands and receives indications from the ARM. 9.2 Device Manager The device manager for TAAD08JU2 is an applications programming interface (API) written in C. This software is referred to as device manager software or simply the device manager, and the interface is called the device manager API. Programmers use this interface to access and control TAAD08JU2 from their applications. The device manager is coded using ANSI C and is compiled into a library of functions for the host platform. TAAD08JU2 is a very complex device. Communication with the ARM processor firmware in TAAD08JU2 is somewhat complicated. The API is designed to provide programmers with a level of abstraction that hides the complexity of this device. Figure 4 illustrates how the device manager API is used to communicate with TAAD08JU2. Agere Systems Inc. Agere Systems - Proprietary 29 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 9 Software Components (continued) HOST PROCESSOR APPLICATION CODE DEVICE MANAGER API TAAD08JU2 EDC REGISTERS ARM—FIRMWARE Figure 4. Device Manager APIs and TAAD08JU2 Communications As shown in Figure 4, the host application code (typically high-level management software) is linked with the device manager library. It uses the C functions exported by the library to access and control TAAD08JU2. TAAD08JU2’s host interface provides for the hardware registers that are used to pass information to and from the firmware software. The device manager layer sends commands to the firmware for services. The firmware uses indications to inform the device manager about events that occur on TAAD08JU2. The device manager interprets the indications and, if required, conveys the results to the application. The device manager hides the details of the hardware from the application. The device manager API deals in highlevel functionality, not device registers. It also hides device-specific details where possible. Applications use the device manager’s C function interface to request services provided by the device manager. The return value of the C function indicates whether the called function was successful or not. Refer to the TAAD08JU2 device manager API document for more details on device manager implementation and the available API functions. In summary, the key points of the TAAD08JU2 device manager are as follows: ! The device manager is a set of C functions that applications use to configure, control, and monitor TAAD08JU2. ! The device manager library is called by the application code. It is not a separate, stand-alone application. ! The device manager is the interface between application code and the TAAD08JU2 device. Application code never accesses TAAD08JU2 registers directly. Application code never sends firmware commands or receives firmware indications. ! The device manager translates application calls into firmware commands, which then are sent to the ARM in TAAD08JU2. ! The device manager handles alarm recording and statistics reported by TAAD08JU2. ! The device manager handles downloading the ARM firmware into TAAD08JU2 during start-up. 9.3 Setup File Utility (SFU) The TAAD08JU2 setup file utility (also known as Newport Setup File Utility, NSFU for short) runs on a Windows® PC and guides the user through the initial configuration of TAAD08JU2. The NSFU software package creates a binary setup file containing configuration commands, which are downloaded to TAAD08JU2 using a device manager API. 30 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 9 Software Components (continued) The user is encouraged to use the setup file utility to generate the sequence of setup commands and store them in a setup file. The setup utility allows the user to choose the mode information by simply clicking on the GUI to select options. It ensures that the user selects a valid combination of options by providing appropriate error messages and, where possible, by restricting the available options to the appropriate ones. This methodology brings the great advantage that the firmware is freed from having to perform error checking on all the provisioning parameters, which reduces the embedded code size. In addition, the GUI handles the translation of provisioning parameters into per-block provisioning parameters, also reducing the complexity of the initialization code. Once the host via device manager has downloaded the setup file to TAAD08JU2, the provisioning of the device is complete. The TAAD08JU2 waits for the next command from the host or provides an indication to the host if an interrupt occurs or an alarm goes off. See the TAAD08JU2 setup file utility software for more details on using this utility. 9.4 TAAD08JU2 Application Code The TAAD08JU2 application code performs the actual initialization, configuration, control, and monitoring of TAAD08JU2 through device manager API function calls (configuration that is done by the NSFU can also be accomplished via the application code). This software is written by users to implement TAAD08JU2 in their system. The device manager is not an autonomous program. It is only an interpreter between the user's application code and TAAD08JU2. It is up to the application code to instruct device manager to initialize TAAD08JU2, download the firmware, download the setup file, add connections, monitor alarms and statistics, etc. Therefore, the programmer of TAAD08JU2 should be very familiar with all of the API functions and data types listed in the TAAD08JU2 device manager API document. The following application pseudocode demonstrates control of TAAD08JU2 using device manager API calls: /* Reset the TAAD08JU2 chip - board specific */ lapiInitialize(..); /* Initialize DevMan */ /* User must allocate all memory used by DevMan *. npGetNewportDeviceMemoryRequirement(..); pDev1Handle = malloc(MemoryRequired); lapiInitializeDevice(pDev1Handle, ..); npLoadFirmware(..); /* download the 2 ARM firmware binary files */ lapiSetupDevice(..); /* Download NSFU file */ /* Now add connections, monitor stats, etc. etc. */ npAddConnection(..); ... ... ... /* When all done, shutdown TAAD08JU2 and DevMan */ lapiFinalizeDevice(..); /* Users must free memory they allocated for Device Manager */ free(pDev1Handle); lapiFinalize(); Complete example applications illustrating the proper use of device manager API functions are provided in the TAAD08JU2 software release documentation. Agere Systems Inc. Agere Systems - Proprietary 31 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 9 Software Components (continued) 9.5 System Software The system software is the customer's software controlling the entire system in which a TAAD08JU2 device operates. This software is responsible for configuring peripheral devices around TAAD08JU2 and determining the operating modes of TAAD08JU2. Working examples of system software, running on the Agere TAAD08JU2 evaluation board set, are provided in the Agere software release. 9.6 Software Development Environment Figure 5 is an example of a TAAD08JU2 software development environment. The TAAD08JU2 SOC is totally controlled by software. TAAD08JU2's internal functional blocks are controlled by software (firmware) running on the embedded ARM processor. The firmware must be downloaded by host software into TAAD08JU2 in order for the ARM to boot. Even then, there are no specific directives in the firmware to configure TAAD08JU2 and have it begin executing autonomously. TAAD08JU2 must be initialized, configured, and controlled by an external microprocessor; hence, some form of a software development effort is required in order to use the TAAD08JU2 device. APPLICATION CODE COMPILER DEBUGGER RTOS TARGET BOARD APPLICATION & DEVMAN CPU NSFU TAAD08JU2 SOFTWARE (DEVMAN LIBRARY & FIRMWARE DEVELOPMENT PC SETUP FILE TAAD08JU2 FIRMWARE Figure 5. TAAD08JU2 Software Development Environment Figure 5 details a typical TAAD08JU2 software development environment. The hardware components in the example are the following: ! TAAD08JU2: the TAAD08JU2 device. ! CPU: the external host CPU to which TAAD08JU2 is connected. TAAD08JU2 is usually a memory mapped device on the processor's bus. ! Target Board: a stand-alone, embedded, single-board computer. The board usually has a real-time operating system (RTOS) running on it to provide resource management (processor start-up, serial communications, networking, etc.). The target board has RAM to hold the TAAD08JU2 DevMan and application code and possibly nonvolatile storage for the firmware and setup file(s). ! Development PC: a Windows-based PC used for developing the TAAD08JU2 software that will run on the target board. The PC is connected to the target board by some means (serial, network, or other). 32 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 9 Software Components (continued) The software components are the following: ! Application and DevMan: the executable application and TAAD08JU2 device manager code, linked together to run in the target board's environment and access the TAAD08JU2 EDC registers and interrupt. ! Firmware: the ARM binary files that must be downloaded into TAAD08JU2 by the application using DevMan calls. ! Setup file: the binary configuration file created by the NSFU, loaded into TAAD08JU2 by the application using DevMan calls. ! Compiler/debugger/RTOS: a software development tool kit to compile C code into executable form that will run on the target board CPU. A real-time operating system would probably be included with these tools to manage the target board. A source-level debugger would also assist in debugging application code running on the target board. ! TAAD08JU2 software: the device manager source code ported to the particular compiler and RTOS being used on the target board. The two binary firmware files that need to be downloaded into TAAD08JU2 are also provided in the software package. ! NSFU: the TAAD08JU2 setup file utility that runs on a Windows PC and generates the setup file. ! Application code: the user's code. This code is responsible for initializing DevMan, making the firmware and setup files available to DevMan, and calling DevMan functions to configure and control TAAD08JU2. It is usually linked with DevMan and downloaded to the target board where it runs. 9.7 Notes The following notes apply to initializing, configuring, and operating TAAD08JU2 Version 3.1 only: 1. A minimum number of static queues must be allocated in each of the two enqueue blocks of the SAR (ISIA and ECA). Assuming a default configuration, the minimum number is three. 2. When AAL5/SSSAR/SSTED packet data is reassembled to the host, the user must allocate an IL2Q for each direction in which data is sent to the host. Both of these queues are in the 2—7 range; therefore, they are both ingress queues. Then the user must allocate level 1 queues for egress and ingress to the host. The egress to the host queue must have a source direction of egress. The ID of the egress to host queue is in the 2—7 range. The IL2Q for egress to host traffic is still associated with portID 31. Agere Systems Inc. Agere Systems - Proprietary 33 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 10 Functional Overview Depending upon the provisioned mode of operation, the cell stream associated with a link may be treated as an ATM user network interface (UNI) or combined along with several other links associated with an inverse multiplexing over ATM (IMA) group. The IMA block is provided to terminate the IMA protocol. An ATM service access point (SAP) is effectively provided between the TC and the ATM layers via an internal UTOPIA-2 multi-PHY (MPHY) interface and associated control logic. The ATM layer functions such that connection management, QoS scheduling, buffer management, and statistics gathering are provided by an ATM port controller (APC) block, a modified version of the Agere ATLANTA® APC device. An ATM-SAP is also provided between the ATM and AAL layers via an internal UTOPIA interface and associated control logic. The AAL engine adapts service-specific convergence sublayer (SSCS) packets into ATM cells, supporting both AAL2 and AAL5 protocols. The AAL engine includes class-of-service multiplexing to enable a single AAL2 VC to transport connections of different traffic types. Finally, the SSCS packet is exchanged with the destination SSCS entity via the system interface operating in one of two modes. TAAD08JU2’s system interface supports the following: ! Standard UTOPIA level 2 (also known as UT2) cell-based MPHY master port. ! UTOPIA level 2 plus packet-over-SONET (also known as UT2+) MPHY master port. Underlying this system-on-chip implementation is an embedded device controller (EDC) that provides an intelligent higher-level interface for provisioning and monitoring as well as alarm correlation and statistics gathering. This higher-level, command-based interface simplifies integration of TAAD08JU2 into end systems by reducing firmware development efforts. 10.1 Receive Direction Data Flow This section describes the basic operation of TAAD08JU2 as data is received from the T1/E1/J1 line interface (shown on the left side of Figure 2) and is processed by TAAD08JU2. 10.1.1 PHY Layer TAAD08JU2 may receive cells from either low-speed interfaces (T1, E1, or J1) or high-speed interfaces (155 Mbits/s) via the following: ! Eight T1/E1/J1 span line interface ports ! One UTOPIA-2 16- or 8-bit MPHY port that bypasses the framer, TC, and IMA functions. 10.1.2 Low-Speed PHY Links In the case of low-speed interfaces, TAAD08JU2 enables flexible link assignments for either IMA or UNI mode on a per-link basis. This provisioning capability enables users to isolate delay-sensitive traffic from delay-insensitive traffic by steering time-critical traffic flows onto UNI mode links while carrying other data on IMA mode links. TAAD08JU2’s scheduler views each logical link independently, whether the logical link consists of a single physical link (UNI mode) or multiple physical links (IMA mode group). If IMA mode is selected, the link is also assigned to an IMA group. TAAD08JU2’s scheduler may also guarantee bandwidth to real-time-critical traffic while in IMA mode (since ATM layer processing is independent of the TC and PHY layers). TAAD08JU2’s eight physical ports may be configured in any configuration of links and groups with up to four IMA groups. TAAD08JU2 also provides the capability to switch nxDS0 channels to/from the span lines from/to a concentrated highway interface (CHI) time-division multiplexed (TDM) bus for legacy applications. This provides the ability to share bandwidth on a span line between ATM and TDM traffic. 34 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 10 Functional Overview (continued) 10.1.3 High-Speed PHY Links TAAD08JU2 provides the ability to connect a high-speed external PHY/TC framer device directly to the ATM layer functions via the UTOPIA expansion port. This port can operate at up to 50 MHz. For this path, the internal framer engine, TC, and IMA blocks are bypassed. This high-speed PHY expansion port can be used in combination with the low- and medium-speed PHY links. Note: This expansion port can also be used as an ATM service access point for connections to AAL and system devices. The UTOPIA expansion bus operates at up to 50 MHz and 16-bit data bus widths. 10.1.4 TC and IMA Layers As data is received by the framer, it delineates the data into time slots and bytes and then passes the data to the TC and IMA blocks. The TC determines the proper ATM cell boundaries. The cells are then passed to the IMA block. When receiving data from the TC/IMA blocks, the APC determines which logical link data is received based on the MPHY addressing presented by the IMA block. The IMA block provides the ability to group multiple physical low-speed links into a single logical high-speed link, approximately equal in bandwidth to the sum of individual low-speed links (less IMA protocol overhead). One to four IMA groups may be specified for TAAD08JU2, ranging from a minimum of two links per IMA group to a maximum of eight links per IMA group. The effect of an IMA group is to reduce the transmission latency of long packets (corresponding to high-speed bursts) by increasing the apparent bandwidth available to the data flow. IMA enables a network operator to scale transport capacity for higher bandwidth flows in a more granular way without having to buy more expensive excess capacity. That is, multiple T1/E1 links may be added as demand grows, rather than T3. In the receive direction, the IMA block detects link failures and automatically rebalances the offered load (in the transmit direction) across the remaining good links. The IMA block provides an indication of a failed link to the APC block as a type of backpressure to redistribute the offered load over the smaller bandwidth of the remaining good links. Any data (cells) currently being transmitted over the bad link may necessarily be lost and cannot be retransmitted across the other links. The IMA block removes IMA-protocol-specific cells from the cell stream (in the receive direction) and verifies the IMA protocol across each group’s links. Thus, only data stream cells are transferred between the IMA and APC blocks. TAAD08JU2 provides these protocol checking functions autonomously (via the EDC block) without requiring intervention by an external host device. Agere Systems Inc. Agere Systems - Proprietary 35 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 10 Functional Overview (continued) 10.1.5 ATM Layer Receive ATM layer functions such as connection management, header look-up translation, ingress queueing, OAM, and performance-monitoring processing are provided by the APC block, which is derived from the standardproduct Agere APC device. In the APC receive direction, cells are transferred from the IMA via a UTOPIA-2 interface. In this case, the APC block provides a UTOPIA-2 master function to poll cells from the PHYs. The APC block buffers the cells (so as to prevent cell loss due to buffer overrun in the TC block) and determines the egress destination port for the cell. The APC block may switch the cell to any of several destinations, as follows: ! Any of 31 APC egress ports ! The EDC block to be processed or sent out the host interface ! Any of 40 ATLANTA switch fabric ports connected to the switch fabric interface Within TAAD08JU2, the APC block provides a novel architecture in which ingress and egress internal data buffer space is shared across PHY links and the AAL engine. This enables TAAD08JU2 to route cells from one span line to another span line based on the ATM cell header for an add/drop multiplexer-like ATM-based switch function. The destination port is determined via a header look-up into internal context memory. 10.1.6 AAL Engine The AAL engine receives cells from the ATM layer (APC block) via the UTOPIA-2 bus. The APC is the bus master. Data is transferred from the TC block, the IMA block, or an external source via the UTOPIA expansion interface, into the APC via the UTOPIA-2 bus. Cells are then transferred to the AAL engine. Following AAL processing, cells or packets are forwarded to their destination via the system interface. Alternatively, cells could be received via the switch fabric interface. In this mode, the data enters the APC block via the switch fabric port and then exits via the UTOPIA-2 bus to the AAL engine. From there it is processed and sent out the system interface via the Rx egress system bus as packets, just as in the previously illustrated flow. The AAL engine provides the following types of services, based upon the SSCS entity pertaining to the connection: ! AAL5 reassembly ! AAL2/I.366.1 frame reassembly from AAL2 CPS packets ! AAL2 demultiplexing (in the case of short CPS packets) In addition, if the system interface is the cell-based UTOPIA-2 MPHY, the AAL engine may demultiplex CPS packets from an AAL2 VC into AAL0 cells so that the AAL2 connections can be routed to different destinations within the system. The packets or cells are forwarded to the system interface port operating in one of two modes: ! UTOPIA-2 (cell transfer) ! UTOPIA-2+ (packet transfer) The AAL engine provides class-of-service packet scheduling onto the system interface port to distribute service to different types of traffic via a weighted round-robin scheduler. The AAL engine provides quality-of-service scheduling onto both system and network interfaces to distribute service to different types of traffic via a hierarchy of schedulers. The AAL engine may also detect that certain packets (CPS-SDUs or reassembled AAL5 packets) are destined for the external host device. In this case, the AAL engine transfers the packet to a buffer for access by the external host via TAAD08JU2's device manager. 36 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 10 Functional Overview (continued) 10.1.7 Embedded Device Controller The embedded device controller (EDC) consists of a microcontroller that manages the general operation of the other blocks and communicates with an external CPU via the host interface. 10.2 Transmit Direction Data Flow The transmit direction refers to data transfer from the system interface/switch fabric interface/host microprocessor, through TAAD08JU2, and out through the network interface. 10.2.1 SSCS/AAL Layer Interaction Data to be exchanged with the PHY link may be received from one of the following four sources: ! System interface ! ATLANTA switch fabric interface ! External host microprocessor ! UTOPIA expansion port Data from the system interface may be formatted as cells or packets. In the case of a cell stream, TAAD08JU2's AAL engine may provide AAL0 CPS-packet or AAL2 VC multiplexing; AAL5 cells are passed directly to the ATM layer. In the case of a packet stream, TAAD08JU2's AAL engine may be programmed to provide the following processing (via the AAL engine): ! AAL5 segmentation ! AAL2/I.366.1 frame SARing into AAL2 CPS packets ! AAL2 multiplexing (in the case of short CPS packets) ! AAL2 class-of-service scheduling (when multiple traffic classes share a common AAL2 VC) Data from the ATLANTA switch fabric may be switched to the AAL engine (for AAL0 to AAL2 multiplexing or AAL2 to AAL0 demultiplexing) and then to a PHY link. Also, data may be transferred from the system interface through the AAL engine and routed by the APC to the switch fabric interface. Data from the external host microprocessor may undergo either of the following: ! AAL5 SARing for an SSCOP service ! AAL2 multiplexing as a CPS packet ! AAL2/I.366.1 SARing 10.2.2 ATM Layer The APC block performs normal queue management and scheduling (on a VC basis) corresponding to destination PHY link speed. The APC block moves cells to the IMA block via the APC_TX_Egress UTOPIA bus where the MPHY ID determines the destination for the cell as either one of four IMA group PHYs or one of eight UNI PHYs. The APC UTOPIA-2 bus block acts as a master, while the IMA UTOPIA-2 bus block acts as a slave. Alternatively, the APC block may schedule delivery to an external PHY/TC via the expansion UTOPIA-2 MPHY port. Again, the APC is the UTOPIA master. Agere Systems Inc. Agere Systems - Proprietary 37 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 10 Functional Overview (continued) 10.2.3 IMA/TC Layer The IMA block receives a cell from the APC block via the internal UTOPIA bus and determines the destination port from the MPHY address. If the MPHY address indicates an IMA group, the IMA block routes the cell stream to the appropriate group state machine, which distributes cells in a round-robin fashion across the subtending links. ICP and filler cells are inserted under control of the IMA block, based on IMA frame synchronization and ATM layer traffic rates. In the case of a UNI link, the cells are routed from the ATM layer (APC macrocell) to the TC directly by bypassing the IMA processing. In the IMA transmit direction, the APC block views IMA groups and UNI links as independent logical paths. Thus, the APC block schedules traffic onto the j-th logical PHY as an n x 1.5 Mbits/s (n x 2 Mbits/s) link; n = 1 in the case of a UNI link. Data is transferred between the IMA/TC block and the APC via a UTOPIA-2 MPHY bus where each IMA group or UNI link constitutes a single-destination PHY. The APC block schedules flows onto a maximum of eight PHYs (corresponding to the eight links) or a minimum of one PHY (corresponding to a single eight-link IMA group). The IMA block to framer block interface is via a TC function. Thus, the IMA block provides the ability for the APC block to issue/receive cells directly to/from the framer when links are not provisioned in IMA mode. In this case, each such link is provisioned in UNI mode. 10.2.4 PHY Layer Cells are mapped onto the associated PHY link based on the frame type (T1, E1, or J1). Direct mapping is used to map cells in an octet-aligned fashion into the frame payload. 38 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 11 Modes of Operation This section provides a high-level description of TAAD08JU2’s device operating modes, as well as its interface operating modes. The first section describes the various modes for each interface. The sections following the interface mode section describe the basic modes of operation for TAAD08JU2. 11.1 Interface Modes To provide flexibility, the line interface port, UTOPIA expansion port, and the system interface port pins are multiplexed interfaces sharing different functions that are programmed during device configuration. UTOPIA EXPANSION INTERFACE LINE INTERFACE SWITCH FABRIC BUS TAAD08JU2 CHI TDM INTERFACE SYSTEM INTERFACE HOST PROCESSOR INTERFACE 5-9963(F) Figure 6. TAAD08JU2 Interfaces 11.1.1 UTOPIA-2 Expansion Port Multiplexing Modes The UTOPIA-2 expansion port may be configured in one of the following two operating modes: ! UTOPIA-2 MPHY master, 16-bit mode, 25 MHz or 50 MHz: — Internal or external PHY mode ! UTOPIA-2 MPHY slave, 16-bit mode, 50 MHz: — SAR-only mode 11.1.2 System Interface Port Multiplexing Modes The system interface port may be configured in one of two operating modes: ! UTOPIA-2 MPHY master, 8-bit or 16-bit mode, 25 MHz or 50 MHz ! UT2+ MPHY master, 8-bit or 16-bit mode, 25 MHz or 50 MHz Agere Systems Inc. Agere Systems - Proprietary 39 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 11 Modes of Operation (continued) 11.1.3 Line-Interface Modes The following two line-interface modes of TAAD08JU2 can each be configured in three different areas: ! Line-formatting modes: — T1, E1, or J1 — T1 and E1 can be mixed — IMA groups must be composed of links with the same protocol ! Transmit-clocking modes: — Loop timing — Common transmit clock (CTC) — Independent transmit clock (ITC) 11.2 Device Operating Modes By configuring TAAD08JU2 via pin settings and commands, TAAD08JU2 can be enabled to operate in several modes, as follows: ! Internal PHY mode ! External PHY mode ! SAR-only mode These device operating modes are briefly illustrated on the following pages. 11.2.1 Operating Mode 1: Internal PHY Mode LINE INTERFACE TAAD08JU2 APC LIU CHI FRAMER DEVICE FRAMER B TC AAL ATM SWITCH FABRIC A SYSTEM INTERFACE CHI TDM INTERFACE IMA SWITCH FABRIC BUS UTOPIA EXPANSION INTERFACE EDC SYSTEM HOST PROCESSOR INTERFACE HOST CPU/RAM/ ROM, ETC. 0146(F) Figure 7. Mode 1: Internal PHY Mode Operation The internal PHY mode provides termination of low- or medium-speed span lines carrying ATM cell traffic. Each PHY link may be individually configured as either UNI or IMA mode. In the case of IMA operation, TAAD08JU2 supports up to four IMA groups. A group may comprise two to eight IMA links. 40 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 11 Modes of Operation (continued) TAAD08JU2 implements all IMA group and link state machine behavior within the device with no real-time intervention required by an external host processor. This includes differential delay measurement and mitigation, link synchronization, IMA frame synchronization, and insertion/checking of IMA control protocol (ICP) and filler cells. The IMA virtual link and UNI link cell streams are provided to the ATM layer function within TAAD08JU2 for connection management, buffer management, OAM and PM cell processing, and QoS scheduling. The ATM layer function provides the capability to switch cells from one span line to a different span line for ATM-based grooming. The ATM layer function may also forward cells to the AAL engine function for further processing. If the UTOPIA-2 cell-based system interface is used, the AAL engine may pass AAL5 cells directly to the system interface, while demultiplexing individual packets from AAL2 VCs into individual AAL0 cells. In this mode, TAAD08JU2 provides the ability to route cells onto a cell-based backplane to provide an access function for an ATM switch. The AAL0 and AAL5 cells may be routed through the switch fabric to an egress port, where converging AAL0 cells are remultiplexed into AAL2 VCs. If the UT2+ packet-based system interface is used, the AAL engine provides AAL5 reassembly and AAL2 SARing before forwarding the packet onto the system interface. AAL2 VCs are decomposed into individual CPS packets, prepended with a simple layer 2 packet header, and transmitted over the packet system interface. TAAD08JU2 provides the ability to reassemble frames from multiple AAL2 SDUs in support of ITU I.366.1 SSSAR functionality. TAAD08JU2 also provides the ability to schedule packet transmission onto the system interface via traffic class scheduling to share service across different traffic types. An additional function in the internal framer mode capabilities provides the capability to map ATM traffic into fractional (nx64) logical channels and TDM traffic into the remaining (mx64) logical channels for simultaneous transport of ATM and TDM traffic. This is an enhancement to the internal framer mode in which the CHI interface is enabled. One example of an application where this feature is useful is colocated 2G and 3G base transmission sites (BTS). In this case, the 2G traffic is backhauled on a separate logical channel within a shared physical channel (span line). TAAD08JU2 enables multiplexing (on an nxDS0 basis) of TDM traffic onto a span line along with ATM traffic via logical pipes. ANALOG BTS CHI TDM INTERFACE 3G BTS SPAN LINE LIU BSC TAAD08JU2 ATM CELL MAPPING, 1-n OCTETS TDM CELL MAPPING, n-24/31 OCTETS T1 OR E1 FRAME 5-9966(F) Figure 8. Example of Sharing Span Line with TDM and ATM Data Agere Systems Inc. Agere Systems - Proprietary 41 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 11 Modes of Operation (continued) In Figure 8, the DS0 time slots designated as TDM are multiplexed to/from the CHI bus interface, through the framer block, to/from the T1 or E1 span line. The size of the TDM logical pipe may vary from 0 to N, where N is the number of time slots in the PHY frame structure. The CHI bus is a 32-DS0 frame structure. Regardless of the size of the ATM channel, only 64 kbits/s clear channel span lines are supported. Also, only a single ATM logical channel is supported per span line port; i.e., the logical channel must be composed of consecutive time slots and cannot be split among time slots within a frame. Due to the capacity of the CHI buses, this mode is only available for T1/E1/J1 line interface modes. Other optional traffic routing in this mode includes the use of the switch fabric interface via the APC block instead of the system interface and use of the UTOPIA expansion port in conjunction with the network interface. MPHY addresses not used for the internal connections can be configured for use by the UTOPIA expansion interface. 11.2.2 Operating Mode 2: External PHY Mode OC3/ STS-3 LINK 155 Mbits/s ATM PHY INTERFACE LINE INTERFACE TAAD08JU2 B TC APC IMA AAL FRAMER ATM SWITCH FABRIC A SYSTEM INTERFACE CHI TDM INTERFACE NOT USED SWITCH FABRIC BUS UTOPIA EXPANSION INTERFACE NOT USED EDC SYSTEM HOST PROCESSOR INTERFACE HOST CPU/RAM/ ROM, ETC. 0147(F) Figure 9. Mode 2: External PHY Mode TAAD08JU2 also supports bypass of the low-speed framer/TC and IMA blocks for direct access to the ATM layer and AAL engine functions via a high-speed UTOPIA-2 MPHY interface. In this mode, an external high-speed PHY/TC function may be used to provide access to an OC-3 ring. The ATM layer can filter and process only drop cells. A typical application may be any RAC or BTS that requires access to a high-speed ring, yet terminates only a portion of the total traffic at any particular node. In this mode, TAAD08JU2 may provide SARing or cell switching onto the corresponding system interface. TAAD08JU2 supports up to 155 Mbits/s of data traffic on this interface when operating with a 50 MHz global clock. Even though the switch fabric would not typically be used in this mode, TAAD08JU2 allows full use of this interface. 42 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 11 Modes of Operation (continued) 11.2.3 Operating Mode 3: SAR-Only Mode UTOPIA MASTER DEVICE TC B NOT USED APC IMA AAL FRAMER SWITCH FABRIC BUS LINE INTERFACE UTOPIA EXPANSION INTERFACE CHI TDM INTERFACE TAAD08JU2 SSCS FUNCTION SYSTEM INTERFACE A NOT USED EDC HOST PROCESSOR INTERFACE HOST CPU/RAM/ ROM, ETC. 5-9968(F) Figure 10. Mode 3: SAR-Only Mode The SAR-only mode refers to the capability in which TAAD08JU2 is used as a SAR device only. In this mode, the AAL2 cells entering the UTOPIA expansion interface, which is operating in slave mode, are sent to the AAL engine, bypassing the APC. The AAL engine would provide SARing functions for the supported AAL processing. In this mode, all blocks except SAR and EDC are disabled. When instructed to do so, the AAL engine processes data received on AAL2 connections at the CPS layer. For VCs received from the expansion port, the CPS packets may be queued for either the system interface or looped back to the expansion port. This capability enables the SAR to implement a CID switching capability where data on a given terminated VC/CID may be demultiplexed from the source VC and remultiplexed into a new destination AAL2 VC with a configurable CID. 11.2.4 Operating Mode Summary TAAD08JU2's flexible architecture readily accommodates many ATM access termination challenges while providing open, telecom application-friendly interfaces. The basic modes covered in this document do not represent the only possible operating configurations. The ability to route traffic is not limited to these configurations, and many other combinations are possible by using loopbacks in various blocks and the switching capabilities of the APC. Complementing its APC-based ATM layer architecture with powerful PHY/TC and AAL layer functions, TAAD08JU2 offers a powerful solution to address the convergence of voice and data traffic through a system-onchip approach that minimizes time to market. Agere Systems Inc. Agere Systems - Proprietary 43 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 12 Applications TAAD08JU2 can be used in several applications where low-speed ATM access is required. Target applications include the following: ! BTS network interface termination ! Voice traffic over ATM (VToA) trunking application ! Low-speed ATM access ! AAL2 crossconnect The following subsections describe TAAD08JU2 in each of these applications. 12.1 BTS Network Interface Termination In the BTS network interface termination, TAAD08JU2 provides several useful features. First, third-generation systems require ATM transport for a variety of traffic types, including compressed speech, video, data, and signaling information. TAAD08JU2 supports this through features such as AAL2 and AAL5 transport, AAL2 multiplexing/demultiplexing, and QoS scheduling. AAL2 is especially useful for these wireless systems because most speech packets are short packets, and AAL2 enables multiplexing of several such packets into a single ATM cell, realizing efficient transmission to the BSC. For example, assuming a CDMA-based EVRC speech compression algorithm, from 160 to 200 user connections may be supported on a single T1/E1 span line. Moreover, each of the user connections pertain to a unique channel element (within the BTS) and must be demultiplexed from the ATM VC into individual packets. TAAD08JU2 provides this function by mapping AAL-CIDs into AAL0 cells and forwarding the cells to the destination modem channel card. Third-generation wireless systems also anticipate higher user access speeds, ranging up to 2 Mbits/s. For a 20 ms air frame, this corresponds to ~5000 octets per user. Typically, such data will be transported as ATM/AAL5 flows with different QoS requirements, depending on the source data. BTS IWF N=8 SPANS BSC LOW-SPEED NIC MODEM CARD TAAD08JU2 SAR CE CE CE CE CE CE RF BTS 5-9971(F) Figure 11. BTS Application TAAD08JU2 supports this data traffic by passing through AAL5 cells (which will be reassembled at their destination) with corresponding egress QoS scheduling utilizing the APC scheduling services. Moreover, such traffic tends to be bursty, yet has maximum latency requirements to ensure delivery in time for soft-hand-off diversity combining. TAAD08JU2 supports this need by the use of the IMA block, which reduces the transport latency of long packets while allowing reduced operating costs by not requiring migration to higher-speed transport links (since such traffic is bursty and such pipes may tend to be underutilized). 44 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 12 Applications (continued) Finally, the wireless application necessitates economic and efficient interconnection of microcell or picocell sites so that overall transport costs are minimized. Typically, this may be ensured either by daisy-chaining BTSs or by concentrating traffic in a gateway application. TAAD08JU2 supports this need through the internal prioritizing of passthrough traffic and switching (in the daisy-chain application) through an ATM drop-and-insert feature utilizing the APC’s switching capabilities. TAAD08JU2 further supports such topologies with an egress system interface UTOPIA-2 MPHY bus that enables routing traffic to multiple destinations (in the gateway application). PSTN MSC BSC 5-9972(F) Figure 12. BTSs Require ADM Functions IN-BUILDING PICOCELL CONFIGURATION GATEWAY CONTROLLER RP RP 8-T1/E1 SPAN LINES BSC TAAD08JU2S RP HDLS UTOPIA RP 5-9973(F) Figure 13. Gateway Controller Agere Systems Inc. Agere Systems - Proprietary 45 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 12 Applications (continued) 12.2 VToA Trunking Application Typically, ATM is used to transport CBR services such as voice services via AAL1. However, for low-speed access and PCM traffic, AAL1 is inefficient due to cell header and AAL overhead. ATM/AAL1 may map either multiple samples (n x 64 kbits/s) pertaining to a single connection (with up to 5.875 ms worth of speech) or multiple samples pertaining to multiple connections in a single ATM cell. However, efficient VToA is provided by efficiently transporting compressed speech. In this application, speech is first compressed, possibly utilizing silence-interval suppression. Then the speech packet is multiplexed into an ATM VC utilizing AAL2. TAAD08JU2 supports this application in its normal AAL2 cross connect mode, in which cells are forwarded to TAAD08JU2 from DSPs and compressed into AAL2 by TAAD08JU2. This relieves the DSP from having to provide transport protocol processing. Note: An interworking function may be provided between the DSP and TAAD08JU2 to map the speech packets into ATM. For example, a SAR could be used. However, concentration of the voice packets into a VC should be performed as close to the egress port as possible for maximum statistical multiplexing across many sources. This is why TAAD08JU2 provides an advantage over localized stat-MUXing on a single DSP card. Also, most SARs do not provide AAL2 support. RAC T1/E1 64 Kbits/s PCM PSTN CO VARIABLE LENGTH PACKETS OR AAL0 CELLS FRAMER DSP FRAMER DSP IWF N = 8 T1/E1 ATM/AAL2 COMPRESSED VOICE MUX TAAD08JU2 UTOPIA FRAMER DSP 5-9974(F) Figure 14. Remote Access Concentrator Application EDGE SWITCH LOW-SPEED NIC CPE TAAD08JU2 N=B SPANS ATM SWITCH C-ATM HIGHSPEED NIC C-ATM C-ATM 5-9975(F) Figure 15. Edge/Access Switch Application 46 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 12 Applications (continued) 12.3 Low-Speed ATM Access In this application, data from a variety of sources are multiplexed into AAL2 using I.366.1 QoS MUXing. I.366.1 pertains only to the AAL type 2 and defines SSCS, CPCS, and SAR services necessary for mapping larger packets (e.g., >64 octets) to/from AAL2 VCs. TAAD08JU2 provides support for I.366.1 concentration to enable ATM transport of a variety of low-speed services. 12.4 AAL2 Cross Connect CID switching is provided by the following architecture. One (west) TAAD08JU2 device terminates an AAL2 VC, demultiplexes CIDs from the CID into AAL0 cells, and routes the AAL0 cells through a space switch fabric to a second (east) TAAD08JU2 device. The east TAAD08JU2 device accepts the AAL0 cells, relates them to specific AAL2 VCs, and maps the SSCS into CIDs and subsequently into an AAL2 VC and onto the physical transmission link. In this fashion, TAAD08JU2 may function as a scalable AAL2 cross connect by expansion in ports and (cell) switching fabric. AAL2 VC_1 CIDs: A, C, B TAAD08JU2 (WEST) TAAD08JU2 (EAST) AAL2 VC_3 CIDs: A, E, B TAAD08JU2 (EAST) AAL2 VC_4 CIDs: F, G, B (SWITCH FABRIC/ ASE) AAL2 VC_2 CIDs: E, F, G TAAD08JU2 (WEST) AAL0 VC_E AAL0 VC_F AAL0 VC_G Note: East CIDs need not match west CIDs. 5-9976(F) Figure 16. AAL2 Cross Connect In a fashion similar to the APC stand-alone ATM switching configuration, TAAD08JU2 may provide a stand-alone AAL2 cross connect, as illustrated in Figure 17. AAL2 V C_1 CIDs: A, C, B TAAD08JU2 AAL2 V C_2 CIDs: E, F, G 5-9977(F) Figure 17. Stand-Alone AAL2 Cross Connect Agere Systems Inc. Agere Systems - Proprietary 47 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 13 Embedded Device Controller (EDC) (continued) 13 Embedded Device Controller (EDC) 13.1 Introduction The embedded device controller is based on an ARM9 RISC architecture. The EDC is responsible for the control functions of TAAD08JU2 and is responsible for configuration, on-chip resource management, and compilation of statistical information for performance monitoring. The EDC also provides alarm correlation among the blocks for faster fault detection and isolation. The EDC enables TAAD08JU2 to be controlled via high-level, simple, device-specific commands issued from the external host device. The commands and associated parameters are converted into a series of register-level transactions for OAMP of blocks within the device. The EDC also decouples host signaling insertion/extraction from the dataflow to enable scheduling of service to the external device. Through the use of embedded application code running on the EDC and APIs running on the external host device, the TAAD08JU2 user achieves faster time to market with a powerful software architecture to control the highly integrated system-on-a-chip device. 13.2 Features ! Enables an abstract command/indication interface to TAAD08JU2. ! Performs alarm correlation and fault isolation without requiring external host intervention to minimize data loss. ! Contains all necessary application code to provide its functions. ! Allows data transfers between the host interface and the APC and SAR blocks to support implementation of OAM functions on the host processor. 13.3 EDC Functional Description The EDC contains an ARMT core and a number of peripherals: ! The arbiter and decoder. ! Interfaces to embedded memories including two 32K x 32 SRAMs. ! A programmable interrupt controller (PIC). ! A timer block that provides a watchdog timer and seven additional general-purpose timers. ! The host controller that can transfer between the ARM bus and its own data buffers. 13.4 Host Interface The host interface controller provides the interface between TAAD08JU2 and a host processor. This interface implements the host signal interface and registers through which the host communicates with TAAD08JU2. The interface also provides the synchronization between the host clock (HCLK) and the chips global clock (GCLK). The host registers are located in the HCLK domain to allow for very fast access reads. The input buffer is also written using this clock. Any other signals that must pass to TAAD08JU2’s global clock are appropriately synchronized. The synchronization scheme allows the global chip clock to be completely independent of HCLK. 48 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 13 Embedded Device Controller (EDC) (continued) 13.5 Host Interface Signals and Timing The host interface mimics the timing of a sync-burst SRAM. All inputs and outputs are timed relative to the rising edge of the host interface clock (HCLK). The data is delayed (by two clock cycles) relative to the corresponding address so that reads and writes have exactly the same timing. This allows the user to easily interleave reads and writes to TAAD08JU2. Another feature of the host interface is the HADV input, which allows the user to autoincrement the previously provided address rather than specifying a new address. This may be useful for users who wish to use the input and output buffers in their linearly addressed mode. HCLK HCEN HADV HA WA1 RA1 RA2 WA2 RA3 HWEN HD(31:0) W1 R1 R2 W2A W2B W2C W2D R3A R3 1417 (F) Figure 18. Standard Host Interface Timing 13.6 Host Interactions The host interface controller has several registers and two data buffers that are used for all external host communication. Addresses are shown per 32-bit word. The host registers are shown in Table 14. Agere Systems Inc. Agere Systems - Proprietary 49 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 13 Embedded Device Controller (EDC) (continued) Table 14. Host Registers Register Name CONF (configuration) HA(9:0) Bits 0x000 0 1 2 3 IS (interrupt status) 0x001 4 31:5 0 1 3:2 4 5 QIS (qualified interrupt status) 0x002 31:6 31:0 IE (interrupt enable) 0x003 31:0 IES (interrupt enable set) (write only) IEC (interrupt enable clear) (write only) 0x004 31:0 0x005 31:0 CMD (command register) 0x006 31:0 50 Reset Description Value 0 SOFT_RESET. Places entire TAAD08JU2 device (except for the host interface) into a reset state. 0 ARM_ENABLE. Allows the embedded ARM processor to come out of reset. This bit should remain 0 until the ARM’s firmware has been loaded. 0 DCLK_RATE. Sets the rate of the clock to the framer, TC, and IMA blocks (called DCLK): 0 = DCLK is equal to GCLK. 1 = DCLK is 1/2 of GCLK. Normally, this bit is sent to 0. 0 DCLK_DISABLE. 1 disables the clock to the framer, TC, and IMA blocks. 0 Reserved; should be set to 0. 0 Reserved. 0 IBRI (input buffer ready interrupt). Becomes active when the host can write to the input buffer and the command register. 0 OBRI (output buffer ready interrupt). Becomes active when the output buffer and/or the indication register has data for the host. 0 Reserved. 0 ASI0 (application-specific interrupt). Becomes active when the ARM sets the register. Application-specific interrupts are reserved for future use by the firmware. 0 ASI1 (application-specific interrupt). Becomes active when the ARM sets the register. 0 Reserved. 0 This register contains the same bits as the interrupt status register, after they have been qualified (ANDed) by the corresponding bits in the interrupt enable register. 0 Each bit in this register is an enable for the corresponding bit in the IS register. The IRQN signal becomes active if the OR of each bit of (IS AND IE) is active. — During a write, a 1 in any data bit will cause the corresponding bit in the interrupt enable register to be set. — During a write, a 1 in any data bit will cause the corresponding bit in the interrupt enable register to be cleared. 0 The host should write commands to this register, which will cause an interrupt to the ARM. Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 13 Embedded Device Controller (EDC) (continued) Table 14. Host Registers (continued) Register Name HA(9:0) Bits AIB (autoincrement input buffer) 0x007 31:0 AWA (ARM processor bus write address) 0x008 31:0 AWP (ARM processor bus write parameters) 0x009 8:0 16:9 IBL (input buffer level) 0x00A 31:17 8:0 IR (indication register) 0x00B 31:9 31:0 AOB (autoincrement output buffer) 0x00C 31:0 ARA (ARM processor bus read address) 0x00D 31:0 Agere Systems Inc. Reset Description Value 0 This address provides an autoincrementing addressing of the input buffer. This can be used in place of the HADV external signal to write multiple 32-bit words to TAAD08JU2’s host input buffer. The first write to this location (after reset and after the input buffer ready interrupt has been sent) will be equivalent to a write-to location 0x100. Subsequent writes will be to subsequent word addresses. 0 This register is used for downloading EDC firmware to the ARM processor. A write to this register initiates a write transfer from the host buffer to the address specified. The AWP register should be programmed before writing to this location. 0 This register is used for downloading EDC firmware to the ARM processor. ARM processor bus transfer size: specifies amount of data (in words) that should be transferred to the ARM processor bus. 0 Input buffer start location. Starting location (in words) of the input buffer from where the transfer should originate. 0 Reserved. 0 Read-only. The host can read this register to determine how much data is present in the input buffer. This is only updated when the host is writing data and after the entire buffer of data has been transferred out of the buffer. 0 Reserved. 0 Read-only. The host can read this register after the ARM has set the output buffer ready interrupt. 0 Read-only. This location provides an autoincrementing addressing of the output buffer. This can be used in place of the HADV external signal to read multiple 32-bit words from TAAD08JU2’s host output buffer. The first read from this location (after reset and after the output buffer ready interrupt has been set) will be equivalent to a read from location 0x200. Subsequent reads will be from subsequent word addresses. 0 This register is a diagnostic register that allows the host to read back the contents of the ARM processor memories. A write to this register initiates an ARM processor bus read transfer from the ARM processor bus address specified to the host output buffer. The AWP register should be programmed before writing to this location. Agere Systems - Proprietary 51 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 13 Embedded Device Controller (EDC) (continued) Table 14. Host Registers (continued) Register Name ARP (ARM processor bus read parameters) HA(9:0) Bits 0x00E 8:0 16:9 0x00F 31:17 8:0 VR (version register) (This entire 32-bit register provides the same value as the device ID accessible through the JTAG.) 0x010 31:9 11:0 BIST_INPUT_BUFFER IB (0:255) (input buffer) 0x011 0x100— 0x1FF 0x200— 0x2FF OBL (output buffer level) OB (0:255) (output buffer) 52 27:12 31:28 31:0 31:0 31:0 Reset Description Value 0 This register is a diagnostic register that allows the host to read back the contents of the ARM processor memories. ARM processor bus transfer size: specifies amount of data (in words) that should be transferred from the ARM processor bus. 0 Output buffer start location. Starting location (in words) of the output buffer to where the transfer should go. 0 Reserved. 0 Read-only. The host can read this register to determine how much data is present in the output buffer. This is only updated after the entire transfer of data into the buffer. 0 Reserved. — ID number + LSB of 1. This field value represents the device version ID. The field value is 0x321 (ID number 0x190) for version 3.1 and 0x03B (ID number 0x01D) for version 2.1. 0x5A94 Part number. X Version number. Varies with each mask set. 0 Reserved; must be set to 0. — Linearly addressed input buffer. — Linearly addressed output buffer. Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 14 Framer Block 14.1 Introduction The framer block implements the physical layer function in the ATM protocol stack. The block interfaces with the T1/E1/J1 lines on the line side and the TC block on the system side. It also has a CHI interface for carrying TDM traffic. The block receives and transmits data traffic available from the ATM or TDM interfaces over the physical line. It also receives and transmits data link information over the line. The framer block is a flexible framer engine that enables integration of a feature-rich, yet power-efficient framer function into TAAD08JU2. 14.2 Features ! Framer features: — T1 framing modes: ESF, D4, SLC®-96, T1 DM DDS, SF (FT-only) — E1 framing modes: G.704 basic and CRC-4 multiframe consistent with G.706 — J1 framing modes: JESF (Japan) — E1 signaling modes: • Transparent • Register and system access for entire TS16 multiframe structure per ITU G.732 • Alarm reporting and performance monitoring per AT&T, ANSI, ITU-T, NTT, TTC, and ETSI standards ! Facility data link features: — HDLC or transparent access for either ESF or DDS+ FDL frame formats — Register/stack access for SLC-96 transmit and receive data — Extended superframe (ESF) — Automatic transmission of the ESF performance report messages (PRM) — Automatic transmission of the ANSI T1.403 ESF performance report messages — Automatic detection and transmission of ANSI T1.403 ESF FDL bit-oriented codes — Register/stack access for all CEPT Sa-bits transmit and receive data ! Framer interface: — Concentration highway interface: • Single clock and frame sync signals • Programmable clock rates at 2.048 MHz, 4.096 MHz, 8.192 MHz, and 16.384 MHz • Programmable data rates at 2.048 Mbits/s, 4.096 Mbits/s, and 8.192 Mbits/s • Programmable clock edges and bit/byte offsets — LIU interface: • Eight T1/E1/J1 channels • Line coding: B8ZS, HDB3, AMI, and CMI (JJ20-11) Agere Systems Inc. Agere Systems - Proprietary 53 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) 14.3 Framer-to-Line Interface Unit Physical Interface The network interface of the framer consists of eight groups of six connections. The six connections for each framer are LTXNDATA, LTXPDATA, and LTXCLK, driven from the transmit framer (receive path), and LRXPDATA, LRXNDATA, and LRXCLK (transmit path), sourced from the external line interface device. The line interface may operate in single-rail or dual-rail mode. The default mode of the line encoder is single-rail. In this mode, the input signals are passed transparently through the line encoder. In single-rail mode, the link’s framer internal bipolar line encoder/decoder is disabled and monitoring of received line format violation is accomplished with the use of the LRXNDATA input. When LRXNDATA = 1 on the rising edge of LRXCLK, the line format violation counter increments by one. The link’s transmit framer transmits data via the LTXPDATA output pin while LTXNDATA is forced to a 0 state. In dual-rail mode, the internal line encoder/decoder and monitoring are enabled. The line code may be selected from the following choices: ! Alternate mark inversion (AMI). ! High-density bipolar of order 3—G.703, A.1 (HDB3). ! Binary 8 zero code suppression—G.703, A.2 (B8ZS). Line format violations due to excessive zeros will be optionally monitored as follows: ! B8ZS—eight consecutive zeros cause a violation. ! HDB3—four consecutive zeros cause a violation. 14.3.1 Line Interface References/Standards 1. ITU-T Recommendation G.703, Physical/Electrical Characteristics of Hierarchical Digital Interfaces, 1991. 2. ANSI T1.403-1995, Network-to-Customer Installation—DS1 Metallic Interface, March 21, 1995. 14.3.2 Clocking Modes This section lists all the Tx line clocking modes for framer. ! Mode A1: either all E1s or all T1s running off a single clock (just ATM traffic). In this mode, the CRXCLK input will be used to drive the Tx line clocks. The CRXCLK input can be 1.544 MHz (T1) or 2.048 MHz (E1). Only ATM traffic is supported in this mode. ! Mode A2: a mix of E1s and T1s—all E1s running off a single clock, all T1s running off another single clock (just ATM traffic). In this mode, a reference 2.048 MHz clock input on CRXCLK is used to feed into a PLL that generates 1.544 MHz and 2.048 MHz. The two clocks will then be selected on a per-link basis to run the line at either E1 or T1 rate. Only ATM traffic is supported in this mode. ! Mode A3: independent timing—eight independent clock inputs will be used to drive the eight lines (just ATM traffic). In this mode, eight independent (mutually asynchronous) clocks will be input to the framer. These clocks will be used to drive the Tx line clocks for each line (LTXCLK). Note: In this mode, LTXCLK is an input. Only ATM traffic is supported in this mode. ! Mode A4: loop timing—the Rx line clock will be used to drive the Tx line clocks (just ATM traffic). In this mode, the Rx line clocks (LRXCLK) get internally looped to the Tx line clock (LTXCLK). Only ATM traffic is supported in this mode. ! Mode B1: either all E1s or all T1s running of a single clock or a mix of E1s and T1s—all E1s running off a single clock, all T1s running off another single clock (ATM plus CHI traffic). In this mode, the receive CHI clock (CRXCLK) is used to feed in to the PLL, which generates a 1.544 MHz clock and a 2.048 MHz clock. These clocks are then selected on a per-link basis to run the line at either E1 or T1 rate. Both ATM and CHI traffic are supported. 54 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) 14.3.3 Frame Formats The eight framers support the following frame formats: ! DS1 superframe D4. ! DS1 superframe J-D4 with Japanese Remote Alarm. ! DS1 superframe DDS. ! DS1 superframe SLC-96. ! DS1 extended superframe (ESF). ! Japanese extended superframe J-ESF (J1 standard with different CRC-6 algorithm). ! Nonaligned DS1 (transparent 193 bits). ! CEPT basic frame {ITU G.706}. ! CEPT CRC-4 multiframe with 100 ms timer (ITU G.706). ! CEPT CRC-4 multiframe with 400 ms timer (automatic CRC-4/non-CRC-4 equipment interworking) (ITU G.706 Annex B). ! Nonalign E1 (transparent 256 bits). ! 2.048 coded mark inversion (CMI) coded interface (TTC Standards JJ-20.11). ! DS1 superframe (FT bits only). 14.3.4 Transmit Framer Functions ! Transmits alarm indication signal (AIS) to the line automatically and on demand. ! Transmits remote alarm indication (RAI) to the line automatically and on demand. Conditions for transmitting RAI include loss of received-frame alignment, CEPT loss of received-time slot 0 multiframe alignment, CEPT CRC-4 timer expiration, CEPT loss of received-time slot 16 signaling multiframe alignment, CEPT received Sa6 = 8, and received Sa6 = C. ! Transmits auxiliary test pattern (AUXP) to the line automatically and on demand. ! Transmits CEPT E bits based received CRC-4 errors. ! Support the CEPT double not-FAS system mode. ! Transmits line loopback on and off codes to the line on demand (T1.403 Section 9.3.1). ! When not in frame alignment, to optionally send AIS or transparently pass data. 14.4 DS1 Transparent Framing Format The transmit framer can be programmed to transparently transmit 193 bits of CHI data to the line. When configured for transparent framing, the transmit framer extracts bit 8 of time slot 1 from the receive CHI data and inserts this bit into the framing bit position of the transmit line data. The other 7 bits of the receive system time slot 1 are ignored by the transmit framer. The receive framer will insert every 193rd bit of the receive line data into bit 8 of time slot 1 of the CHI data. The other bits of time slot 1 are set to 0. Frame integrity is maintained in both the transmit and receive framer sections. Agere Systems Inc. Agere Systems - Proprietary 55 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) TIME SLOT 1 TIME SLOT 2 TIME SLOT 3 TIME SLOT 31 TIME SLOT 32 32 TIME-SLOT CHI FRAME (STUFF TIME SLOT) 0 0 0 0 0 0 0 F BIT F BIT TIME SLOT 1 TIME SLOT 2 TRAMSMIT FRAMER’S 193-bit FRAME DS1 = 125 µs TIME SLOT 24 5-5989(F).ar.1 Figure 19. DS1 Transparent Frame Structure In transparent framing mode 1, the receive framer is forced not to reframe on the receive line data. Other than bipolar violations and unframed AIS monitoring, there is no processing of the receive line data. The receive framer will insert the 193rd bit of the receive line data into bit 8 of time slot 1 of the transmit system data. Bit 8 of time slot 1 of the receive system interface is inserted as the 193rd data bit into the transmit line data. In transparent framing mode 2, the receive framer functions normally on receive line data. All normal monitoring of receive line data is performed, and data is passed to the transmit CHI as programmed. The receive framer inserts the extracted framing bit of the receive line data into bit 8 of time slot 1 of the transmit system data. The remaining bits in time slot 1 are set to 0. Bit 8 of time slot 1 of the receive system interface is inserted in the transmit line framing bit position. 14.5 CEPT 2.048 Basic Frame Structure Transparent Framing Format The transmit framer can be programmed to transparently transmit 256 bits of CHI data to the line. The transmit framer must be programmed to transparent framing mode 1. In transparent mode, the transmit framer transmits all 256 bits of the system payload unmodified to the line. Time slot 1 of the CHI interface, determined by the system frame sync signal, is inserted into the FAS/NOTFAS time slot of the transmit line interface. Frame integrity is maintained in both the transmit and receive framer sections. TIME SLOT 1 TIME SLOT 2 TIME SLOT 3 TIME SLOT 31 TIME SLOT 32 TIME SLOT 1 TIME SLOT 2 TIME SLOT 3 TIME SLOT 31 TIME SLOT 32 32 TIME-SLOT CHI FRAME 32 TIME-SLOT LINE FRAME 5-5988(F) Figure 20. CEPT Transparent Frame Structure 56 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) In transparent framing mode 1, the receive framer is forced not to reframe on the receive line data. Other than bipolar violations and unframed AIS monitoring, there is no processing of the receive line data. The entire receive line payload is transmitted unmodified to the CHI. In transparent framing mode 2, the receive framer functions normally on the receive line data. All normal monitoring of receive line data is performed, and data is transmitted to the CHI as programmed. 14.6 Receive Framer Nonalignment Mode (DS1/E1) In the nonalign framing modes, the receive frame aligner does not frame to the receive line data. Other than bipolar violations, AIS, and AUXP monitoring, there is no processing of the receive line data. The entire receive line frame is given unmodified to the system interface. 14.6.1 Loss of Frame Alignment Criteria There are two criteria for declaring loss of frame: frame bit errors and CRC errors. 14.6.1.1 Frame Bit Errors ! T1: two frame bit errors out of 4 frame bits (FT and FS bits checked). ! T1: two frame bit errors out of 5 frame bits (FT and FS bits checked). ! T1: two frame bit errors out of 6 frame bits (FT and FS bits checked). ! T1: three frame bit errors out of 12 frame bits—DDS only (FT, FS, and time slot 24 F bits). ! T1: two frame bit errors out of 4 frame bits (only FT bits checked). ! T1: two frame bit errors out of 5 frame bits (only FT bits checked). ! T1: two frame bit errors out of 6 frame bits (only FT bits checked). ! T1: four frame bit errors out of 12 frame bits—DDS only (FT, FS, and time slot 24 FAS pattern). ! E1: three consecutive incorrect frame alignment signals. ! E1: three consecutive incorrect frame alignment signals or three consecutive incorrect non-FAS frames as indicated by bit 2 in time slot 0 in frames not containing the frame alignment signal. ! E1: three consecutive incorrect FAS or non-FAS frames. ! 2.048 Mbits/s CMI: two consecutive missing code rule violations (CRVs). 14.6.1.2 CRC Errors ! The use of CRC errors to declare loss of frame is optional. CRC errors are monitored in the performance monitor block. ! In DS1 mode, ESF and J-ESF formats only, N or more CRC-6 errors in a 1 second interval results in loss-offrame alignment. N is provisionable. N defaults to 320 in DS1 mode. ! In CEPT mode, N or more CRC-4 errors in a 1 second interval results in loss-of-frame alignment. N is provisionable. N defaults to 915 in CEPT modes. 14.7 Frame Alignment Criteria Table 15 describes the frame alignment criteria for the formats supported by the framer. Agere Systems Inc. Agere Systems - Proprietary 57 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) Table 15. Frame Alignment Criteria Frame Format Alignment Procedure SF Frame alignment is established when six consecutive error-free superframes are received. Only the FT framing bits are checked (36 bits checked). D4 and J-D4 Frame alignment is established when six consecutive error-free superframes are received (72 bits checked in D4, 66 bits checked in J-D4). DDS Frame alignment is established when six consecutive error-free frames are received (42 bits checked: FT, FS, and time slot 24). SLC-96 The FT frame position is established when four consecutive error-free superframes are received (24 FT bits checked). After establishing the FT frame position, SLC-96 superframe alignment is established on the first valid FS sequence of 000111000111. All the while, the FT frame position must remain error-free. ESF and J-ESF Frame alignment is established when three consecutive error-free superframes are received (18 bits checked). CEPT Basic Frame Uses the strategy outlined in G.706 paragraph 4.1.2. CEPT CRC-4 100 ms Timer Uses the strategy outlined in G.706 paragraphs 4.1.2 and 4.2. CEPT CRC-4 400 ms Timer Uses the strategy outlined in G.706 paragraph 4.1.2 and Annex B. 2.048 Mbits/s CMI Coded Interface Frame alignment is established on the first detection of the CRV violation. Multiframe alignment is achieved the first time the 01111111 multiframe alignment pattern is detected. 14.8 Performance-Monitoring Functional Integration Into Framer The framer monitors the recovered line data for alarm conditions and errored events. To a lesser degree of importance, the framer also monitors the receive system data when in the switching mode and presents the information to the system through the embedded device controller. In the transport mode, both directions are monitored for alarm conditions and error events. Table 16 shows the functions provided by the performance monitor and establishes the functions’ validity in particular framing modes. 58 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) Table 16. Performance Monitor Functional Descriptions Function Description 1 Performance report messages (PRMs) as per G.704 Section 2.1.3.1.3.3, G.963, T1.231 Section 6.3, and T1.403 Section 9.4.2. Provides status for errored seconds, bursty errored seconds, severely errored seconds, and unavailable seconds at ET, ET-RE, NT, and NT-RE. Maintains a count of errored seconds, bursty errored seconds, severely errored seconds, and unavailable seconds at the ET. Provides a status indication for a loss of signaling frame alignment condition. Provides a status indication for an out-of-frame condition. Provides a status indication for a loss of time slot 0 CRC-4 multiframe alignment. Provides a status indication for a time slot 0 CRC-4 multiframe alignment signal bit error. Provides a status indication for auxiliary pattern detection. Provides a status indication for detection of the DS1 idle signal. Provides a status indication for detection of an alarm indication signal. Provides a status indication for detection of remote alarm indication. Provides a status indication for detection of time slot 16 AIS. Provides a status indication for detection of remote multiframe alarm in time slot 16 (RTS16MFA). Provides a status indication for the loss of CEPT biframe alignment (LBFA). Provides a status indication for detection of remote Japanese yellow alarm (RJYA). Provides a status indication for continuous E-bit reception. Provides a status indication for detection of Sa6 states. Provides a status indication for detection of line format violations. Provides a status indication for detection of frame bit errors. Provides a status indication for detection of CRC errors. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Valid Framing Modes for Functions ESF and J-ESF only All modes All modes All modes All modes CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only All modes except CEPT CRC-4 All modes All modes CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only J-D4 only CEPT CRC-4 only CEPT CRC-4 only All modes All modes ESF, J-ESF, and CEPT CRC-4 only Provides a status indication for detection of excessive CRC errors. ESF, J-ESF, and CEPT CRC-4 only Provides a status indication for detection of an E bit equal to 0. CEPT CRC-4 only Provides a status indication for expiration of CRC-4 multiframe CEPT CRC-4 only alignment timer. Provides a status indication for new frame alignment. All modes Provides a status indication for detection of Sa7 link identification CEPT CRC-4 only code. Provides a status indication for detection of an SF line loopback on SF only code. Provides a status indication for detection of an SF line loopback off SF only code. Agere Systems Inc. Agere Systems - Proprietary 59 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) Table 16. Performance Monitor Functional Descriptions (continued) Function 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 60 Description Provides a status indication for detection of an overflow in the receive elastic store. Provides a status indication for detection of an underflow in the receive elastic store. Provides a status indication for detection of loss of signal. Maintains a count of received CRC errors. Maintains a count of received bipolar violations, line code violations, and excessive zeros. Provides a status indication for detection of a bit-oriented message in the ESF data link bits. Provides a status indication of a test pattern detector lock. Provides a status indication for detection of a test-pattern bit error. Provides a status indication for detection of an ESF-FDL RAI/yellow alarm code. Provides a status indication for detection of the ESF-FDL payload loopback enable code. Provides a status indication for detection of the ESF-FDL payload loopback disable code. Provides a status indication for detection of the ESF-FDL line loopback enable code. Provides a status indication for detection of the ESF-FDL line loopback disable code. Maintains a 16-bit count of received framing bit errors. Maintains a 16-bit count of received E bit = 0 events. Maintains a 16-bit count of received Sa6 = 00x1 events. Maintains a 16-bit count of received Sa6 = 001x events. Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (x, x, AIS). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (0, 1, 1111). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (1, 1, 1111). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (x, x, AUXP). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (1, 1, 1000). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (0, 1, 1000). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (0, 1, 1110). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (0, 1, 1100). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = (1, 0, 0000). Agere Systems - Proprietary Valid Framing Modes for Functions All modes All modes All modes ESF/J-ESF and CEPT CRC-4 only All modes ESF only All modes All modes ESF only ESF only ESF only ESF only ESF only All modes CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only CEPT CRC-4 only Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) Table 16. Performance Monitor Functional Descriptions (continued) Function 54 55 56 Valid Framing Modes for Functions Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = CEPT CRC-4 only (1, 1, 1110). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = CEPT CRC-4 only (1, 1, 00xx). Provides a status indication for detection of an (A, Sa5, Sa6[1:4]) = CEPT CRC-4 only (x, 0, xxxx). Description 14.9 Performance Report Message A performance report message is assembled by the performance monitoring block in the framer. This message can be either sent automatically in the ESF data link or sent as a result of a command. The performance monitor block monitors for errored second events and generates the one-second data for the extended superframe (ESF) performance report message (PRM) (G.704 Section 2.1.3.1.3.3, G.963, T1.231 Section 6.3, and T1.403 Section 9.4.2). The form of the PRM message is shown in Table 17 below. The definition of the fields is given in Table 18. A severely errored frame (SEF) defect is determined by examining contiguous time windows for frame bit errors. In ESF, the window size is 3 ms, and only the frame pattern sequence bits are checked. An SEF defect occurs when two or more frame bit errors in a window are detected. An SEF defect is terminated when the signal is in frame and there are less than two frame bit errors in a window. Table 17. Performance Report Message Format Octet Number 1 2 3 4 5 6 7 8 9 10 11 12 13—14 15 Agere Systems Inc. PRM B7 G3 FE G3 FE G3 FE G3 FE PRM B6 LV SE LV SE LV SE LV SE PRM B5 G4 LB G4 LB G4 LB G4 LB PRM B4 PRM B3 Flag SAPI TEI Control U1 U2 G1 R U1 U2 G1 R U1 U2 G1 R U1 U2 G1 R FCS Flag Agere Systems - Proprietary PRM B2 G5 G2 G5 G2 G5 G2 G5 G2 PRM B1 PRM B0 C/R EA EA SL Nm SL Nm SL Nm SL Nm G6 NI G6 NI G6 NI G6 NI 61 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) . Table 18. Performance Report Message Field Definition Field G1 = 1 G2 = 1 G3 = 1 G4 = 1 G5 = 1 G6 = 1 SE = 1 FE = 1 LV = 1 SL = 1 LB = 1 U1, U2 = 0 R=0 Nm, NI = 00, 01, 10, 11 Definition CRC Error Event = 1 1 < CRC Error Event ≤ 5 5 < CRC Error Event ≤10 10 < CRC Error Event ≤100 100 < CRC Error Event ≤ 319 CRC Error Event ≥ 320 Severely Errored Framing Event ≥ 1 (FE will = 0) Frame Synchronization Bit Error Event ≥ 1 (SE will = 0) Line Code Violation Event ≥ 1 (BPV ≥ 1 or EXZ ≥ 1) Slip Event ≥ 1 Payload Loopback Activated Reserved Reserved (default value = 0) One Second Report Modulo 4 Counter 14.10 ESF Data Link When the framer is in ESF mode, several options are available for use of the 4 kbits/s data link that is part of the SF framing structure. The npSetLineAutoPrmConfig1 device manager API allows for the enabling and disabling of automatic sending of the PRM and the selection of threshold levels for the receive and transmit ESF data link FIFOs from two choices programmed using the npSetDataLinkThresholds. If the PRM is set to be transmitted automatically, then the data link cannot be used for other types of messages. If the automatic PRM is disabled, the ESF data link channel can be used to send bit-oriented messages (BOMs), PRM, and any other data message the user wishes to send. Messages to be sent are specified using the npSendData in the case of non-BOM or the npSetLineAutoBomConfig for BOM. Non-BOM messages can be up to 128 bytes long. Received messages can be generated by npGetData for the host to extract the data. 14.11 Facility Data Link The facility data link is available in SLC-96, DDS and CEPT framing formats. FDL messages are sent using the npSendData and received using the npGetData. 14.11.1 Facility Data Link References/Standards ANSI T1.403-1995-Bit-Oriented Messages (BOM). 1. Device manager function calls throughout this document are indicated by boldface type. 62 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) 14.11.2 Receive Data Link Functional Description This block extracts facility data links bits, as follows: ! D bits from the SLC-96 multi-superframe. ! Sa bits from time slot 0 in CEPT basic and CRC-4 multiframes. ! Data link bits from DDS frames. The respective bits will always be extracted from the framed-aligned receive line payload and stored in the facility data link stack regardless of other configuration bits. All frame types support clear-on-read status and interrupt bits based on the setting of the input select signal. 14.11.3 SLC-96 Superframe Receive Data Link ! Delineates the SLC-96 data link in the Fs signaling frame, extracts the 24 D bits, and stores them in word 1 of npGetData. ! Provides interrupt for stack ready. ! Provides host access to stack using processor clock. ! Supports loss-of-frame status. Both basic frame alignment and multiframe alignment must be established before the data can be assumed valid. 14.11.4 DDS Receive Data Link Stack ! Extracts data link bit (bit 6) from time slot 24 and stores it in words 1 and 2 of npGetData. ! Provides interrupt via indication processing. ! Provides host access to stack using npGetData. ! Supports loss-of-frame status. DDS frames are numbered 1 through 12 with the data link bits located in bit 6 of time slot 24 in every frame. Only basic frame alignment must be established for the data link bits to be extracted. 14.11.5 CEPT; CEPT CRC-4 (100 ms); CEPT CRC-4 (400 ms) Multiframe Sa Bits Receive Stack ! Extracts two multiframes of Sa bits from CEPT links and stores them in words 1 through 3 of npGetData. ! Supports loss-of-frame status. ! Provides host access to the stack using npGetData. ! Provides interrupt via indication processing. CEPT frames are numbered 0 through 15 with the Sa bits located in time slot 0 of the odd numbered frames. The Sa bits can only be extracted from CEPT links when the proper alignment has been established. For basic CEPT frames, the Sa bits will be extracted given the arbitrary alignment selected by the frame aligner block when basic frame alignment is established. For CEPT CRC-4 links the Sa bits will be extracted based on the alignment determined by the frame aligner block when multiframe frame alignment is established. Optionally, the Sa bits will be extracted from CEPT CRC-4 links only after basic frame alignment is established (RxCRCSM). Agere Systems Inc. Agere Systems - Proprietary 63 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) 14.11.6 Receive Data Link Stack Idle Modes ! No data link stack features for the following frame formats: — D4 — J-D4 — ESF — J-ESF — CMI 14.11.7 Transmit Facility Data Link Functional Description This block performs the transmission of D bits into SLC-96 superframes, Sa bits into CEPT multiframes, and datalink bits into DDS frames using npSendData. 14.11.8 SLC-96 Superframe Transmit Data Link ! Provides for sending D bits and delineator bits on SLC-96 bits via npSendData. ! Provides interrupt and initiates npGetData process using a Tx threshold. ! Performs retransmission of stack when update is yet to be performed. The 12 frame SLC-96 superframe is composed of a terminal frame (FT) alternating with a subframe that consists of a combined signaling (FS) frame and data link. The subframe shares establishing the signaling frame (FS) and SLC-96 data link. The FDL stack bits are inserted into the signaling and data link subframe position in the superframe. Seventy-two frames (six superframes) are required to deliver the 24 D bits and 12-bit delineator. The frontend delineator is 00111, which is followed by 24 D bits and trailed by 0001110. The alignment of the FS bits within the superframe is determined and indicated by the frame aligner block. The transmission of the SLC-96 stack will take 9 ms to complete, during which time the host should refill the stack using the npSendData if the D bits need to change. 14.11.9 DDS Transmit Data Link Stack ! Provides for sending three superframes of data link bits via npSendData. ! Provides interrupt and initiates npGetData process using a Tx threshold. ! Performs retransmission of stack when update has yet to be performed. ! Provides host access to stack using npSendData. The transmission of the SLC-96 stack will take 9 ms to complete, during which time the host should refill the stack using the npSendData if the D bits need to change. 14.11.10 Transmit ESF Data Link Bit-Oriented Messages ! Provides capability to transmit bit-oriented messages. When enabled, bit-oriented messages will be transmitted on the data link channel of the frame bit for ESF links. The ESF superframe is numbered 1 through 24 with the data link channel transmitted in the odd numbered frames (4 kbits/s). The BOM is a 16-bit message defining an alarm or command and response action, and sent repeatedly for a period of time determined by the event. The message consists of eight 1s, a 0, a 6-bit code to identify the alarm or action, and a 0 (1111_1111_0 in front and 0 behind the 6-bit code). 64 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) The message can occur at any point in the extended superframe without respect to boundaries. The BOM format is as follows: 0 X X X _ X X X 0 _ 1111_1111 (right-most bit being transmitted first). When the BOM pattern is enabled, it will be transmitted until disabled. When disabled, the pattern will cease to be transmitted immediately. When enabled, the BOMs should only be inserted when the proper alignment has been reached. For ESF links, both BFA and MFA are required for insertion. This condition affects the insertion of BOMs bits and the reporting of stack empty to the host. 14.11.11 CEPT, CEPT Multiframe Transmit Data Link Sa Bits Stack ! Provides two multiframes of Sa-bit storage for transmission on CEPT links. ! Provides interrupt for stack empty. ! Performs retransmission of stack when update has yet to be performed. ! Provides capability to source Sa bits from blocks other than the data link block. This block will always present the Sa bits stored in the Tx stack to the TDM data stream. In CEPT, the Sa bits are located in time slot 0 of the NOTFAS frames (odd-numbered frames). CEPT multiframe format frames are numbered 0 through 15 with the Sa bits located in time slot 0 of the odd-numbered frames (NOTFAS frames). The Sa bits are stored in the Tx stack as follows. Table 19. Shared Tx Stack Format for CEPT Frame Word 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 SA41 SA43 SA45 SA47 SA49 SA411 SA413 SA415 SA41 SA43 SA45 SA47 SA49 SA411 SA413 SA415 1 SA51 SA53 SA55 SA57 SA59 SA511 SA513 SA515 SA51 SA53 SA55 SA57 SA59 SA511 SA513 SA515 2 SA61 SA63 SA65 SA67 SA69 SA611 SA613 SA615 SA61 SA63 SA65 SA67 SA69 SA611 SA613 SA615 3 SA71 SA73 SA75 SA77 SA79 SA711 SA713 SA715 SA71 SA73 SA75 SA77 SA79 SA711 SA713 SA715 4 SA81 SA83 SA85 SA87 SA89 SA811 SA813 SA815 SA81 SA83 SA85 SA87 SA89 SA811 SA813 SA815 Transmission of the Sa stack will take 4 ms, during which time the host should refill the system stack if the Sa bits need to change. Near the beginning of each CEPT double multiframe, the Tx data link block will determine whether a new set of Sa bits is available to be transmitted. If this is the case, the new Sa bits will be transmitted; otherwise, the previous Sa bits will be retransmitted. When enabled, the Sa bits will only be inserted when the proper alignment has been reached. For CEPT with no CRC-4 links, only biframe alignment (BFA) is required for insertion. For CEPT with CRC-4 links, both biframe alignment (BFA) and CRC-4 multiframe alignment (MFA) are required for insertion for the insertion of Sa bits. Before enabling a link for CEPT format, the host should initialize the stack. If not, the data link block will transmit the reset state of the stack, which is arbitrary. Agere Systems Inc. Agere Systems - Proprietary 65 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) 14.11.12 Transmit Data Link Stack Idle Modes ! D4 ! J-D4 ! CMI ! No data link features 14.11.13 SLC-96, DDS, or CEPT ESF Frame Alignment For CEPT, DDS, or SLC-96 frames, loss-of-frame alignment is not an issue since the framer is the source of timeslot 0 or the F bits. Once a link is enabled, the frame sequence always starts at the beginning. In the case of the system being the source of multiframe alignment, the data link block will simply deliver what is requested. 14.12 Concentration Highway Interface (CHI) The CHI can be programmed to operate at 2.048 MHz, 4.096 MHz, 8.192 MHz, or 16.384 MHz clock rates (data rates up to 8.192 Mbits/s only). A pair of global system clock and system frame sync (one for the transmit and one for the receive direction) are required. The offset between the frame sync and bit 0 of time slot 0 is programmable. This interface can be used, for example, to interface with the Agere TSI devices. 14.12.1 CHI References/Standards 1. ITU G.783 Characteristics of Synchronous Digital Hierarchy (SDH) Equipment Functional Blocks. 2. ITU Q.511 Exchange Interfaces Towards Other Exchanges. 14.12.2 Transmit/Receive CHI Features The features supported on the CHI are summarized below: ! Data rates of 2.048 Mbits/s, 4.096 Mbits/s, 8.192 Mbits/s. ! Clock rates of 2.048 MHz, 4.096 MHz, 8.192 MHz, 16.384 MHz. ! A global input clock and frame sync. ! Byte offset—2.048 Mbits/s, 0—31 bytes. ! Byte offset—4.096 Mbits/s, 0—63 bytes. ! Byte offset—8.192 Mbits/s, 0—127 bytes. ! Bit offset. ! 1/2-bit offset. ! 1/4-bit offset. ! Clock mode select. ! Double time-slot mode, CHIDTS. ! Double NOTFAS system time slot, FRM_DNOTFAS. ! Sampled clock edge for transmit system frame sync. ! Global programmable stuffed time-slot position in DS1 mode. ! Global programmable stuffed byte in DS1 mode. 66 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) ! Global single time-slot loopback address for system or line. ! Programmable automatic system AIS (loss-of-frame alignment). ! Programmable automatic system AIS (CEPT CRC-4 multiframe alignment timer expiration). ! On-demand transmission of system AIS. 14.12.3 Double NOTFAS System Time-Slot Mode In the default case (FRM_DNOTFAS = 0), both the FAS and NOTFAS time slots are transmitted by the transmit system interface and expected by the receive system interface. Setting FRM_DNOTFAS to 1 enables the NOTFAS time slot to be transmitted twice on the transmit system interface in the NOTFAS and FAS time-slot (TS0) positions. Similarly, the receive system interface assumes time slot 0 to carry NOTFAS data that is repeated twice. 14.12.4 Transparent Mode In the transparent DS1 mode, the transmit system interface inserts the 193rd bit of the DS1 frame in bit 7 (LSB) of the first stuffed time slot. The receive system interface takes bit 7 of the first stuffed time slot and inserts it into the framing bit position (193rd bit on the TDM data bus). In the transparent E1 mode, the transmit system maps 32 received time slots into the CHI time slots. Similarly, the receive system maps the CHI time slots into the TDM bus time slots. The transmit frame formatter inserts TS0 of the CHI (FAS/NOTFAS) into the TS0 of the frame based on the biframe alignment. 14.12.5 Loopbacks Two forms of loopbacks are supported: single time-slot system loopback and single time-slot line loopback, as shown in Figure 21. When in single time-slot system loopback, a single time slot from the receive system interface is looped back to the system. An idle code is transmitted to the line in place of the looped-back time slot. When in single time-slot line loopback, a single time slot from the transmit system interface is looped back to the line. The programmable idle code is transmitted to the system in place of the looped-back time slot. SYSTEM LINE ES SINGLE TIME-SLOT SYSTEM LOOPBACK SYSTEM LINE ES SINGLE TIME-SLOT LINE LOOPBACK ES = ELASTIC STORE 5-9030(F)r.1 Figure 21. System Loopbacks Agere Systems Inc. Agere Systems - Proprietary 67 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) 14.12.6 Nominal CHI Timing Figure 22 illustrates nominal CHI frame timing. Double time-slot mode (CHIDTS) is disabled. The frames are 125 µs long and consist of 32 contiguous time slots when the 2.048 MHz data rate mode is selected. In DS1 frame modes, the CHI frame consists of 24 payload time slots and eight stuffed (unused) time slots. In CEPT frame modes, the CHI frame consists of 32 payload time slots. ! CTXDATA: output data to system. ! CRXDATA: input data to system. ! CTXFS: transmit CHI frame sync. ! CRXFS: receive CHI frame sync. 125 µs CTXFS/ CRXFS DS1 FORMAT 2.048 Mbits/s CHI: FRAME 1 24 VALID TIME SLOTS CTXDATA FRAME 2 7 STUFFED SLOTS* 1 STUFFED SLOT 24 VALID TIME SLOTS CRXDATA FRAME 2 CEPT FORMAT 2.048 Mbits/s CHI: 32 VALID TIME SLOTS CTXDATA or CRXDATA FRAME 1 FRAME 2 4.096 Mbits/s CHI: CTXDATA FRAME 1 CRXDATA FRAME 1 HIGH IMPEDANCE FRAME 2 FRAME 2 8.192 Mbits/s CHI: CTXDATA FRAME 1 CRXDATA FRAME 1 HIGH IMPEDANCE FRAME 2 FRAME 2 5-8978(F) * The position of the stuffed time is controlled by register bit FRM_STUFFL. FRM_STUFF = 1 is shown. Figure 22. Nominal Concentration Highway Interface Timing 68 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 14 Framer Block (continued) 14.12.7 CHI Timing with CHI Double Time-Slot Timing (CHIDTS) Mode Enabled Figure 23 illustrates the CHI frame timing when CHIDTS is enabled. In the CHIDTS mode, valid CHI payload time slots are alternated with high-impedance intervals of one time-slot duration. This mode is valid only for 4.096 Mbits/s and 8.192 Mbits/s CHI data transfer rates. 125 µs CTXFS/ CRXFS FRAME 2 FRAME 1 4.096 Mbits/s CHI TIME TIME SLOT SLOT CTXDATA TS0 TS1 TS2 TS3 TS4 TS30 TS31 TS0 TS0 TS1 TS2 TS3 TS4 T30 TS31 TS0 TS0 TS1 TS2 TS30 TS31 TS0 TS1 TS2 TS30 TS31 8 bits CRXDATA 8.192 Mbits/s CHI CTXDATA CRXDATA HIGH IMPEDANCE TS0 TS0 5-8979(F) Figure 23. CHIDTS Mode Concentration Highway Interface Timing 14.12.8 Clocking Scheme LTXCLK—Line Transmit Clock. This can be either input or output. When programmed as output, it is used to generate line clock for each tx line. The clock source for this clock is either the LRXCLK when lines are set in Rx CLK loop mode, or the CRXCLK pin. When programmed as input, the system designer must provide the Tx line clocks to . It is assumed that these clocks will also drive the Tx LIUs. Typically, LTXCLK is set as an input to implement independent transmit clock (ITC) mode. This pin either inputs or outputs 1.544 MHz or 2.048 MHz. CRXCLK—CHI Receive Clock. This pin is an input, but also has multiple functions as listed below: ! It can be set up as a CHI clock in which 2.048 MHz, 4.096 MHz, 8.192 MHz, or 16.384 MHz can be applied. In this case, transmit clock must be generated either by line Rx clock in loop timing mode, or LTXCLK in input mode. ! This pin can also be set to REF2M048 mode. In this mode, a fixed 2.048 MHz is applied and generates both 1.544 MHz and 2.048 MHz clocks that drive the various framer ports based on user-programmed setup. This pin can be set as a common transmit clock (CTC) for all lines. In this mode, this pin could have 1.544 MHz or 2.048 MHz, and this will be the clock reference for all eight lines. The LRXCLK and the CTXCLK each have the one function as defined in the pinout. Agere Systems Inc. Agere Systems - Proprietary 69 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 15 Transmission Convergence (TC) Block 15.1 Introduction The transmission convergence (TC) block provides the cell delineation function and physical layer mapping of ATM cells to/from the PHY link. In the receive direction, the TC receives cells as a byte-aligned payload from the framer. The TC locates the ATM cell boundary via the common I.432-based cell delineation technique. The TC then passes ATM cells to the ATM layer for subsequent processing and routing. In the transmit direction, the TC receives cells from the ATM layer via an internal UTOPIA-2 MPHY interconnection. The TC inserts a correct HEC in the ATM cell header and maps the cell in a byte-aligned fashion into a data stream toward the framer. If the link is configured for ATM UNI mode, the TC performs cell rate decoupling between the physical layer and the ATM layer by discarding idle cells in the receive direction and by inserting idle cells in the transmit direction. The TC also discards ATM cells with uncorrectable HEC errors. If the link is configured for ATM IMA mode, the TC does not discard any cells for rate decoupling. ATM cells with a bad header are discarded and an indication is provided to the IMA sublayer in order to preserve IMA frame synchronization. 15.2 Features ! Supports up to eight PHY channels. ! Provides cell delineation per ITU I.432. ! ATM UNI mode support for idle cell discard. ! ATM IMA mode support by detecting and/or discarding errored cells. ! Performs ATM cell HEC checking in the receive direction. ! Optional support for error detect/correct mode. ! Performs ATM cell HEC generation in the transmit direction. ! Optional support for x55 coset addition to ATM header. ! Optional support for cell payload scrambling. ! Supports TC onto T1, E1, and J1 PHY links. ! Supports fractional ATM logical channels on T1, E1, and J1 links. ! Mapping of ATM cells onto DS1 per direct mapping scheme of G.804. ! Mapping of ATM cells onto E1 per G.704 and G.804. 70 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 15 Transmission Convergence (TC) Block (continued) 15.3 TC—Receive Direction The receive section consists of the following blocks: ! Receive line interface. This block interfaces to the framer. It determines which channel is to be processed next and provides the receive cell delineator with the data word to process. ! Receive cell delineator. This block implements the cell delineation state machine and generates the loss of cell delineation event. ! Receive cell processor. This block implements the header error control state machine, includes a self-synchronizing descrambler, and writes passed (unfiltered) cells into the receive data buffer. ! Transmit IMA interface. This implements the interface to the IMA block. Cell data is written into the transmit data buffer from the IMA block. ! Receive IMA interface. This block implements the interface to the IMA block. Cell data is read from the receive data buffer and output to the IMA block from this interface. Parity is generated for each byte read out of the receive data buffer. ! Receive line side context RAM. This block, indexed by the channel number from the receive line interface block, stores the state variables (generated on the line clock) for the receive cell delineator, receive processor, and receive buffer manager. ! Receive UTOPIA side context RAM. This block, which is indexed by the port number from the receive UTOPIA slave interface block, stores the state variables (generated on the UTOPIA clock) for the buffer manager. ! Receive data buffer. This block stores received ATM cells together with the start-of-cell (SOC) signal. 15.4 TC—Transmit Direction The transmit section consists of the following blocks: ! Transmit line interface. This block interfaces to the framer; it determines which channel is to be processed next and gets the data from the transmit cell processor. ! Transmit cell processor. This block reads cells from the transmit data buffer and performs the payload scrambling and HEC generation. If no cells are available in the buffer, the transmit cell processor generates idle cells. ! Transmit buffer manager. This block manages the read and write pointers for each channel (which are generated on separate clocks), as well as the various FIFO flags. ! Transmit UTOPIA slave interface. This block implements the MPHY transmit UTOPIA-2 interface. Cell data and the SOC bit from the UTOPIA interface are written into the transmit data buffer and output on the UTOPIA interface. Parity is checked for each byte input from the UTOPIA interface. ! Transmit line-side context RAM. This block, which is indexed by the channel number from the transmit line interface block, stores the state variables (generated on the line clock) for the transmit cell processor and buffer manager. ! Transmit UTOPIA side context RAM. This block, which is indexed by the port number from the transmit UTOPIA slave interface block, stores the state variables (generated on the UTOPIA clock) for the buffer manager. ! Transmit data buffer. This block stores transmit ATM cells together with the start-of-cell (SOC) signal. Agere Systems Inc. Agere Systems - Proprietary 71 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 15 Transmission Convergence (TC) Block (continued) 15.4.1 HEC Generation/Checking The HEC octet is generated in the transmit direction by calculating a CRC with the polynomial x8 + x2 + x + 1 over the first four header octets and then adding the coset 01010101. In the receive direction, the HEC is checked by calculating the CRC according to the above method on the received header octets and subsequently adding the received HEC. The syndrome thus generated is equal to 0 if no errors are in the header. The CRC provides the capability of single error correction and multiple error detection. The error control mechanism can be in one of two states: correction or detection. When an error occurs, a cell may be corrected or discarded depending on the state of operation. 15.5 Cell Delineation Cell delineation means finding the cell boundaries in a cell stream with a good degree of confidence. The TC implements HEC-based cell delineation. Initially, the cell delineation state machine is in hunt. When a correct HEC is found, it is assumed that a candidate cell delineation is found and the state machine goes to presync. In presync, if an incorrect HEC occurs, the hunt state is resumed. Otherwise, if delta (normally = 6) consecutive correct HECs occur, it is assumed that the cell delineation is correct and the state machine goes to the sync state. The state machine remains in sync unless alpha (normally = 7) consecutive incorrect HECs are found—in which case the state machine goes back to the hunt state. 15.6 Cell Payload Scrambling/Descrambling In the transmit direction, each cell payload is scrambled by a self-synchronizing scrambler using the polynomial x43 + 1. Cell payloads are scrambled in order to avoid repeating fixed bit patterns and improve the robustness of the cell delineation algorithm. In the receive direction, the cell payload is descrambled by a descrambler using the same polynomial (x43 + 1). The descrambler is only active when the cell delineation state machine is in presync or sync. 15.7 Cell Mapping The direct cell mapping method is used with most line framing formats. In this method, cells are octet aligned (with respect to the line framing and overhead) and consecutive with no gaps between cells. 15.8 Facility Maintenance Two maintenance functions are defined: detection of the out-of-cell delineation (OCD) condition and detection of the loss-of-cell delineation (LCD) condition. OCD is declared when hunt state is entered and it is cleared when sync state is entered. ! LCD is declared when OCD persists continuously for a time greater than that set in a programmable timer. The timer can be programmed for up to 1 s. The default timer setting is 4 ms. LCD is cleared when OCD is cleared continuously for the same programmed time. 72 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 15 Transmission Convergence (TC) Block (continued) 15.9 Cell Rate Decoupling Cell rate decoupling, in the transmit direction, means inserting idle or unassigned cells to keep the cell stream continuous when no other cells are available for transmission. In the receive direction, it means discarding idle, unassigned, or invalid cells. These types of cells are defined in Table 20. Cell rate decoupling is a physical layer function according to the International Telecommunication Union (ITU) I.432 standard and an ATM layer function according to the ATM Forum's User-Network Interface (UNI) Specification Version 3.1*. * When transmitting idle or unassigned cells, the correct HEC must be generated. In addition, I.432 requires that the payload of an idle cell be 48 octets equal to 01101010. Table 20. Cell Headers of Idle, Unassigned, and Invalid Cells Cell Type Cell Header Definition Idle (ITU I.432) 00000000 00000000 00000000 00000001 Unassigned (ATM Forum UNI 3.1) 00000000 00000000 00000000 0000XXX0 Invalid (ATM Forum UNI 3.1) XXXX0000 00000000 00000000 0000XXX1 Note: X means don't care. 15.10 Functionality Table 21 lists the functions and requirements met by the TC block. . Table 21. TC Functionality Function HEC Generation/ Verification Cell Delineation Cell Scrambling/ Descrambling Cell Mapping Maintenance Functions Cell Rate Decoupling Agere Systems Inc. Subfunction Error detection Error correction HEC generator polynomial Coset HEC-based State transition to hunt after loss of cell delineation State transition to sync after presync State transition to presync from hunt Specification Multiple-bit, cell optionally discarded. Single-bit, cell header corrected and accepted if in correction mode. Cell discarded if in detection mode. x8 + x2 + x + 1. 01010101. Programmable on/off. ATMF and ITU state diagram. <7 cell times. <6 cell times. 1 valid HEC. x43 + 1, active only in presync and sync state and on ATM payload bytes. Programmable ON/OFF. — Direct mapping. Out-of-cell delineation (OCD, LOC for ITU) On = Transition out of sync state. OFF = Transition to sync state. Loss-of-cell delineation (LCD) On = OCD persisting for 4 ms. OFF = OCD not present for 4 ms. Idle cell insertion Programmable on/off. Idle cell pattern Programmable header and payload. Self-synchronizing scrambler/descrambler Agere Systems - Proprietary 73 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block 16.1 Introduction IMA provides modular bandwidth, using existing physical links (e.g., DS1/E1), to access ATM networks and to interconnect ATM network elements. Inverse multiplexing groups a number of physical links to form a logical link whose bandwidth is approximately the sum of the bandwidth of the individual links. In the transmit direction, a single stream of ATM cells is inverse multiplexed across physical links in a group, in a round-robin fashion. In the receive direction, ATM cells are inverse demultiplexed from the various links in a group, in a round-robin fashion, in order to reconstruct the original cell stream. An IMA example is illustrated in Figure 24 for a single group of three links. IMA GROUP PHY PHY PHYSICAL LINK PHYSICAL LINK IMA GROUP PHY PHY CELL STREAM ATM LAYER ORIGINAL CELL TO ATM LAYER PHY PHYSICAL LINK PHY IMA VIRTUAL LINK 5-9980(F) Figure 24. IMA Application 74 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) 16.2 Features ! Supports up to four IMA groups. ! Supports from two to eight links per IMA group within TAAD08JU2. ! Provides IMA group and link state machine behavior per ATM Forum AF-PHY-0086.001, version 1.1. ! Provides ICP cell insertion and filtering. ! Provides enhanced alarm correlation and fault isolation to rapidly react to link losses. ! Performs rate decoupling between the PHY and ATM layers to maintain IMA frame synchronization. ! Supports link differential delay compensation for up to 25 ms via internal memory buffer. ! Default mode is UNI mode, which will transparently transfer cells between the TC and the APC. All of the IMA functions (state machines, link addition and deletion, ICP cell processing, etc.) are implemented in hardware, thus requiring microprocessor support only in non-real-time critical operations such as configuration, error handling, and performance parameter accumulation. The IMA is divided into a link section and a group section, each handling a configurable number of links and groups, respectively. The IMA layer interfaces to the ATM layer via a UTOPIA Level 2 interface. The IMA functionality is divided between the link and group processors, which are interconnected via an internal UTOPIA level 2 interface on which the link processor is the slave and the group processor the master. The microprocessor interface provides access to the IMA cores internal registers. Communication between the link processor and the group processor occurs in-band by overwriting fixed fields of ICP and filler cells. In the transmit direction, the link processor restores the fixed information in the ICP and filler cells before transmitting them to the physical layer. TO TC LINK PROCESSOR GROUP PROCESSOR MPHY UTOPIA SLAVE UTOPIA-2 TO APC DELAY COMPENSATION BUFFER 5-9981.a (F) Figure 25. IMA High-Level Interconnect Block Diagram Agere Systems Inc. Agere Systems - Proprietary 75 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) 16.3 Multi-PHY UTOPIA Slave Interface The multi-PHY UTOPIA slave interface performs the following two functions: ! It monitors the fill level of all of its port FIFOs, and reports the cell available status of the ports polled by the multiPHY UTOPIA master. ! It transfers cells to/from the port selected by the APC, which is the multi-PHY UTOPIA master. The multi-PHY UTOPIA slave interface monitors the fill level of each group’s receive and transmit FIFO, and reports the cell available status of the link polled by the ATM layer. When a port is selected, the multi-PHY UTOPIA slave interface controls the reading/writing of cells from/to the port receive/transmit FIFO. 16.4 Link Processor The receive and transmit link state machines, the IMA frame synchronization mechanism, and the IMA error/maintenance state machine are driven by events triggered by changes in the link configuration, by the reception of ICP cells, and by faults detected by the link processor. The link control section selects the link to receive or transmit cells. In reception, a link is selected if the status from the TC indicates that it has a cell available in its FIFO and if the delay compensation buffer (DCB) can hold one more cell. In transmission, a link is selected if the status from the TC indicates that its FIFO can hold one more cell, and if the Tx FIFO in the link processor has one cell to transmit. The Rx cell processor section monitors incoming cells, determines with the CRC monitor section the validity of ICP cells, and processes the ICP cell information. It computes the initial value of the DCB write pointer, reports the reception of ICP cells to the IMA frame synchronization section, and informs the Rx LSM and Tx LSM sections of the far-end link status. The DCB control and the DCB constitute a FIFO that is large enough to absorb the required link differential delay (approximately 8 Kbytes per T1/E1 link). All received cells are written into the DCB, except for stuffed ICP cells that are used by the IMA frame synchronization section and then discarded. 76 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 16 Inverse Multiplexing for ATM (IMA) Block (continued) 16.5 Group Processor The group and group traffic state machines are implemented in the group state machine section. These state machines are driven by events triggered by changes in the group configuration, by the reception of ICP and filler cells (which contain in-band information from the link), and by faults detected by the group processor. The Rx FIFO section implements multiple independent receive FIFOs for each group, using a contiguous section of memory. The FIFO depth is 128 bytes (2.4 cells) for each receive FIFO. The receive FIFO stores the group's reconstructed cell stream that is passed to the ATM layer via the multi-PHY UTOPIA slave interface. The receive FIFO also serves as a smoothing buffer for removing the cell delay variation attributed to ICP and stuff ICP cells. The Rx cell processor receives cells from the link and performs the following functions: ! It monitors incoming ICP cells and processes the group-related ICP cell information. ! It informs the group state machine of the acceptance/rejection of the group parameters and the far-end group status. ! It extracts in-band information sent by the link processor, computes the DCB read pointer, and checks the link difference delay. ! It passes the read pointer to the Tx cell processor for transmission back to the link processor. ! It writes ATM cells into the group’s receive FIFO and discards ICP and filler cells. Agere Systems Inc. Agere Systems - Proprietary 77 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) 16.6 Delay Compensation Buffer (DCB) Figure 26 through Figure 32 illustrate the logical operation of the IMA’s DCB. The DCB for each link is represented as an individual memory in which data is stored sequentially in an upward direction, starting near the bottom and incrementing toward the top. The data storage is handled by a set of pointers that effectively make the RAM operate as a FIFO. Once the data storage hits the top of the memory, it wraps back around to the bottom (circular store). Read and write pointers keep track of where data is read or written for each DCB, and also whether the memory is near full or near empty. Figure 26 shows an example just after a three-link group has been started. Once started, the IMA looks at incoming cells. When it encounters an ICP cell on each link, it uses the frame addressing in the cell to calculate where to start loading data into the DCB the subsequent cells. Once all three links have received ICP cells, all three links will be storing data in the DCB. At this point, however, no data will be read out and forwarded to the ATM layer. DCB FOR LINK #1 DCB FOR LINK #2 DCB FOR LINK #3 TOTAL SIZE OF DCB DETERMINES TOTAL RANGE OF LINK DELAY THAT CAN BE TOLERATED BY IMA. LINK START DICB_GUARDBAND THRESHOLD TO START READING SET BY DETERMINING THE PROGRAMMABLE DELAY. e.g., 5 ms CURRENT WRITE POINTERS (AFTER SOME DATA HAS COME IN FROM LINES). THE INITIAL READ POINTER IS SET BASED ON WHERE THE EARLIEST LINK’S STARTING WRITE POINTER WAS ORIGINALLY SET. SHADING REPRESENTS DATA WRITTEN TO DCB. INCREASING ADDRESS LATEST LINK (LONGEST DELAY) DIFFERENCE BETWEEN THESE ADDRESSES CORRESPONDS TO DELAY OF READING OUT DATA, E.G., COULD BE 5 ms. MIDDLE DELAY LINK EARLIEST LINK (LEAST DELAY) INITIAL DCB WRITE POINTER BASED ON INCOMING ICP CELLS PER LINK. 1609 (F) Figure 26. Logical View of Three-Link Group’s DCB Shortly After Starting to Receive Data from the Line 78 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) Data is sent to the ATM layer once enough data has been received on all three links, which is determined by the programmable threshold, as illustrated in Figure 27. The threshold is programmable via a register and is an offset from the starting point of the slowest link. Once all three links in Figure 27 pass this threshold, the DCB starts to be read beginning at the starting read pointer of the earliest link. DCB FOR LINK #1 DCB FOR LINK #2 DCB FOR LINK #3 CURRENT WRITE POINTERS WHEN LATEST LINK PASSES THIS THRESHOLD, READING OF DCB IS STARTED. ONCE LINK #1 HAS PASSED THE START READING THRESHOLD, THE DCB IS READ STARTING AT INITIAL READ POINTER. e.g., 5 ms INCREASING ADDRESS LATEST LINK (LONGEST DELAY) MIDDLE DELAY LINK EARLIEST LINK (LEAST DELAY) INITIAL DCB WRITE POINTER. DARKER SHADED DATA WILL EVENTUALLY BE DISCARDED. 1610 (F) Figure 27. Logical View of Three-Link Group’s DCB When It Starts Reading DCB The reason for this programmable offset value is that it determines two operating behaviors of the DCB. First, it sets the range of link delays that the DCB can tolerate, both for earlier links and later ones. Second, it sets the latency of the DCB. For example, in Figure 27, the latency is 5 ms minimum. With this delay of 5 ms set, the DCB can add links that are up to 5 ms later than the current latest link. It can add links earlier by (total DCB delay) – 5 ms. Agere Systems Inc. Agere Systems - Proprietary 79 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) As can be seen in Figure 28, the DCB is read by reading across each link. In this example, link #1 is followed by link #2, etc. Because data is read out at the IDCC rate (the sum total of the line rates minus overhead), as the write pointers advance the read pointers follow, and the difference between them remains constant. DCB FOR LINK #1 DCB FOR LINK #2 DCB FOR LINK #3 CURRENT WRITE POINTERS. THRESHOLD TO START READING SET BY DETERMINING THE PROGRAMMABLE DELAY. e.g., 5 ms CURRENT READ POINTER—READS LINKS IN ROUND-ROBIN FASHION. INITIAL READ POINTER. INCREASING ADDRESS LATEST LINK (LONGEST DELAY) MIDDLE DELAY LINK EARLIEST LINK (LEAST DELAY) INITIAL DCB WRITE POINTER. DARKER SHADED DATA WILL EVENTUALLY BE DISCARDED. 1611 (F) Figure 28. Logical View of Three-Link Group’s DCB After It Starts Reading DCB 80 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) In Figure 29, during normal operation, the DCB pointers are just following each other. As they reach the top of memory, the pointer just wraps around to the bottom. DCB FOR LINK #1 DCB FOR LINK #2 DCB FOR LINK #3 CURRENT WRITE POINTERS e.g., 5 ms CURRENT READ POINTER: READS LINKS IN ROUND-ROBIN FASHION. DATA IS READ AT THE IDCC RATE. INCREASING ADDRESS LATEST LINK (LONGEST DELAY) MIDDLE DELAY LINK EARLIEST LINK (LEAST DELAY) 1612 (F) Figure 29. DCB During Normal Operation In Figure 30, a new link is about to be added. Prior to adding a link to the group, the link is enabled and data is received. Once an ICP cell has been received on this new link, the IMA calculates where in the DCB the data is to be stored. In Figure 30, a fourth link has been enabled, an ICP has been found, and data reception and storage into the DCB has started. This link will not be added to the round-robin reading until the read pointer has moved to an area memory containing valid data. This is shown in Figure 31. Agere Systems Inc. Agere Systems - Proprietary 81 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) DCB FOR LINK #1 DCB FOR LINK #2 DCB FOR LINK #3 DCB FOR LINK #4 THE DATA WILL WRAP AROUND ONCE THE WRITE POINTER REACHES THE END OF THE BUFFER. THIS AREA OF MEMORY, FROM THE FASTEST LINK’S WRITE POINTER TO THE READ POINTER, DETERMINES EARLIEST LINK THAT CAN BE ADDED. LINK #4 ICP CELL FOUND, DATA STARTING TO BE WRITTEN INTO BUFFER. THE AMOUNT OF MEMORY BETWEEN THE SLOWEST LINES READ AND WRITE POINTER DETERMINES THE LATEST LINE THAT CAN BE ADDED. E.G., 5 ms INCREASING ADDRESS LATEST LINK (LONGEST DELAY) MIDDLE DELAY LINK EARLIEST LINK (LEAST DELAY) SLIGHTLY LATER THAN LINK #1 CURRENT READ POINTER. 1613 (F) Figure 30. Starting to Add a Link to a Group 82 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) DCB FOR LINK #1 DCB FOR LINK #2 DCB FOR LINK #3 DCB FOR LINK #4 CURRENT WRITE POINTERS THE READ POINTER HAS CAUGHT UP TO THE STARTING WRITE PONTER OF LINK #4, AND THE FOURTH LINK NOW JOINS THE GROUP. CURRENT READ POINTER INCREASING ADDRESS LATEST LINK MIDDLE (LONGEST DELAY) DELAY LINK EARLIEST LINK SLIGHTLY LATER (LEAST DELAY) THAN LINK #1 1614 (F) Figure 31. Link Now Being Read Agere Systems Inc. Agere Systems - Proprietary 83 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) Figure 32 illustrates what happens in the DCB when the links fall out of their delay window. A fault on link #3 causes it to nearly catch up to the read pointer. A fault on link #4 causes it to run too slow. The faults shown in this figure are not yet causing a loss of delay synchronization (LODS). If either link #3 catches the read pointer or if link #4 is caught by the read point, a LODS condition will occur. DCB FOR LINK #1 DCB FOR LINK #2 DCB FOR LINK #3 DCB FOR LINK #4 LINK #3’s DCB IS NEARLY FULL. OUT OF DELAY SYNCHRONIZATION IS DETERMINED WHEN THE DIFFERENCE BETWEEN THE WRITE POINTER AND READ POINTER AND READ POINTER GROWS TO A PROGRAMMED AMOUNT. LINK #4’s DCB IS NEARLY EMPTY. IF THE READ POINT CATCHES UP WITHIN 4 CELLS OF THE WRITE POINT, THEN THE LINK WILL BE CONSIDERED OUT OF DELAY SYNCHRONIZATION. CURRENT WRITE POINTERS CURRENT READ POINTER: INCREMENTING AT THE IDCC RATE. INCREASING ADDRESS LATEST LINK (LONGEST DELAY) MIDDLE DELAY LINK EARLIEST LINK SLIGHTLY LATER (LEAST DELAY) THAN LINK #1 1615 (F) Figure 32. Effects of Link #3 and Link #4 Faults 84 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 16 Inverse Multiplexing for ATM (IMA) Block (continued) 16.7 Programming the DCB 16.7.1 Link Start-Up Guardband Field This value is set in the setup file utility and is the number of cells in the DCB of the slowest link (i.e., the link with the most transport delay) before the IMA round robin is started. It must be greater than four. This value should be kept as small as possible, since all links will incur the delay programmed into this field. When adding a link, the value programmed here also corresponds to the value of the slowest link that can be added. For example, if the guardband delay is 5 ms, a link can be added up to 5 ms slower than the slowest link. When programming the delay for these fields: delay = guardband value * cell rate (53 bytes in one ATM cell). If a delay of 5 ms is desired: ! T1 cell rate 276 µs: 18 * 276 µs = 4.96 ms. Program the guardband field to 18. ! E1 cell rate 221 µs: 23 * 221 µs = 5.083 ms. Program the guardband field to 23. (Two of the 32 time slots in E1 are used for signaling. 30 bytes arrive at a rate of 125 µs.) 16.7.2 Link Maximum Operational Delay This value is called LddSpread in npIMAGroupParameters. This field is the threshold in the number of cells above which a loss of delay synchronization is declared: the total window of time allowed between the slowest and fastest link. A link can be added that is the link maximum operational delay field: link start-up guardband delay. If the link maximum operational delay is 25 ms and the link start-up guardband is 5 ms, a link can be added that is up to 20 ms faster than the current slowest link. The formula for determining the value in this field is the same as the link start-up guardband fields. 16.8 Features Not Supported in IMA Table 22 lists the ways in which the TAAD08JU2 device differs from the IMA PICS Proforma, as given in IMA specification AF-PHY-0086.001, dated April 1999. Table 22. TAAD08JU2 Exceptions to the IMA PICS Proforma PICS Proforma Mandatory/ Optional BIP 21, R-15 M The TAAD08JU2 IMA implementation complies with BIP 21. The IMA can be programmed to operate either in a v1.1 or a v1.0 operating mode. When programmed to operate as a v1.1 IMA controller, the TAAD08JU2 IMA follows the definitions for bytes 18 and 19 outlined in Figure 36 of the IMA specification (AF-PHY-0086-001). When programmed to operate in v1.0 IMA mode, TAAD08JU2 will follow the definitions noted in Figure 37 of the IMA specification. BIP 70-71, O-5 and 6 O TAAD08JU2 supports two of the three symmetry modes: Comment ! Symmetrical configuration and operation mode ! Symmetrical configuration and asymmetrical operation mode Asymmetrical configuration and operation mode is not supported. Agere Systems Inc. Agere Systems - Proprietary 85 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 Table 22. TAAD08JU2 Exceptions to the IMA PICS Proforma (continued) PICS Proforma Mandatory/ Optional IDC.2, IDC.7 R-64, R-69 M LDD.2, R-75 M The TAAD08JU2 buffer has a depth of 128 cells. At E1 operation, the buffer size is 128 cells x 221 µs = 28.29 ms, and at T1 the buffer size is 128 cells x 276 µs = 35.33 ms. Therefore, if the number of cells in the DCB of the slowest link (dcbGuardband) is set to 5 cells, at E1 operation the LDD is equal to 28.29 ms – (5 cells x 221 µs) = 27.19 ms; for T1 operation with the same dcbGuardband setting, the link differential delay (LDD) is equal to 33.95 ms. It is possible to have an LDD less than 25 ms if the dcbGuardband is set to a higher value or the LDD_SPREAD (number of cells above which a loss-of-delay synchronization is declared) is set to a lower value. OAM.52-53 O-26, O-27 O TAAD08JU2 performance parameters are only accumulated over a 1 s interval. OAM.76-77, O-30 O Only the default value of 2.5 s is supported by TAAD08JU2. IPM.4 R-158 M The TAAD08JU2 IMA does not supply the host processor with cell rate information. TAAD08JU2 does notify the host when links have been added or deleted. The host can then compute the impact on cell rate. This satisfies R-158. MIB.1-3, O32, O-33, CR-17 O The IMA MIB structure is not built or populated by TAAD08JU2. CIT.14 M If a timing mismatch occurs during startup of an IMA group, the IMA state machine will go to the Startup state. It should go to the Insufficient-Links state. This violates R-63 of the IMA specification. OAM.32—33 M OAM.40,42 O If an IMA link is pulled at the hub (i.e., both Rx and Tx at the same time), the FELSR1 register for that link is not updated (it shows Active when it should be Unusable). When the link is reconnected to the hub, the RX_UUS_FE alarm is quickly set and cleared. Likewise, no far-end IMA_RDI indication is given until the link is plugged back in. This violates the IMA specifications R-132, R-133, R-143, and R-144. 86 Comment IDCC is not implemented in the TX direction. The rate at which cells are played out is governed by the line clock. Further, a caveat to TAAD08JU2’s computation of cell rate in the RX direction assumes the use of nonfractional T1, J1, or E1 links in the IMA group. This also means that fractional formats such as T1-DDS cannot be used for IMA. Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block 17.1 Introduction The APC block is a highly integrated module that provides the ATM layer functions of an ATM switching system. The block is based upon Agere’s ATM port controller (APC) IC, which is a part of the ATLANTA R2 chip set. The core of the APC is integrated along with the memories to store connection tables, cells, pointers, and control information in the APC block. The APC block can support 2K VCs and has a cell buffer capacity of 4K cells. There are three distinct differences between the APC block in TAAD08JU2 and the APC device, as follows: ! The smaller number of connections supported by TAAD08JU2 due to the smaller on-chip memories compared to the external memories used with the APC device (see Figure 33). VCRAM 64K x 64 bits BRAM 64K x 32 bits CRAM 32K x 32 bits PRAM 16K x 8 bits UTOPIA A INTERFACE SWITCH FABRIC INTERFACE APC BLOCK APC EXTERNAL STATISTICS INTERFACE UTOPIA B INTERFACE 1940 (F) Figure 33. APC Block Integrated Memory Configuration ! The presence of the on-chip ARM processor and associated firmware that allow the user to interact with the APC block through commands and indications rather than direct register reads and writes. ! Only the APC's UTOPIA B interface is available externally. UTOPIA B connects to TAAD08JU2's expansion interface and to the SAR. UTOPIA A connects internally to the IMA block. Also note that because the SAR's UTOPIA is 16-bit only, the APC's UTOPIA B is limited to 16-bit only. The APC block operates in one of three distinct modes: ! Single APC switch mode. This mode corresponds to the case in which the switch fabric interface is not connected to anything. In this mode, all connections in the APC are between MPHYs. ! Dual APC switch mode. In this mode, the switch fabric interface on TAAD08JU2 is connected back-to-back with the switch fabric interface on either another TAAD08JU2 or an APC device. Connections in this mode can either be between an MPHY and the fabric (in which case the cells are destined for the other TAAD08JU2/APC) or between two MPHYs (indicating that the cells on this connection enter and leave the same TAAD08JU2 device by its MPHYs). ! Port card mode. In port card mode, the switch fabric interface is connected to a switch fabric device. In this mode, all connections in the APC are between an MPHY and the fabric. These modes are set via the setup file utility. Note: Advisories are issued as needed to update product information. When using this data sheet for design purposes, please contact your Agere Account Manager to obtain the latest advisory on this product. Agere Systems Inc. Agere Systems - Proprietary 87 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block (continued) In both port card mode and dual APC switch mode, setting up a complete unidirectional connection requires issuing two device manager function calls, as follows: ! First, the npAddConnection() is given to one TAAD08JU2/APC to set up a connection in the fabric-to-MPHY direction (set by the ConnDir field). This command returns a 64-bit connection tag in an indication if successful. ! Second, the npAddConnection() is sent to a different TAAD08JU2/APC to set up a connection in the MPHY-tofabric direction. This is the second half of the complete connection, and the command takes as a parameter the connection tag returned by the command issued to the first TAAD08JU2/APC. TRANSMIT PORT A TRANSMIT PORT B TAAD08JU2 1 RECEIVE PORT A RECEIVE PORT B FABRIC SIDE PHY SIDE TRANSMIT PORT A TRANSMIT PORT B TAAD08JU2 2 RECEIVE PORT A RECEIVE PORT B 1616 (F) Figure 34. Switch Fabric Connections for Dual TAAD08JU2 Switch Mode 17.2 Architecture Refer to the ATLANTA ATM Port Controller Advance Data Sheet, the APC User Manual, or the ATLANTA architecture specification for detailed information about the APC architecture. Except where required for essential TAAD08JU2 implementation-specific details of the APC block, only short descriptions will be provided in this document, or references will be made to appropriate ATLANTA APC documentation. Since the entire APC block is imported into TAAD08JU2, the subblock level descriptions are not documented here. The APC block provides support for ATM layer functions including connection management, buffer management, class-of-service VC scheduling, and VP or VC switching. The APC also provides considerable statistics gathering functions. 88 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block (continued) 17.3 Features ! Supports up to 2k connections with no external memory required. ! Supports multiple traffic classes: — CBR (up to 32 rates per port). — rt-VBR (up to 16 rates per port). — nrt-VBR. — ABR. — UBR. ! Provides I.610 and ATM TM4.1 traffic management. ! Supports per-VC dual leaky bucket UPC policing per generic cell rate algorithm (GCRA). ! Provides performance monitoring and OAM cell processing (end-segment behavior): — Up to 128 flows may be monitored. ! Buffer management per adaptive threshold algorithm: — Per-VC thresholds for optimum buffer utilization. — Adapts excess buffer allocation to current buffer occupancy. ! Provides per-port hierarchical scheduler. ! Enables support for up to 155 Mbits/s PHY links (bypassing the low-speed framer and IMA). ! Provides a comprehensive single-chip solution for implementing all ATM layer functions needed at an ATM switch port. ! Can be configured in a variety of switching modes for flexible operation: — Performs as an ATM switch port card by supporting linear aggregation of up to 311 Mbits/s of ATM traffic at the physical layer interface (full duplex). — Operates as a stand-alone single-chip 32 x 32 shared memory switch or a N:1 concentrator. — Operates in conjunction with the Agere APC (ATM port controller device) or with another TAAD08JU2 as a 2 x 2 dual APC-based switch (no separate external switch fabric needed). ! Performs ATM layer user network interface (UNI) and network node interface (NNI) management functions: — Controls up to 31 full-duplex MPHY (multiple physical layer) ports on the physical layer side. — Allows any MPHY to be configured as a UNI or an NNI. — Optionally, translates or passes the generic flow control (GFC) field of the egress ATM cell header for NNI or UNI applications. — Performs virtual path identifier (VPI)/virtual channel identifier (VCI) translation for up to 4K connections on egress while allowing reusability of same VPI/VCI on different UNIs. ! Maintains a variety of optional per-connection, per-port, and per-device statistics counters in external memory and on-chip. ! Provides dual interfaces to ATLANTA switch fabrics to facilitate construction of redundant systems for fault tolerance. ! Implements a flexible, efficient buffer/congestion management scheme based on a novel Bell Labs patented adaptive dynamic thresholding (ADT) algorithm: — Supports selective cell discard and early/partial packet discard (EPD/PPD). — Provides capability for minimum buffer reservations on per-connection, per-class, and per-port basis. ! Performs ATM forum-compliant available bit rate (ABR) explicit rate flow control using the highly efficient and sophisticated Bell Labs patented ALBERTA algorithm. Provides optional support for EFCI marking as well. Agere Systems Inc. Agere Systems - Proprietary 89 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block (continued) ! Provides OAM fault management functions for loopback, continuity check, defect indication on all connections, and performance monitoring for up to 127 processes. ! Provides an enhanced services interface (ESI) to support operation of an optional external adjunct device for comprehensive statistical data collection and virtual source/virtual destination (VS/VD) ABR functionality support. ! Queues up to 4096 cells in on-chip memory. 17.4 Summary of Commands The commands associated with the APC block fall into the following categories: ! Initialization: used only at device initialization to set modes, enable features, configure ports, and set identifier ranges and scheduler limits. ! Configuration: used to configure the schedulers at initialization and dynamically. ! Masks: masks alarm interrupts and statistics reporting. ! Connection: used to set up and tear down individual connections and set their associated parameters, statistics gathering, fault state, and OAM behavior. ! Cell insertion: used to insert cells onto bidirectional connections. ! Statistics: used to retrieve global or per-connection statistics. 17.5 Buffer Management The following buffer management thresholds exist in the APC: ! CLP1 discard threshold ! CLP0+1 discard threshold ! Partial packet discard (PPD) ! Early packet discard (EPD) threshold ! Selective EFCI (SEFCI) marking threshold ! Ingress fabric backpressure (IFBP) alarm threshold These thresholds, with the exception of SEFCI and IFBP, are dynamic and provided per-VC for each traffic class. Except for IFBP, which exists only on the ingress, all other thresholds are provided both on the ingress and egress sides. Furthermore, SEFCI and IFBP are provided per-port, per-traffic-class on the ingress, and per-subport and per-traffic-class on the egress. The per-VC thresholds are dynamic because they change depending on the amount of free buffer space available. The larger the free buffer space, the higher the threshold. The APC provides the option to configure the per-VC dynamic thresholds as static thresholds, as well as the capability to partition the buffer space between the five traffic classes. Thus, there is flexibility for using the buffer space differently for different traffic classes. For example, large buffers can be used for nrt-VBR traffic class, implying smaller effective bandwidths for given traffic descriptors. For CBR and rt-VBR traffic classes, CDV and CTD constraints dictate smaller buffer requirements. The APC also provides the facility to limit the buffer usage per-port, per-traffic-class on the ingress, and per-subport and per-traffic-class on the egress. The APC buffer management scheme makes use of the concepts of effective scheduling bandwidths and buffer allocations to guarantee (to a given statistical probability) delay and cell loss ratios for connections admitted on the condition of their conformance to leak-bucket traffic regulation, and it provides those effective bandwidth and buffer values on a per-VC basis within the VC parameter table entry. 90 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 17 ATM Port Controller (APC) Block (continued) The effective buffer allocation for a VC is not used immediately as the discard threshold, but conditioned on the available reserve of the common buffer pool relative to a congestion threshold established for the traffic class to which the VC belongs. The overall reserve and specific congestion conditions add to and subtract from the value of the buffer allocation, to continuously adjust the threshold to balance the losses of the given connection against the overall state of the system. This concept is elaborated for use in APC with its multiple classes of service. Additional supported objectives and features include the following: ! A VC that is not being policed and is violating its leaky bucket regulator contract is prevented from consuming excessive buffer space. ! Different traffic classes support buffer reservation independent of other classes. The reservation mechanism supports partial or complete partitioning of the total buffer space. ! Within a traffic class, different ports/subports support buffer reservation independent of other ports/subports. ! For a given VC, different kinds of loss determinations (e.g., CLP1, CLP0+1, EPD) are provided, as well as thresholds for other purposes such as SEFCI. ! Within a class, it is possible to provide different loss priorities to different VCs and to provide any or all VCs a guaranteed minimum buffer space during periods of heavy buffer use. ! The APC supports 64 per-VC guaranteed levels per traffic class and threshold type. SEFCI and IFBP thresholds are based on static thresholds. The CLP1 and CLP0+1 thresholds trigger cell discarding based on the value of the CLP1 field in the ATM header. CLP1 thresholding operates on CLP1 cells only, while CLP0+1 thresholding operates on both CLP0 and CLP1 cells. CLP transparent mode is also supported. PPD occurs if a user cell is discarded because of a policing violation, a CLP1 threshold violation, a CLP0+1 threshold violation, or no free buffer space available. The remaining user cells are dropped (up to but not including the next end-of-frame cell). EPD threshold is only evaluated on the first user cell of a VC after the (AAL5) end of frame is detected using the PTI field of the cell header. All user cells are dropped (up to and including the next end-of-frame cell). The ingress fabric backpressure (IFBP) alarm threshold does not explicitly cause cell discard. It is used to set an interrupt indicating that the ingress buffer is being congested by cells for a particular destination fabric port and traffic class. When the threshold is exceeded, status information is generated to indicate the congested destination fabric port and traffic class. Selective explicit forward congestion indication threshold (SEFCI) marking occurs only if enabled. If the SEFCI threshold is violated and the cell's ATM header PTI field indicates a user data cell, it is marked to indicate congestion was experienced. SEFCI is supported only for NVBR, ABR, and UBR classes. Dynamic threshold function permits allocation of buffers to individually overloaded connections when there are large reserves of unoccupied buffers, avoiding losses that would be incurred by the overloaded circuits under a static threshold policy. Conversely, as overall class occupancy approaches a preconfigured trigger point (B), the extra allocation is withdrawn, supporting the cell loss probability (CLP) performance associated with using the effective buffer (b_e) requirement for each connection. Similarly, between B and a second preconfigured trigger point (R), the slope of the dynamic threshold function moves the threshold toward each connection's guaranteed value (b_g), providing a controlled reduction of all connections to their guaranteed values for those cases where general system overload peaks make it impossible to support the b_e values for all connections. Agere Systems Inc. Agere Systems - Proprietary 91 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block (continued) 17.6 Scheduling Cells that arrive and are not dropped by the APC are stored in memory while they await transmission. The stored cells are organized into FIFO linked lists called virtual connection (VC) queues. There is one VC queue for each VC. The virtual connection could be a virtual channel connection (VCC) or a virtual path connection (VPC). Empty memory locations available for cell storage are contained in the free cell queue. The free cell queue is organized as a LIFO stack. All memory configured for cell storage belongs either to a VC queue or the free cell queue. When a cell arrives after VC table look-up, and if the cell is admissible (e.g., by the policer at the ingress only), it is passed for enqueueing to the VC queue structure that is part of the associated VC table entry. If necessary, that VC is then scheduled in the flow queue for the associated traffic class. A flow queue is a collection of VCs organized as a queue for the purpose of scheduling. A flow is characterized by the parameters output port, traffic class, and bit rate (or weight). 17.6.1 Ingress Scheduling APC's scheduling mechanism selects which VC queue to service at each cell dispatch time, thus enabling it to meet the QoS requirements of different VCs while simultaneously promoting high utilization of both bandwidth and buffers in the switch. APC scheduling functions support five traffic classes: CBR, rt-VBR, nrt-VBR, ABR, and UBR. VCs are configured at call setup to belong to one of the five traffic classes. Service is divided into three levels of priority, as follows: ! The highest priority is accorded to CBR traffic. ! The second priority level is for guaranteed traffic services (GTS), which include the rt-VBR, nrt-VBR, ABR, and UBR classes. Within GTS, the different services are shaped according to programmable rates. ! The lowest priority level is called excess bandwidth service (EBS) and is available in weighted shares for rt-VBR, nrt-VBR, ABR, and UBR, as determined by programmable weights. Scheduling is implemented in a hierarchy of three tiers. At the first level, VCs within a traffic class are scheduled among each other. At the second level, traffic classes within a priority are scheduled. At the third level, service is scheduled among the three priorities. Within a traffic class, VCs are grouped into flows that are organized with queues. Algorithms used to schedule service to VCs depend on the traffic class to which the VC belongs. These scheduling algorithms are based on manipulating the flow queues belonging to the traffic class. The state of the flow queues is kept in flow tables managed by the scheduling algorithms. CBR is scheduled in a non-work-conserving manner by a shaper algorithm according to the rate assigned to each VC. rt-VBR is scheduled in a work-conserving manner by a starting potential fair queueing (SPFQ) algorithm according to the rate assigned to each VC. nrt-VBR is scheduled in a work-conserving manner by a weighted round-robin (WRR) algorithm according to the weights assigned to each VC. ABR (UBR) is scheduled in a workconserving manner by a weighted round-robin scheduler and is provided bandwidth proportional to MCR above MCR. GTS provides non-work-conserving service via a shaper algorithm to its four associated VC schedulers according to the rate assigned for the aggregate of all VCs in the class associated with each of those first-level schedulers. EBS provides work-conserving service via self-clocked fair queueing (SCFQ) to its four associated VC schedulers according to the weight assigned for the aggregate of all VCs in the class associated with each of those first-level schedulers. 92 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 17 ATM Port Controller (APC) Block (continued) 17.6.2 Fabric Backpressure Ingress scheduling is augmented by backpressure from the fabric. For each of the fabric ports, there are 5 bits of backpressure status that apply to the traffic classes CBR, rt-VBR, nrt-VBR, ABR, and UBR. When backpressure is asserted for a particular traffic class and port, no traffic of that flow is transmitted by the APC. Backpressure status is continuously updated via the egress cell stream from the fabric to the APC. The APC provides a mechanism to map the four classes in the fabrics to the five classes used by the APC. Backpressure controls the effects of output congestion in the fabric. This functionality offers the important advantage of having a completely lossless switching fabric. Furthermore, it precludes the necessity of large and expensive buffering in the switch fabric by shifting the congestion to the ingress port where there is sufficient buffering. That is, only a small amount of buffering is provided per fabric element and shared by all input ports, and there are three orders of magnitude more buffers in the port that can be shared between the ingress and egress. 17.6.3 Egress Scheduling The APC provides 32 independent schedulers that can be individually mapped to any of the 31 subports and the microprocessor subport. There are two different scheduling arrangements, based on the subport number. The first 16 subports (0 through 15) are provided with scheduling structures similar to that used at ingress (Type 1 subports), with some modifications. The remaining subports (16 through 31) are provided with a somewhat simpler configuration, using weighted round robin (WRR) for all classes at the first level (Type 2 subports). However, at the egress there is an additional level of scheduling (common to both Type 1 and Type 2 subports) that governs the distribution of service to the 32 subports. It consists of a UTOPIA rate scheduler (URS) and UTOPIA excess bandwidth scheduler (UES) and the associated priority selector policy. The URS provides guaranteed bandwidth to fixed rate UTOPIA subports, while the UES offers unused bandwidth, leftover from URS, in a workconserving manner to the MPI (address 31) and to one other specially designated subport (e.g., one occupied by a SAR). The MPI has strict priority over the specially designated UTOPIA subport. The URS/UES can be operated in two alternative modes: one applicable to direct connections to the UTOPIA MPHY devices; the other applicable to operation with an external port inverse multiplexer (PI-MUX) device. In the direct connection case, multiple subport schedulers may be logically mapped to a single MPHY subport destination, thus providing what is called a virtual PHY partitioning of one physical link into separate, rate-controlled flows (for example, to construct multiple virtual paths, each with guaranteed bandwidth, within a physical link). 17.7 ABR Flow Control The APC supports an ABR flow-control mechanism based on adaptive load/buffer explicit rate algorithm (ALBERTA). ALBERTA is fully compliant with the ATM Forum standards, has fast transient response, provides fast convergence, supports two types of fairness criterion (proportional to MCR and MCR plus equal share), requires few parameters, and allows high link utilization. In addition, the APC's ABR solution provides the flexibility to support two kinds of switch behavior: explicit rate (ER) marking and selective EFCI. ALBERTA measures both traffic load and queue length to control the rate of an ABR connection. RM cells are marked with congestion information both in the forward and backward directions; bidirectional ER marking improves convergence time. On ingress, congestion is per-APC based, and on the egress, congestion is per-subport based. The APC supports ABR point-to-multipoint connections and provides consolidation of RM cells in the backward direction. Agere Systems Inc. Agere Systems - Proprietary 93 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block (continued) 17.8 Control Plane Functions Control plane functions are triggered during the connection setup phase and are composed of the following functions: generic call admission control (GCAC) and call admission control (CAC). Working in conjunction with the PNNI routing protocol, select a path through the network from the source to the destination that has a high likelihood of meeting the network resource requirements of the connection. GCAC is an important part of this functionality. Because the path selected during connection setup phase is only a best guess due to the latencies and periodicity involved in the PNNI routing protocol, and also because it is based on a limited set of advertised nodal and link state attributes and metrics, GCAC alone is not enough to guarantee the QoS for the connection. The switching system also performs local CAC upon receipt of a connection setup request. The decision to accept or reject the connection is the fundamental objective of CAC and is based on the connection's traffic descriptors, QoS requirements, and the currently available resources (buffer and bandwidth) in the switch after accounting for resources already committed to guarantee QoS of existing connections. In addition, CAC must satisfy conflicting requirements of promoting efficient use of switch resources while making its decisions on-the-fly. This imposes severe limitations on the computational complexity of the algorithm. 17.8.1 APC Support for Control Plane Functions The APC provides the necessary functions to configure connections and establish connection records. Its built-in hardware counters provide a set of statistic collections that can be used to support PNNI functions (routing topology update) and CAC, precluding the need for reliance on any particular traffic model. The APC facilitates the design of an admissible region that is coupled to its shared memory buffer management, leading to fair, efficient, and robust algorithms. Furthermore, opportunity is provided to integrate the CAC with the measured state of buffer and bandwidth obtained from the measurement function, leading to another degree of enhancement. 17.9 Management Plane Functions 17.9.1 Operation Administration and Maintenance (OAM) The APC supports the following OAM functions: ! Connections can be configured independently for either F4 or F5 flows as connection endpoints, segment endpoints, and intermediate points. ! The detection of OAM cells is used to identify end-to-end and segment OAM cells for F4/F5 flows, recognize OAM cell type and function type, and optionally discard/capture (under microprocessor control) OAM cells with invalid or nonsupported functions. ! The error detection and protection are done by CRC-10 error-detection code check on incoming OAM cell payloads, discarding OAM cells with payload errors, and CRC-10 generation on transmitted OAM cells. ! Alarm indication signal (AIS) and remote defect indication (RDI) are used to support fault management. End-toend VP and VC AIS/RDI defect indication cell generation and processing are supported on a per-connection basis. ! Continuity check is supported on all VCs, configured as either a source, a sink, or a combination thereof, with a continuity check cell insertion performed in the absence of user traffic. Activation and deactivation functions are performed by cell insertion/extraction through the microprocessor interface. 94 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block (continued) ! Detection and looping back of end-to-end and segment loopback cells at the corresponding endpoints and optional loopback cell detection and looping back at the intermediate points (i.e., neither connection nor segment endpoints) based on loopback location ID for all virtual connections at the rate of up to 100 cells per second is supported in the APC. The generation and insertion of the loopback cells is performed via the microprocessor interface. ! The APC provides performance monitoring support through automatic generation, insertion, and capture of forward monitoring and backward reporting cells for VP/VC segment or end-to-end flows. Performance monitoring support is provided for up to 127 flows. Standard block sizes of 1024, 512, 256, and 128 user cells are used. ! The APC supports generation of forward monitoring flow at the connection or segment endpoint, and receipt of the associated backward reporting flow with optional data collection, where the connection or segment endpoint is configured as PM source. ! The APC supports receipt of forward monitoring flow at a connection or segment endpoint with data collection, receipt of forward monitoring flow at the connection or segment endpoint, and generation of associated backward reporting flow with optional data collection. 17.10 Statistics Counters The APC provides the unique capability for on-line estimation of the QoS parameters such as CDV and CLR for different traffic classes on a VC/port/subport basis. This capability is made possible by measurements based on built-in hardware counters that APC provides for statistics collection. APC's measurement capability does not need to rely on any particular traffic model because it is based on continuously monitoring the cell flow intensities and queue occupancies to infer desired information on target parameters. By supporting novel concepts of effective buffer occupancy and effective service rate, this unique capability overcomes the fundamental limitations of real-time, in-service performance monitoring in complex switching architectures with multistage queuing, backpressure, adaptive dynamic thresholding, and WFQ scheduling. APC's collection of on-chip statistics counters also supports thresholding for buffer management, ABR explicit rate calculation, and performance management. On the ingress, counters are provided per VC, per destination fabric port, per traffic class, and per APC. On the egress, counters are provided per VC, per egress subport, per traffic class, and per APC. 17.11 Ingress Enqueue Operations The APC performs ingress enqueue operations on the cell path that begins from the ingress PHY side interfaces and ends at the APC ingress cell buffer. These ingress enqueue operations include the following: ! PHY interface termination ! Connection look-up ! OAM processing ! Policing ! Buffer thresholding ! Cell and VC enqueue Cells traverse the ingress enqueue path in a serial manner. Cells removed from the cell stream by a function are generally not processed by downstream functions except where noted. Modifications to the cell by upstream functions are seen by downstream functions. If an upstream function captures or turns around the cell, subsequent downstream functions will use the capture or turnaround VC information.The APC supports two UTOPIA II interfaces (A and B) and a microprocessor cell insertion capability. Agere Systems Inc. Agere Systems - Proprietary 95 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block (continued) Each of the UTOPIA II interfaces can control up to a maximum of 31 physical layer devices (MPHYs). Each physical layer device feeds into an internal FIFO that is used to synchronize cell arrivals to the time slot. In each time slot, a cell may be selected from one of the three FIFOs to produce a single stream of cells. Selection of a cell is based on the status of these three FIFOs, the status of the ingress OAM insertion request, and the selection mode as indicated by the value of uservice. FIFOs are classified as either high priority or low priority. High-priority FIFOs are always served before low-priority FIFOs. FIFOs at a given priority are served in a work-conserving, round-robin fashion. The UTOPIA II interface A FIFO is always high priority, and the microprocessor insertion FIFO is always low priority. The UTOPIA II interface B FIFO can be configured as high or low priority by the value of uservice. In addition, ingress OAM insertion requests preempt low-priority cells. If no high-priority cells are available when the OAM insertion is pending, no cell will be selected for processing (idle ingress enqueue time slot). In order to insert ingress OAM cells, an idle ingress enqueue time slot is required. Naturally occurring idle time slots are expected in an ingress cell stream consisting of high-priority cells (e.g., the ATM rate resulting from an OC-12c PHY). OAM ingress cell insertions are performed during these idle time slots. When ingress OAM cell insertions are being supported, it is required that a sufficient idle time-slot rate occur in the high-priority ingress cell stream. The resulting ingress cell stream is processed by subsequent downstream operations. The effect of interface termination on the cell is either of the following: ! Pass (P). The cell is accepted into the cell stream. ! Discard (D). A protocol or parity error occurred, and the cell is not accepted into the ingress cell stream. Cells discarded at this point are not seen by any downstream function. 17.11.1 Connection Look-Up Look-up is performed at the arrival time of the cell. The ingress cell stream received from the PHY interface termination function is mapped to an ingress virtual-connection table (IVT) entry using user configured look-up tables and ATM header and additional extended header information received with the cell. The extended header contains additional fields to enable special features used in connection look-up. Cells received from the microprocessor insertion FIFO always use the extended cell format. Cell type recognition is also performed. A three-level look-up is performed. Look-up table 1 (LUT1) is accessed using the MPHY port number. It contains 32 entries to support 31 external MPHY ports and the microprocessor insertion FIFO. Look-up table 2 (LUT2) is accessed using virtual path information from the ATM header. It supports up to 4096 entries. Look-up table 3 (LUT3) is accessed using virtual channel information from the ATM header. It also supports up to 4096 entries. An additional look-up compression table (LUCT) is used between LUT2 and LUT3. This table allows the APC to globally handle selected VCI values less than 32 without allocating LUT3 and VC table entries for them. This saves CRAM and VCRAM memory. For example, if only three VCI values are supported below 32, the LUCT can be configured to compress the VCI range from 0 to 31 into only three LUT3 entries. This saves 29 LUT3 entries for each VP supported. LUCT is not used if the cell's VCI value is >31. If the cell's VCI value is <32, the LUCT offset value is used instead of VCI when computing the index into LUT3. As a prelude to accessing the ingress VC table (IVT), located in VC RAM, a VC index (VCX) is obtained as a result of LUT3 look-up. The VCX is used to map the cell to one of 64K entries in IVT corresponding to 2048 ingress virtual connections that are supported by APC. The resulting VC entry contains connection information used to complete the look-up process and support other operations performed on the cell. The effect of the connection look-up function is one of the following: ! Pass (P). The cell is successfully mapped into an enabled VC entry. 96 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 17 ATM Port Controller (APC) Block (continued) ! Discard (D). The cell is removed from the cell stream. Reasons for discard include: — Idle/unassigned cell — Invalid PTI field: VPI/VCI out-of-range — Look-up compression table (LUCT) entry invalid — Connection inactive — OAM or RM cell payload CRC error ! Turnaround (T). The cell is mapped to a special turnaround VC entry for subsequent processing. Turnaround is initiated by a bit in the cell's extended header. ! Capture (C). The cell is mapped to a special capture VC entry for subsequent downstream processing. Capture is initiated by a bit in the cell's extended header. Reasons for capture include: — Cell capture flag set — Invalid PTI field — VPI/VCI out-of-range — LUCT entry invalid — Connection inactive — OAM or RM cell payload CRC error 17.11.2 OAM Processing Cells recognized by the connection look-up function as ATM OAM cells are processed by this function. The type of processing depends on the OAM cell type, whether or not the particular cell is supported, and whether or not the on-chip OAM processing is enabled. The disposition of OAM cells is determined prior to VC table access based on APC configuration, and look-up table information. The effect of the OAM function is one of the following: ! Pass (P). The cell is not an OAM cell or should not be terminated at this APC. OAM state information in the lookup or VC table may be updated. ! Discard (D). The terminated OAM cell is removed from the cell stream. OAM state information in the look-up or VC table may be updated. Reasons for discard include: — Unsupported OAM cell — Undefined OAM cell/function type — Terminated sink point for OAM flow ! Insert (I). The OAM function has inserted a new cell into the cell stream (downstream of PHY interface termination and connection look-up) during an idle time slot. ! Turnaround (T). The terminated and supported OAM cell is removed from the cell stream, and another OAM cell is generated for the upstream path (using the turnaround VC). ! Capture (C). The OAM cell must be rerouted using the capture VC for processing. Reasons for capture include the following: — Unsupported OAM cell — Undefined OAM cell/function type — Supported by capture for processing external to APC Cells discarded or captured by the connection look-up function are not seen by the OAM function. On-chip OAM processing can be disabled by a configuration register bit. Agere Systems Inc. Agere Systems - Proprietary 97 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 17 ATM Port Controller (APC) Block (continued) 17.11.3 Policing APC uses the VC table information to perform policing if necessary. The policing function is used to monitor a connection to ensure that it conforms to the negotiated traffic contract, established at call setup. The policing function provided by the algorithms implemented in the APC fully meets the requirements and recommendations of the ATM Forum Traffic Management Specification Version 4.0 and ITU-T I.317. The effect of policing function is one of the following: ! Pass or Tag (PT). Policing is not enabled for the connection, or the cell is accepted or tagged as conforming by the policing function. ! Discard (D). Policing is enabled for the connection and the nonconforming cell is removed from the cell stream. 17.11.4 Buffer Thresholding The buffer thresholding function uses VC table and buffer state information to manage congestion. The following thresholds exist: ! CLP1 discard threshold ! CLP0+1 discard threshold ! Partial packet discard (PPD) (triggered by either of the above thresholds) ! Early packet discard (EPD) threshold ! Selective EFCI (SEFCI) marking threshold ! Ingress fabric backpressure (IFBP) alarm threshold The effect of the thresholding function is one of the following: ! Pass (P). The cell is admitted to the buffer. ! Discard (D). The cell is not admitted and is removed from the cell stream. ! Notice (N). The cell was discarded upstream but packet discard state in the VC table must be updated. 17.11.5 Egress—APC VC Queueing Structure Cells that arrive and are not dropped by the APC are stored in memory while they await transmission. The stored cells are organized into FIFO linked lists called VC queues. There is one VC queue for each VC. Empty memory locations available for cell storage are contained in the free cell queue. The free cell queue is organized as a LIFO stack. All memory that is configured for cell storage belongs either to a VC queue or the free cell queue. A logical element of a VC queue contains cell data and a pointer to the next element off the queue. An element of the free cell queue contains a pointer to the next cell but no cell data. For each VC queue, a HeadCellPtr contains a pointer to the head element of the queue and a TailCellPtr points to the tail element. The HeadCellPtr and TailCellPtr are stored in the VC table entry of each VC. A FreeCellPtr contains a pointer to the head of the free cell queue. Operations applied to VC queues are initialize, enqueue, and dequeue. An active VC is one whose VC queue is not empty, i.e., it contains cells that are awaiting service. A VC session begins when the VC queue goes from empty to nonempty and ends when the VC queue returns to empty. To summarize, when a cell arrives at ingress, after VC table look-up, and if the cell is admissible by the policing on that VC, it is passed for enqueuing to the VCQ structure that is part of the associated VC table entry, and, if necessary, that VC is then scheduled in the flow queue for the associated traffic class. 98 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 17 ATM Port Controller (APC) Block (continued) 17.12 Connection Management 17.12.1 Connection Admission Control 17.12.1.1 CBR A CBR connection is characterized by the traffic descriptor (PCR) and the QoS requirements CLR and CDV. Generally, the CDV constraint dominates. This can be transformed to a queue length constraint as follows: a simple CBR effective bandwidth that is PCR of the connection. 17.12.1.2 rt-VBR The rt-VBR connections are characterized from their traffic descriptor (PCR, SCR, MBS) and QoS objectives CLR and CDV. The computation of the effective bandwidth for the rt-VBR and nrt-VBR connections is based on the theory of effective bandwidth of leaky-bucket regulated and extremal periodic on-off sources. 17.12.1.3 nrt-VBR The nrt-VBR is treated like rt-VBR, with the exception that there is no CDV QoS metric defined. 17.12.1.4 ABR The ABR connections provide guarantees only for MCR (minimum cell rate). ABR sources will adjust their rates based on feedback they obtain from the ABR flow control algorithm. 17.12.1.5 UBR UBR connections do not provide any QoS guarantees. Thus, there is no need for any bandwidth or buffer allocation for a UBR connection. Agere Systems Inc. Agere Systems - Proprietary 99 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block 18.1 Introduction The AAL engine block provides ATM adaptation layer (AAL) services as listed in Section 18.2. The AAL engine also provides enhanced AAL-derived services, as well as QoS functions, including latency and bandwidth guarantees. Figure 35 shows the interfaces to the AAL engine. Three interfaces are provided for user data transport. The NIF is a UTOPIA 2 slave interface to the network. The AAL engine implements a single Rx slave PHY and four Tx PHYs. The system interface (SIF) is a UTOPIA 2+ master interface to the system. This interface is configurable either as a standard UTOPIA 2 or as a UTOPIA-derived packet interface. The AAL engine implements all 31 Rx and Tx master PHYs. Finally, the host interface provides an interface to the AAL engine for command and control as provided by the EDC block. The host interface can also be used to extract/insert packets from/to the AAL engine for adaptation. In this respect, the host operates as an additional PHY. A fourth interface—the external statistics interface (ESI)—is provided for statistics reporting, and may be utilized for tariffing applications. Note that Figure 35 does not detail flows through the data path; this is done later. DATA PATH EG R E S S UT2 TX UT2/UT2+ RX 4 31 SYSTEM INTERFACE (SIF) NETWORK INTERFACE (NIF) NETWORK MASTER SLAVE UT2 RX 1 SYSTEM SLAVE MASTER UT2/UT2+ TX IN G R E S S 31 HOST INTERFACE HOST EXTERNAL STATISTICS INTERFACE (ESI) 1641 (F) R.01 Figure 35. AAL Engine Block Diagram 18.2 Features ! Adaptation support for AAL5, AAL2 (including SSSAR and SSTED). ! VC switching support for ATM; CID switching support for AAL2; VC MUX-deMUX support for AAL2. ! Support for up to 4064 unidirectional connections1 or 2032 bidirectional connections. ! Reassembly service support for up to 64 AAL2 VCs per direction2 up to a maximum of 124 VCs for both directions. 1. See Connection: Section 18.3 on page 101. 2. Reassembly services are limited to flows sourced at the NIF (or at the SIF when SIF is in cell mode). See Section 18.4.7 on page 109, Section 18.4.8 on page 110, and Section 18.4.19 on page 117. 100 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) ! Segmentation service support for up to 2032 layer-2 bidirectional VCs into as many as 124 destination AAL2 VCs1. ! Reassembly service support for multiple AAL5 VCs2. ! Segmentation service support for multiple AAL5 ingress VCs3. ! Support for insertion/extraction of flows from/to host from/to any service4. ! Support for QoS across four classes of service within each AAL2 formatted ATM connection. ! Bidirectional 155 Mbits/s average throughput. ! Latency and discard guarantees. 18.3 Definitions ! SAR: the AAL engine is also called the SAR. SAR and AAL engine will be used interchangeably in this document. ! Connection: a unidirectional stream of data identified with a VCI. ! Channel: a bidirectional stream of data identified with an AAL2 CID. ! Flow: a unidirectional stream of data as configured in the SAR, at its lowest level of abstraction. For instance, if the SAR is configured to demultiplex an AAL2 VC into its constituent CPS packets, the CIDs within the AAL2 VC are the flows; on the other hand, if the SAR is configured to route the AAL2 VC as ATM with no demultiplexing, the VC itself is the flow. The SAR provides for up to 4094 flows. Internally, there are 4096 possible flows; however, two flows are set aside for SAR internal configuration. A bidirectional stream of data would consume two flows. ! ICID: internal connection identifier is a unique number (0—4095) that identifies a flow. Note that ICID 0 and 1 are used for configuration purposes. ! Ingress: (loosely) the direction from terminal to TAAD08JU2 (SIF to NIF). See Section 18.4.4 on page 107. ! Egress: (loosely) the direction from network to terminal (NIF to SIF). See Section 18.4.4. ! Enqueue: the direction (for either an ingress or egress flow) from flow source to deposit into the SQASE shared memory (SM). See Section 18.4.4. ! Dequeue: the direction (for either an ingress or egress flow) from SQASE SM to flow destination. See Section 18.4.4. ! Enqueue block, dequeue block, adaptation block: See Section 18.4.3 on page 106. ! Rx: same as ingress. Typically used when describing the physical interface (SIF or NIF). The usage of Rx is consistent with its definition in the ATM Forum Technical Committee, UTOPIA Level 2, Version 1.0, AF-PHY0039.000. ! Tx: same as egress. Typically used when describing the physical interface (SIF or NIF). The usage of Tx is consistent with its definition in the ATM Forum Technical Committee, UTOPIA Level 2, Version 1.0, AF-PHY0039.000. 1. Segmentation services are limited to flows sourced at the SIF (when SIF is in packet mode) or at the host. See Section 18.4.7 on page 109, Section 18.4.8 on page 110, and Section 18.4.19 on page 117. 2. Roughly limited by the maximum number of flows across all services, 4k unidirectional or 2k bidirectional. See Section 18.5 on page 122. Reassembly services are limited to flows sourced at the NIF, or at the SIF (when SIF is in cell mode). See Section 18.4.7, Section 18.4.8, and Section 18.4.19. 3. Roughly limited by the maximum number of flows across all services, 4k unidirectional or 2k bidirectional. See Section 18.5 on page 122. Segmentation services are limited to flows sourced at the SIF (when SIF is in packet mode) or at the host. See Section 18.4.7, Section 18.4.8, and Section 18.4.19. 4. Roughly limited by the maximum number of flows across all services, 4k unidirectional or 2k bidirectional. See Section 18.5. Agere Systems Inc. Agere Systems - Proprietary 101 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) ! Provisioning: configuration of SAR parameters for allocation of buffer space, routing for Il2Qs, etc., as distinguished from configuration for specific flows. ! User data: (loosely) any data provisioned by the user for adaptation/transport through the AAL engine via any of the ports, as distinguished from, for example, control data or ESI messages. ! Management data: (loosely) any user data that is inserted from or extracted to the host. ! Layer 2: packets arriving as NPAAL or HPF contain a header VCI. This VCI in combination with the source PHY is termed the layer-2 address. ! Subpacket: a 12-octet unit of data, plus descriptor. This is the basic unit of data manipulated by SQASE. ! Message mode: when a flow is configured to be in message mode, an entire SDU is enqueued from the source into the SAR’s subpacket buffer before it can be dequeued to the destination. ! Streaming mode: when a flow is configured to be in streaming mode, an entire IDU is enqueued from the source into the SAR’s subpacket buffer before it can be dequeued to the destination. ! PPD: partial packet discard. When a flow is configured to be in streaming mode, it may happen that the head portion of an SDU is dequeued to its destination, whereas the tail is dropped. This is called partial packet discard (PPD). 18.4 Architecture To correctly provision and configure the SAR, an understanding of the SAR subblock architecture is required. 18.4.1 Datapath Flows Figure 36 through Figure 40 build on Figure 35 to illustrate the various datapath flows for which the AAL engine (SAR) may be configured. DATA PATH EG R E S S UT2 TX NIF NETWORK UT2 RX UT2/UT2+ TX SIF INGRESS SYSTEM UT2/UT2+ RX 1642 (F) R.01 Figure 36. SIF-to-NIF, NIF-to-SIF The bold arrow from SIF to NIF in Figure 36 indicates a single ingress flow. The arrow from NIF to SIF indicates an egress flow. 102 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) DATA PATH EG R E S S UT2 TX UT2/UT2+ TX SIF NIF NETWORK UT2 RX SYSTEM UT2/UT2+ RX IN G R E S S 1643 (F) R.01 Figure 37. SIF Loopback, NIF Loopback The arrow from SIF to SIF in Figure 37 indicates a SIF loopback flow. The arrow from NIF to NIF indicates an NIF loopback flow. DATA PATH EG R E S S UT2 TX NETWORK UT2 RX SIF NIF SYSTEM IN G R E S S 1644 (F) R.01 Figure 38. NIF Adaptation Loopback The arrow in Figure 38 indicates adaptation loopback. Adaptation loopback is a service whereby, for instance, an AAL5 service at NIF Tx is readapted into AAL2-SSTED at the NIF Rx. This service is available at both the NIF and the SIF, and requires two flows; Figure 38 thus illustrates two flows: one ingress and one egress. Agere Systems Inc. Agere Systems - Proprietary 103 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) DATA PATH EG R E S S NETWORK UT2/UT2+ TX SYSTEM SIF NIF I NG R E S S UT2/UT2+ RX HOST INTERFACE HOST 1646 (F) R.01 Figure 39. Host-to-SIF, SIF-to-Host The arrow from SIF to host in Figure 39 indicates a single ingress flow; in other words, this flow is extracted to the host. The arrow from host to SIF indicates an egress flow; in other words, this flow is inserted from the host. DATA PATH EG R E S S UT2 TX NETWORK UT2 RX SIF NIF SYSTEM IN G R E S S HOST INTERFACE HOST 1645 (F) R.01 Figure 40. Host-to-NIF, NIF-to-Host The arrow from NIF to host in Figure 40 indicates a single egress flow. Again, this is an extracted flow. The arrow from host to NIF indicates an ingress flow. Again, this is an inserted flow. 104 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.2 Subblock Architecture Datapath flows are summarized in Figure 41. Now the SAR subblocks are shown. Subblocks are labeled with acronyms, which are described in Section 18.4.3. (D) (C) (B) (A) ICA ISIA SM (E) MEMI NIF SIF (H) (A) (B) (C) (D) SQASE ESIA ECA (G) PI (F) STAT 1418 (F) Figure 41. SAR Subblock Diagram Agere Systems Inc. Agere Systems - Proprietary 105 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.3 Subblock Definition ! SIF is the system interface subblock. SIF implements the UTOPIA 2 and UTOPIA 2+ master for all 31 Rx and Tx PHYs. ! NIF is the network interface subblock. NIF implements the UTOPIA 2 slave for a single Rx PHY and up to 4 Tx PHYs. ! ISIA is the ingress system interface adaptation subblock. ISIA performs the address translation and adaptation functions on data arriving from the SIF, PI, or ESIA. ISIA enqueues this data as subpackets to SQASE. ISIA is an enqueue block. ! ECA is the egress cell adaptation subblock. ECA performs the address translation and adaptation functions on data arriving from the NIF or PI. ECA enqueues this data as subpackets to SQASE. ECA is an enqueue block. ! SQASE is the subpacket queueing and scheduling engine. SQASE buffers subpackets from the enqueue blocks in the shared memory (SM), a 18k x 128 bit memory, and schedules subpackets for dequeue to the dequeue blocks. ! ICA is the ingress cell adaptation subblock. ICA gets subpackets dequeued from SQASE and sends data to NIF or PI. ICA is a dequeue block. ! ESIA is the egress system interface adaptation subblock. ESIA gets subpackets dequeued from SQASE and sends data to SIF, PI, or ISIA. ESIA is a dequeue block. ! PI is the processor interface. PI provides an interface between each of the subblocks and the TAAD08JU2 EDC for control and host user data purposes. ! MEMI is the adaptation blocks’ shared memory and interface. MEMI contains a 9k x 64 bit memory (MEMI-SM) used by ISIA, ECA, ICA, and ESIA for LUTs and state variables. ! STAT is the statistics manager. STAT accumulates exception information in internal memories and collects and writes out ESI messages from ISIA, ECA, ICA, and ESIA to the ESI interface. 106 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.4 Subblock Flows Figure 41 illustrates the fundamental behavior of the SAR, as follows: 1. Cells or packets are sourced at an interface subblock (SIF or NIF) (A). The interface subblock checks and removes the cell/packet HEC and forwards the cell/packet to the enqueue block (ISIA or ECA). 2. The enqueue block extracts the cell/packet header (B). The enqueue block then uses the header information to perform a series of look-ups in tables set up in MEMI-SM by the various configuration commands: {PHY, VCI} determines the adaptation and {PHY, VCI, (CID)} determines the service and the unique internal connection identifier (ICID) of the flow. This procedure is called address translation and is overviewed in Section 18.4.5. 3. The enqueue block performs the configured adaptation on the flow data and enqueues the data to the SQASE in a proprietary internal packet format. This internal packet format contains an internal connection identifier (ICID), other information relating to how the data is to be queued, and the packet payload. The packet payload is physically transferred to the SQASE in packet fragments called subpackets, where a subpacket is a 12-byte block of a packet. 4. SQASE queues the subpackets into the shared memory (SQASE-SM) (E). The shared memory permits SQASE to adhere to QoS constraints when queueing and scheduling subpackets. It also allows SAR to handle source and destination burstiness. 5. The SQASE scheduler determines the dequeue of subpackets. Subpackets are formatted according to destination adaptation and dequeued to the dequeue subblock (ICA or ESIA) (C). 6. The dequeue block converts the subpackets into the proper cell/packet format, prepending the cell/packet header according to a look-up in MEMI-SM. 7. The dequeue block sends the constructed cells/packets to the interface block (NIF or SIF) (D). In addition, as shown, data may be sourced (inserted) (F)/destined (extracted) (G) from/to the host via PI. Also, data may be looped back from ESIA to ISIA (H). Note that additional loopback paths are possible within SQASE. For instance, a data flow may be enqueued from ECA to SQASE, which then dequeues that flow to ISIA, forming a NIF loopback path. Agere Systems Inc. Agere Systems - Proprietary 107 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.5 Address Translation The enqueue blocks (ISIA and ECA) each contain a 2048-entry VC table and a 2048-entry Connection Table (ConnTable). Address translation is performed so that incoming data units index to the correct VC and Connection Table entries. The address translation procedures are shown graphically in Figure 42. On the left side of the figure, the enqueue procedure is described. It uses a three-level look-up process. 1. The port table is indexed by the source PHY. The look-up verifies VPI and VCI range (minimum VCI, maximum VCI) and produces a VC base address. 2. The VC base address is added to the normalized source VCI (the source VCI minus the minimum VCI configured for this PHY). The resulting address indexes the VC table. 3. The VC table indicates the AAL type (page 110). It also contains the Conn Base. For non-AAL2 services, the Conn Base is used directly to index the Connection Table. For AAL2-services, the Conn Base is added to the normalized CID (the source CID minus the minimum CID configured for this VC). The resulting address indexes the Connection Table. The Connection Table contains the ICID plus other information relating to how the data is to be queued and scheduled. On the right side of Figure 42, the dequeue address translation procedure is described. The ICID is carried with the queued subpackets through the SQASE. On dequeue, it is used to look up the destination PHY, VPI, VCI, and (for AAL2 services) CID. PORT TABLE VC TABLE DEQUEUE ADAPTATION TABLE CONNECTION TABLE DESTINATION MPHY, VPI, VCI (CID) VC BASE + CONN BASE + ICID ICID PHY NORMALIZED (VCI) NORMALIZED (CID) ENQUEUE ADDRESS TRANSLATION DEQUEUE ADDRESS TRANSLATION 1673 (F) Figure 42. Logical View of the Enqueue (Left) and Dequeue (Right) Address Translation Procedure 108 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.6 Queueing and Scheduling A simplified architecture of the SQASE is shown in Figure 43. The input queueing stage consists of level 0 queues (L0Qs). Incoming data from a particular flow is held in an L0Q until a full SDU/IDU unit has been assembled. Once complete, it is transferred from the input queues to the output queues based on its connection information. Once placed into the output queues, it becomes eligible for dequeue scheduling. The output queues are composed of three stages. The three stages are intralevel 1 queues (IL1Q), level 1 queues (L1Qs) and intralevel 2 queues (IL2Qs). There are four IL1Qs permanently associated with each L1Q. These IL1Qs enable four different classes of scheduling within each L1Q. Each L1Q is configured to route to a single IL2Q. In turn, the IL2Qs are configured to send data to a specific system interface or network interface PHY address. Multiple IL2Qs can be configured to drive a specific PHY. IL1Q L0Q L1Q SOURCE DATA IL2Q DESTINATION DATA INPUT QUEUE ROUTING OUTPUT QUEUES 1674 (F) Figure 43. Simplified Diagram of SQASE Queueing Structure 18.4.7 Modes The SAR has only one true mode configuration. The system interface (SIF) subblock must be configured to support either standard UTOPIA 2 (SIF cell mode) or nonstandard UTOPIA 2+ (SIF packet mode). UTOPIA 2 (SIF cell mode) supports transport of ATM cells. UTOPIA 2+ (SIF packet mode) supports transport of packet ATM (PATM) packets. Agere Systems Inc. Agere Systems - Proprietary 109 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.8 User Data Types (UDT) and AAL Types Discussion of SIF mode leads to a discussion of data types at the interfaces. Figure 44 illustrates data types available at all the interfaces for both SIF cell mode and SIF packet mode. AT THE SIF (CELL MODE) USER DATA TYPE: ATM CELL AAL0 AAL TYPE: AAL2 CPS-AAL0 AAL5 AT THE NIF USER DATA TYPE: ATM CELL AAL TYPE: AAL0 AAL2 CPS-AAL0 AAL5 AT THE SIF (PACKET MODE) USER DATA TYPE: PATM AAL TYPE: NPAAL DATA PATH EGRESS SIF NIF INGRESS HOST INTERFACE AT THE HOST USER DATA TYPE: HPF AAL TYPE: NPAAL 1675 (F) R.01 Figure 44. User Data Types and AAL Types at the Interfaces 110 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) As illustrated in Figure 45, three user data types (UDTs) are provided for transport of user data to/from the SAR: ATM cell, packet ATM (PATM), and host packet format (HPF). These UDTs are visible to the user at the interfaces of the SAR. Five AAL types are provided for SAR provisioning/configuration. The AAL type determines how a sourced data unit of data type is interpreted by SAR. As previously defined, UDT is one of ATM cell, PATM, or HPF. The ATM cell type is further split into AAL types AAL0, AAL2, CPS-AAL0, and AAL5. Conversely, UDTs PATM and HPF both map to a single AAL type: NPAAL. Figure 45 shows the mapping between UDT and AAL type. UDT ATM CELL AAL TYPE AAL0 AAL2 SUBTYPE AAL2 SPAAL2 PATM CPS-AAL0 AAL5 HPF NPAAL 1676 (F) Figure 45. User Data Type (UDT) vs. AAL Type Mapping 18.4.9 UDT: ATM Cell Cells as specified in ITU-T Recommendation I.361, B-ISDN ATM Layer Specification are transported across the NIF UTOPIA 2 interface. Cells are also transported across the SIF UTOPIA 2 interface when that interface is configured as cell mode. (See Section 18.4.7 on page 109.) An ATM cell data stream may be used to carry user data adapted according to the AAL types indicated in Figure 44. The ATM cell header VCI determines the AAL type, according to SAR configuration. Interrogation of the ATM VCI is a component of address translation. (See Section 18.4.5, Address Translation, on page 108 and Section 18.5, Provisioning, on page 122.) The following treatment of ATM cell AAL types may be viewed as a summary of AAL support provided by the SAR at the ATM service access point (SAP). 18.4.10 AAL Type: AAL0 The SAR is configurable to switch the VCI, and to route ATM cells from any source PHY (including the host), to any destination PHY (including the host), with no interrogation of the ATM cell payload. AAL0 cells sourced onto flows configured for transparent service are treated in this manner, regardless of any adaptation the cells may carry. (See Section 18.4.19, Service Types, on page 117.) Other services types are not applicable to AAL0. 18.4.11 AAL Type: AAL2 AAL2 support at the ATM layer is fully compliant with ITU-T Recommendation I.363.2, B-ISDN ATM Adaptation Layer Specification: Type 2 AAL. Agere Systems Inc. Agere Systems - Proprietary 111 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.12 AAL2 Subtype: SPAAL2 (Single-Packet AAL2) Cells generated with the single-packet AAL2 mode are compliant with ITU Recommendation I.363.2. Within the TAAD08JU2 context, it describes an AAL2 style adaptation that will only ever contain a single CPS packet. This is guaranteed by appropriate configuration of the scheduler mode within the SAR. To an AAL2 compliant receiver, the SPAAL2 format will appear as a cell containing a single packet aligned to the front of the cell that is packed to the end of the cell with pad octets. The length of an SPAAL2 packet (header and payload) may be in the range of 4 to 47 octets. Packets may not be multiplexed in the cells. See Figure 46. UUI CPS SDU 2 2 2 2 1 1 1 1 1 1 1 1 1 1 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 LI CID UUI CPS LAYER HEC CPS PACKET HEADER CPS SDU PAD CPS PDU LAYER START FIELD 48 OCTETS CELL HEADER ATM LAYER 3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 GFC VPI VCI PTI LP 1419 (F) R.01 Figure 46. SPAAL2 Data Format 112 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.13 CPS-AAL0 CPS-AAL0 is a nonstandard type used to transfer CPS connections across the ATM layer in a manner uniquely identified by the ATM layer VCI. A CPS-AAL0 VC can contain one, and only one, CID throughout its life. The format of a CPS-AAL0 cell is shown in Figure 47. UUI CPS SDU 2 2 2 2 1 1 1 1 1 1 1 1 1 1 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 CID LI UUI CPS LAYER HEC CPS PACKET HEADER CPS SDU PAD CELL HEADER 48 OCTETS ATM LAYER 3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 GFC VPI VCI PTI LP 1422 (F) R.01 Figure 47. CPS-AAL0 Data Format Note that there is no start field (STF) in the CPS-AAL0 CPS-PDU. As a result, the CPS-SDU can be up to 45 octets in length and still fit into a single ATM-SDU; however, the length of a CPS packet (header and payload) may be in the range of 4 to 48 octets. 18.4.14 AAL Type: AAL5 AAL5 support at the ATM layer is fully compliant with ITU-T Recommendation I.363.5, B-ISDN ATM Adaptation Layer Specification: Type 5 AAL, with the exception of support for Annex E, corrupted data delivery option. Agere Systems Inc. Agere Systems - Proprietary 113 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.15 UDT: Packet ATM (PATM) Packets transported across the SIF UTOPIA 2+ interface when that interface is configured as packet mode (see Section 18.4.7 on page 109) are formatted as PATM. The PATM packet consists of an ATM-like header identifying the connection, a variable-length payload for transporting the payload, and a 1-octet trailer for transporting the UUI. PATM is used to transport packet data to be adapted into one of the AAL services supported by SAR. Note: During segmentation services, the user should allow for the fact that the UUI byte is stripped from the last byte of the packet payload and is incorporated as the user-to-user indication for the segmented frame. Likewise, during reassembly services the UUI gets appended as the last byte of the packet payload. 7 6 5 4 3 2 1 0 UUI PATM HEADER PATM TRAILER PATM PAYLOAD 3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 GFC VPI VCI PTI LP 1647 (F) Figure 48. PATM Format Table 23 describes field usage for the PATM format where it differs from the ATM header field descriptions. Table 23. PATM Fields Field Description PATM Payload 1—65536 octets in length. UUI For adaptation into the following: 114 ! AAL2 CPS or SSSAR: bits 4—0 carry the CPS-UUI or SSSAR-UUI. Bits 7—5 are reserved. ! SSTED: bits 7—0 carry the SSTED-UUI. ! AAL5: bits 7—0 carry the CPCS-UUI. Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.16 UDT: HPF Packets/cells transported across the host interface are formatted as host packet format (HPF). It is possible for the host to extract/insert packets or cells from/to the SAR. Packets may be adapted to any of the AAL services supported by SAR; cells may be transported without adaptation from/to either the NIF or the SIF (cell mode). Note: During segmentation services, the user should allow for the fact that the UUI byte is stripped from the last byte of the packet payload and is incorporated as the user-to-user indication for the segmented frame. Likewise, during reassembly services, the UUI gets appended as the last byte of the packet payload. The HPF does not contain a trailing UUI field in the case of AAL0 service (transparent service). 7 6 5 4 3 2 1 0 UUI HPF USED FOR PACKET TRANSPORT HPF HEADER NPAAL TRAILER HPF PAYLOAD 1 1 1 11 1 1 6 5 4 32 1 0 6 6 6 6 5 5 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 2 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 3 2 1 0 9 8 7 6 5 4 3 2 1 D R VPI VCI PTI LP LEN 0 ICID RSVD 1 2 HPF USED FOR CELL TRANSPORT 48 OCTETS HPF HEADER HPF PAYLOAD 6 6 6 6 5 5 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 2 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 D R VPI VCI PTI LP LEN 1 1 1 1 6 5 4 3 1 1 1 0 RSVD ICID 0 1421 (F) Figure 49. HPF Format Table 24 describes field usage for the HPF format where it differs from the ATM header field descriptions. Table 24. HPF Fields Field D R HPF Payload LEN ICID Description Direction: 0 = ingress; 1 = egress. Reserved. 48 octets for ATM cell transport, 1k—64k octets for packet transport. Length: length of packet payload in octets, including the UUI if present. ICID: internal connection identifier for this packet. Agere Systems Inc. Agere Systems - Proprietary 115 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Packets to the host will have LEN values no larger than the programmed maximum SAR IDU length. Packets from the host must not have LEN values larger than the lesser of 1 Kbyte and programmed maximum SAR IDU length, which is set using the device manager function call npAddConnection()/npAAL2AddChannel(). The maximum SAR IDU length is programmed using a combination of the IDU/SDU table in the setup file utility and the device manager function call npAddConnection/npAAL2AddChannel. The host must partition frames larger than this into multiple HPF packets and indicate frame continuation using the PTI field. Likewise, the SAR issues large frames to the host in the form of multiple HPF packets, wherein the PTI field indicates frame continuation. The HPF format is identical to PATM, except for the extended header shown in Figure 49. 18.4.17 AAL Type: NPAAL (No Particular AAL) NPAAL (loosely, no particular AAL) type simply refers to a type recognized by the SAR for provisioning and configuration purposes, to which data types PATM and HPF map. 18.4.18 Nonuser Data Types: ESI Messages The SAR exports a single nonuser data type for reporting external statistics interface (ESI) messages. The SAR reports enqueue and dequeue operations—both successful and unsuccessful—as statistics messages on the ESI. In general, the enqueue, dequeue, or discard of an entire SDU is reported on the ESI, so that a user may keep track of all traffic put through the SAR. This may be done for a tariffing application or simply for debug. 18.4.18.1 ESI Message Format Messages reported on the ESI are formatted as shown in Table 25. Table 25. ESI Message Format (AALXDATA[15:0]) Word AALXDATA bit 0 15 14 Reserved 1 Reserved 2 Packet 13 12 11 ICID 10 9 8 7 6 5 Violation Code1 4 3 2 1 0 SQ2 DR3 ST4 Length5 1. 1. 2. 3. 4. 5. Violation Code: See Section 18.4.18.2 , ESI Violation Code, on page 117. SQ: SQASE operation. 0 => enqueue. 1 => dequeue. DR: Data Direction. 0 => ingress, 1 => egress. ST: Status. 0 => successful enqueue/dequeue operation. 1 => operation not successful. Packet Length: SDU length of enqueued data unit. The dequeue operation reports PDU length for AAL5 and NPAAL and CPS packet length for AAL2. See page 153 for a functional timing diagram. 116 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.18.2 ESI Violation Code The status and violation code field in the ESI message is coded as shown in Table 26. Possible violation codes for ingress and egress are identical, distinguished by the DR field. Table 26. ESI Violation Codings SQ ST 0 0 0 1 Operation Enqueue Success Enqueue Failure Violation Code Reserved Description Operation Successful. 1 000001 Connection Queue Error. 0000101 Intralevel 1 Queue Error. 0000111 Level 1 Queue Error. 1 000100 Intralevel 2 Queue Error. 0001011 Level 2 Queue Error. 0001101 Enqueue Subpacket Error. 1 101011 Exception #11. See Table 34. 1011001 Exception #12. See Table 34. 1011011 Exception #13. See Table 34. 1 101110 Exception #14. See Table 34. 1011111 Exception #15. See Table 34. 1100001 Exception #16. See Table 34. 1 110001 Exception #17. See Table 34. 1100101 Exception #18. See Table 34. 1110001 Exception #24. See Section 18.4.18.3 and Table 34. 1 111001 Exception #25. See Table 34. 1 0 Dequeue Success Reserved Operation Successful. 1 1 Dequeue Failure Reserved Dequeue Packet Abort 2. 1. Reserved for silicon revisions prior to TAAD08JU2 2K V2.0 (revision register 0000 0011). 2. SIF packet mode only. 18.4.18.3 ESI Packet Length The packet length reported in the ESI message is defined in Note 5 of Table 25. In this case, because the SAR could not allocate an L0Q descriptor, there is no ability to report length information to the ESI. For a frame-based service that fails to get an L0Q, multiple ESI messages are reported per frame. In all other cases, just one message is reported, and the packet length is the SDU length. 18.4.19 Service Types Sourced VCs are configured with an AAL type. Sourced flows are configured with a service type. Table 27 defines permissible service types for the AAL types. Service types are described in the following sections. Agere Systems Inc. Agere Systems - Proprietary 117 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Table 27. AAL Type vs. Service Type Compatibility AAL Type Service Type AAL0 AAL2 CPS-AAL0 AAL5 NPAAL CPS_SERVICE Invalid Valid Valid Invalid Invalid SEG_AAL2_SSSAR_SERVICE Invalid Invalid Invalid Invalid Valid SEG_AAL2_SSTED_SERVICE Invalid Invalid Invalid Invalid Valid SEG_AAL5_SERVICE Invalid Invalid Invalid Invalid Valid TRANSPARENT_SERVICE Valid Invalid Invalid Invalid Valid REASS_AAL2_SSSAR_SERVICE Invalid Valid Valid Invalid Invalid REASS_AAL2_SSTED_SERVICE Invalid Valid Valid Invalid Invalid REASS_AAL5_SERVICE Invalid Invalid Invalid Valid Invalid Figure 50 shows an example of how an HPF packet is read/written from the host interface. In this example, the packet has a length of 62 bytes. WRITES 1ST 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 D R 2ND VPI 8 7 6 5 4 VCI LEN 3 2 1 PTI 0 LP ICD (NOTE) 3RD DATA—byte 1 DATA—byte 2 DATA—byte 3 18TH DATA—byte 61 DATA—byte 62 UUI DATA—byte 4 1941 (F) Figure 50. Transferring an HPF Packet over the Host Interface Example Note: When reading an HPF packet from the host, this field will contain the ICID from the AAL engine. When writing an HPF packet to the host, this field is ignored. 18.4.20 CPS_SERVICE CPS_SERVICE offers a CPS layer segmentation and reassembly process. Management and reserved UUI codepoints are observed, but no attempt is made to perform any SEG-SSCS (segmentation service specific convergence sublayer) related functions. CPS_SERVICE may be selected for processing of received AAL2 cells destined for cell-based interfaces. CPS_SERVICE may also be selected for flows destined for the host. Data units requiring CPS_SERVICE are configured to use static L0Qs (see Section 18.5.4.5 ). 118 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.21 SEG_AAL2_SSSAR_SERVICE SEG_AAL2_SSSAR_SERVICE represents the SSSAR segmentation process defined in ITU-T Recommendation I.366.1, Segmentation and Reassembly Service Specific Convergence Sublayer for the AAL Type 2. Data units requiring the SEG_AAL2_SSSAR_SERVICE may be sourced from a packet-based SIF, the host, or the adaptation loopback path. Connections using SEG_AAL2_SSSAR_SERVICE are configured to use dynamic L0Qs (Section 18.5.4.5 on page 129). Connections using SEG_AAL2_SSSAR_SERVICE may be configured to operate in streaming mode with or without partial packet discard (PPD) behavior. Management or reserved data units received on message mode connections using SEG_AAL2_SSSAR_SERVICE, as defined in ITU-T Recommendation I.366.2, AAL Type 2 Service Specific Convergence Sublayer for Narrow-Band Services, may be redirected to the processor. 18.4.22 SEG_AAL2_SSTED_SERVICE SEG_AAL2_SSTED_SERVICE represents the SSTED segmentation process defined in ITU-T Recommendation I.366.1, Segmentation and Reassembly Service Specific Convergence Sublayer for the AAL Type 2. The SEG_AAL2_SSTED_SERVICE incorporates the SSSAR SSCS. Data units requiring the SEG_AAL2_SSTED_SERVICE may be sourced from a packet-based SIF, the host, or the adaptation loopback path. Connections using SEG_AAL2_SSTED_SERVICE are configured to use dynamic L0Qs (Section 18.5.4.5 on page 129). Connections using SEG_AAL2_SSTED_SERVICE may be configured to operate in streaming mode with or without PPD behavior. Management or reserved data units received on message mode connections using SEG_AAL2_SSTED_SERVICE, as defined in ITU-T Recommendation I.366.2, AAL Type 2 Service Specific Convergence Sublayer for Narrow-Band Services, may be redirected to the processor. 18.4.23 SEG_AAL5_SERVICE SEG_AAL5_SERVICE represents the AAL5 segmentation process defined in ITU-T Recommendation I.363.5, BISDN ATM Adaptation Layer Specification: Type 5 AAL. Data units requiring the SEG_AAL5_SERVICE may be sourced from a packet-based SIF, the host, or the adaptation loopback path. Connections using SEG_AAL5_SERVICE are configured to use dynamic L0Qs (Section 18.5.4.5 on page 129). Connections using SEG_AAL5_SERVICE may be configured to operate in streaming mode with or without PPD behavior. Agere Systems Inc. Agere Systems - Proprietary 119 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.4.24 TRANSPARENT_SERVICE The transparent service provides no adaptation-layer processing. In-band indications received on connections (UUI, PTI) that use this service are propagated but otherwise ignored. No accumulation of multiple data units or segmentation of a data unit occurs. For appropriately sized data units, this service may be used to transfer cells transparently between the NIF and a cell based SIF. The host may inject datagrams destined for a packet-based SIF using this service. Data units requiring TRANSPARENT_SERVICE will typically be configured to use dynamic L0Qs (Section 18.5.4.5 on page 129). 18.4.25 REASS_AAL2_SSSAR_SERVICE REASS_AAL2_SSSAR_SERVICE represents the SSSAR reassembly process defined in ITU-T Recommendation I.366.1, Segmentation and Reassembly Service Specific Convergence Sublayer for the AAL Type 2. Data units requiring the REASS_AAL2_SSSAR_SERVICE may be sourced from a cell-based SIF or the NIF. Connections using REASS_AAL2_SSSAR_SERVICE are configured to use dynamic L0Qs (Section 18.5.4.5 on page 129). Connections using REASS_AAL2_SSSAR_SERVICE may be configured to operate in streaming mode with or without PPD behavior. Management or reserved data units received on message mode connections using REASS_AAL2_SSSAR_SERVICE, as defined in ITU-T Recommendation I.366.2, AAL Type 2 Service Specific Convergence Sublayer for Narrow-Band Services, may be redirected to the processor. 18.4.26 REASS_AAL2_SSTED_SERVICE REASS_AAL2_SSTED_SERVICE represents the SSTED reassembly process defined in ITU-T Recommendation I.366.2, AAL Type 2 Service Specific Convergence Sublayer for Narrow-Band Services. The REASS_AAL2_SSTED_SERVICE incorporates the SSSAR SSCS. Data units requiring the REASS_AAL2_SSTED_SERVICE may be sourced from a cell-based SIF or the NIF. Connections using REASS_AAL2_SSTED_SERVICE are configured to use dynamic L0Qs (Section 18.5.4.5 on page 129). Connections using REASS_AAL2_SSTED_SERVICE may be configured to operate in streaming mode with or without PPD behavior. Management or reserved data units received on message mode connections using REASS_AAL2_SSTED_SERVICE, as defined in ITU-T Recommendation I.366.2, AAL Type 2 Service Specific Convergence Sublayer for Narrow-Band Services may be redirected to the processor. 18.4.27 REASS_AAL5_SERVICE REASS_AAL5_SERVICE represents the AAL5 reassembly process defined in ITU-T Recommendation I.363.5, BISDN ATM Adaptation Layer Specification: Type 5 AAL. Data units requiring the REASS_AAL5_SERVICE may be sourced from a cell-based SIF or the NIF. Connections using REASS_AAL5_SERVICE are configured to use dynamic L0Qs (Section 18.5.4.5 on page 129). Connections using REASS_AAL5_SERVICE may be configured to operate in streaming mode with or without PPD behavior. 120 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Table 28. Transport of Congestion Indication and Loss Priority Source Service AAL0 CPS SSSAR SSTED AAL5 NPAAL AAL0 CPS dst.LP = src.LP dst.PTI = src.PTI — Destination Service SSSAR SSTED — — AAL5 NPAAL — dst.UUI = 0 dst.LP = src.LP dst.PTI = src.PTI — dst.UUI = src.UUI — — — dst.UUI = src.UUI dst.LP = 0 dst.PTI = 1 — — — — — dst.UUI = src.UUI1 dst.LP = 0 dst.PTI = 0/12 — — — — — dst.UUI = src.UUI3 dst.LP = src.LP4 dst.PTI = 0/1/2/35 — — — — — dst.UUI = src.UUI6 dst.LP = src.LP7 dst.PTI = 0/1/2/38 dst.LP = src.LP dst.UUI = dst.UUI = src.UUI dst.UUI = src.UUI dst.UUI = src.UUI dst.UUI = src.UUI dst.PTI = src.PTI src.UUI9 dst.LP = src.LP10 dst.LP = src.LP12 dst.LP = src.LP14 dst.CI = src.CI11 dst.PTI = 0,1,2,313 dst.PTI = 0/1/2/315 1. 2. 3. 4. 5. As derived from the final CPS packet comprising the SSSAR PDU (valid on the final IDU of an SSSAR-PDU only). According to streaming status, note that CI portion of the PTI is always clear. As derived from the SSTED trailer UUI field (valid on the final IDU of an NPAAL-PDU only). As derived from the SSTED trailer CPI LP field (valid on the final IDU of an NPAAL-PDU only). According to streaming status, note that CI portion of the PTI is derived from the SSTED trailer CPI CI field (CI is valid on the final IDU of an NPAAL-PDU only). 6. As derived from the CPCS trailer UUI field (valid on the final IDU of a NPAAL-PDU only). 7. As derived from a rolling OR of the current loss priority and the received loss priority of each cell comprising an AAL5 PDU. 8. According to streaming status, note that CI portion of the PTI is derived from the PTI of the cell immediately prior to a streaming/message mode packet operation. 9. The CPS packet UUI is a truncated version of the NPAAL UUI. 10. As derived from a rolling OR of the current loss priority and the received loss priority of each IDU comprising an NPAAL PDU. 11. As derived from the received congestion indication of the last IDU comprising an NPAAL PDU. 12. As derived from a rolling OR of the current loss priority and the received loss priority of each IDU comprising an NPAAL PDU. 13. According to streaming status, note that CI portion of the PTI is derived from the PTI of the final IDU prior to a streaming/message mode packet operation. 14. As derived from a rolling OR of the current loss priority and the received loss priority of each IDU comprising an NPAAL PDU. 15. According to streaming status, note that CI portion of the PTI is derived from the PTI of the final IDU prior to a streaming/message mode packet operation. Agere Systems Inc. Agere Systems - Proprietary 121 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.5 Provisioning Resources within the SAR must be allocated before flows can be configured. This is termed provisioning. Provisioning is static: it is expected to be done once, before connections are set up. Provisioning is done from the source to the destination—the direction of the datapath. Source provisioning is performed first to allocate resources to reflect an expected mix of AAL types, VC ranges, etc., as they will appear at the source ports. Within the SAR, queueing and scheduling resources are then provisioned to provide desired levels of QoS. Finally, destination provisioning is performed to compose outgoing AAL types, and map flows to outgoing ports, VCs, and CIDs. This data sheet refers to a sourced AAL2 VC, for instance, to describe AAL2 adapted ATM cells at the source interface, as distinct from an AAL2 VC being composed for a destination interface. 18.5.1 Some Notes on Terminology and Command Referencing Provisioning is described in terms of the firmware commands issued to perform it. All commands begin with the NPT string. The firmware commands mentioned in Section 18.5.1 through Section 18.5.4.6 are implemented by the initialization of the AAL in the device manager and the setup file utility. Just as SAR datapaths for ingress and egress are very symmetric, there exist many command pairs: one for ingress, one for egress. Such pairs are occasionally identified with a composite term, e.g., NPT_AAL_NIF(SIF)_TRANSMIT(RECEIVE)_CONFIG instead of NPT_AAL_NIF_TRANSMIT_CONFIG/ NPT_AAL_SIF_RECEIVE_CONFIG. 18.5.2 System Interface The system interface block (SIF) will allow the SAR engine to communicate with the external systems in two different ways: ! UTOPIA cell mode (UT2) ! Enhanced UTOPIA packet mode (UT2+) Both interfaces support a maximum of 31 MPHYs. Upon configuration, all MPHYs work in one and only one selected mode. Packet and cell mode may be run as an 8-bit or 16-bit data bus interface (software selectable). Either of the data bus widths can be run at speeds between 25 MHz and 50 MHz. The SIF is a UTOPIA master. When operating in cell mode, the SIF does not support octet-level handshake. The SIF supports an octet-level flow control in packet mode. 122 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) In egress direction, the SIF throttles data from the SAR on different MPHYs using the DUAV(30:0) (data unit available) signal. In ingress direction, the SAR will use a single DUAV signal to backpressure data from the SIF. 18.5.3 Port Table A port table exists for both the ECA and ISIA. For either block, there is a table entry for each port, where a port is one of the following: a PHY from the interface (SIF or NIF), the management port (the host), or the loopback port (for adaptation loopback—ISIA only). Figure 51 illustrates the port table fields relevant to provisioning. PORT TABLE 0 1 2 PortIndex . . . VCIMinOffset VCIMaxOffset M–1 [NOTE 1] NumAal2VCs VPI Note: For ECA, M = 5. For ISIA, M = 33. 1648 (F) Figure 51. Port Table The port table is indexed by the PortIndex. The NPT_AAL_NIF_TRANSMIT_CONFIG(PortIndex) in each NPT_AAL_NIF_TRANSMIT_CONFIG command word is mapped to an ECA port. The NPT_AAL_SIF_RECEIVE_CONFIG(PortIndex) in each command word is mapped to an ISIA port, as shown in Table 29. Table 29. PortIndex to Enqueue Block Port Mapping PortIndex 0 1 2 3 4 5 6—30 31 32 ECA Port PHY[UtopiaStrtAdd] PHY[UtopiaStrtAdd+1] PHY[UtopiaStrtAdd+2] PHY[UtopiaStrtAdd+3] Management Port NA NA NA NA ISIA Port PHY[0] PHY[1] PHY[2] PHY[3] PHY[4] PHY[5] PHY[6-30] Management Port Loopback Port Note that NPT_AAL_SIF_RECEIVE_CONFIG(PortIndex) = [0—30] is hard-mapped to an ISIA PHY, whereas NPT_AAL_NIF_TRANSMIT_CONFIG(PortIndex) = [0—3] is hard-mapped to a normalized PHY, reflecting the fact that NIF Tx PHYs may begin at any multiple of 4 within [0, 4, 8, 12, 16, 20, 24, 28], as indicated by NPT_AAL_NIF_TRANSMIT_CONFIG(UtopiaStrtAdd). The VPI is set by programming VPI and PortIndex appropriately within each even word of NPT_AAL_NIF(SIF)_TRANSMIT(RECEIVE)_CONFIG. Note that the VPI is fixed per port. Data units sourced with a VPI other than NPT_AAL_NIF(SIF)_TRANSMIT(RECEIVE)_CONFIG(VPI) are discarded with an exception. Agere Systems Inc. Agere Systems - Proprietary 123 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Since the VPI is fixed per port, flow identification at the source is referred to as {port, VCI, (CID)}. The total number of VCs sourced on the PortIndex is VCIMaxOffset-VCIMinOffset+1, and the VCIs begin at VCIMinOffset. NumAal2VCs is the number of AAL2 VCs on the PortIndex, where AAL2 VCs must begin at VCIMinOffset, and must fall within a contiguous range. This is discussed further in Section 18.5.4. 18.5.4 MEMI Shared Memory The SAR contains two large memories, each of which contains multiple resources (tables or state). These memories are the adaptation blocks shared memory (MEMI-SM) and the SQASE shared memory (SQASE-SM). The number of resources in each memory is fixed, but the memory allocation given to the resources is programmable. MEMI-SM is a 9K deep (0x2400 entries) memory. Each entry is 64 bits wide. MEMI-SM contains resources as listed in Table 30. Resources have different widths and are packed into MEMI-SM as efficiently as possible. This is illustrated by the width indicator in Table 30. The impact of packing on the user is explained on a case-by-case basis. Table 30. MEMI-SM Resources SAR Subblock Resource Name Width Indicator1 Description Default #Entries3 Base ISIA VC Table 2 Section 18.5.4.2 on page 125. 0x00 2048 ISIA AAL2 VC Table 2 Section 18.5.4.3 on page 126. 0x400 64 ISIA Connection Table 1 Section 18.5.4.4 on page 127. 0x420 2048 ISIA Level 0 Queue Descriptor 0.5 Section 18.5.4.5 on page 129. 0xC20 128 1 State for sourced AAL2-SSTED VCs. 0xD20 64 Pipe2 ISIA SSTED Trailer ECA VC Table 2 Section 18.5.4.2 on page 125. 0x1560 2048 ECA AAL2 VC Table 2 Section 18.5.4.3 on page 126. 0x1960 64 ECA Connection Table/Descriptor 1 Section 18.5.4.4 on page 127. 0x1980 2048 ECA Level 0 Queue Descriptor 0.5 Section 18.5.4.5 on page 129. 0x2160 128 1 State for sourced AAL2-SSTED VCs. 0x2260 4164 2 Section 18.5.4.6 on page 129. 0xD60 4096 ECA SSTED Trailer Pipe2 ICA & ESIA Dequeue Adaptation Table 1. Indicates for each resource the number of resource entries that fit into each MEMI-SM entry. 0.5 indicates that 2 MEMI-SM entries are required for each resource entry. 2. An SSTED trailer pipe entry is required for each allocated dynamic L0Q. 3. The number of entries is the number of resource indices available within the default allocated region, not the number of memory words. 4. MEMI-SM is 9K deep: 0x2400—0x2260 = 0x1a0 = (decimal) 416. Strictly, the SSTED trailer pipe needs to be no larger than the number of allocated dynamic L0Qs. The command NPT_AAL_ADAPBLK_MEM_ALLC is executed once to provision MEMI-SM. There are a number of constraints regarding MEMI-SM provisioning, as listed below. See Section 18.5.4.2 through Section 18.5.4.6 for explanations. 124 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.5.4.1 MEMI-SM Provisioning Constraints ! NifVcCnt + SifVcCnt may not exceed the overall user flow limit of 4064. ! The AAL2 VCs within a port are constrained to occupy the lowermost VC table addresses. ! Neither NifAal2VcCnt nor SifAal2VcCnt may exceed the per-direction AAL2 VC limit of 64. ! NifVcPlusChanCnt + SifVcPlusChanCnt may not exceed the overall user flow limit of 4064. 18.5.4.2 VC Table The VC table contains parameters relating to sourced VCs. Both AAL2 and non-AAL2 sourced VCs require a VC table entry, where AAL2 VCs are constrained to occupy the lower VC table entries. This table is carved up across the ports as illustrated in Figure 52. VC TABLE 0 1 PORT TABLE 0 * 1 2 NumAal2VCs0 VCIMaxOffset0 – VCIMinOffset0 + 1 * PortIndex M–1 [NOTE 1][NOTE 2] NumAal2VCs2 VCIMaxOffset2 – VCIMinOffset2 + 1 N–1 [NOTE 1] 1649 (F) Figure 52. VC Table Notes: 1. For ECA, N = NPT_AAL_ADAPBLK_MEM_ALLC(NifVcCnt), which in turn, must be set to at least: Σ (NPT_AAL_NIF_TRANSMIT_CONFIG(VCIMaxOffset) – NPT_AAL_NIF_TRANSMIT_CONFIG(VCIMinOffset) + 1) Active Egress Ports 2. For ISIA, N = NPT_AAL_ADAPBLK_MEM_ALLC(SifVcCnt), which in turn, must be set to at least: Σ (NPT_AAL_SIF_RECEIVE_CONFIG(VCIMaxOffset) – NPT_AAL_SIF_RECEIVE_CONFIG(VCIMinOffset) + 1) Active Ingress Ports Agere Systems Inc. Agere Systems - Proprietary 125 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) MEMI-SM can fit two VC table entries per memory word. For example, in the default initialization of resource bases shown in Table 30, the ISIA VC table is given a range of 0x400 words, which provides an allocation of 2048 VCs sourced to ISIA. The provisioning constraints are as follows: ! NifVcCnt + SifVcCnt may not exceed the overall VC limit of 4064 (4096—32). There are 32 entries reserved for internal use. ! AAL2 VCs are constrained to occupy the lowermost addresses within the VC range. The VC table is discussed further below. 18.5.4.3 AAL2 VC Table The AAL2 VC table contains parameters and state relating to sourced AAL2 VCs. This table is carved up across the ports as illustrated in Figure 53. AAL2 VC TABLE 0 1 PORT TABLE 0 * NumAal2VCs0 1 2 * PortIndex M–1 NumAal2VCs2 N–1 [NOTE 1] 1650 (F) Notes: 1. The AAL2 VC count is set to 64 and cannot be changed. 2. MEMI-SM can fit two VC table entries per memory word. For example, in the default initialization of resource bases shown in Table 30, the ISIA AAL2 VC table is given a range of 0x20 words, which provides an allocation of 64 AAL2 VCs sourced to ISIA. Figure 53. AAL2 VC Table 126 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.5.4.4 Connection Table The Connection Table contains parameters and state relating to sourced flows. For sourced AAL2 VCs, a flow is identified as {Port, VCI, CID}. For sourced non-AAL2 VCs, a flow is identified as {Port, VCI}. The Connection Table is carved up across the port, VCs as illustrated in Figure 54. PortIndex VC TABLE CONNECTION TABLE VCTableBase2 + i 0, i VCTableBase2 + j 0, j VCTableBase2 + p 2, p VCTableBase2 + q 2, q VCTableBase2 + r 2, r 0 2 NumCIDs(0, i) AAL2 VC Table Aal2VCTableBase0 + i 0, i 0 NumCIDs(2, p) 2 Aal2VCTableBase0 + p 2, p Aal2VCTableBase0 + q 2, q NumCIDs(2, q) AAL2 SOURCE: NON-AAL2 SOURCE: N–1 [NOTE 1] 1651 (F) Figure 54. Connection Table Agere Systems Inc. Agere Systems - Proprietary 127 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Notes: 1. For ECA, N = NPT_AAL_ADAPBLK_MEM_ALLC(NifVcPlusChanCnt), which in turn, must be set to at least: Σ( NPT_ADAPTBLK_MEM_ALLC(NifVCCnt) – NPT_ADAPTBLK_MEM_ALLC(NifAal2VcCnt) Active Egress Ports ) + Σ Σ( ) Expected Number of CIDs on VC . Active AAL2 Egress VCs Ports on Port For ISIA, N = NPT_AAL_ADAPBLK_MEM_ALLC(SifVcPlusChanCnt), which in turn, must be set to at least: Σ( NPT_ADAPTBLK_MEM_ALLC(SifVCCnt) – NPT_ADAPTBLK_MEM_ALLC(SifAal2VcCnt) Active Ingress Ports ) + Σ Σ( ) Expected Number of CIDs on VC . Active AAL2 Ingress VCs Ports on Port The Connection Table must be sized for one entry per non-AAL2 VC, plus an entry for every used CID within each AAL2 VCs. NumCIDs indicates a contiguous range of used CIDs. NumCIDs is programmed during connection configuration, not provisioning (and thus is not a NPT_AAL_ADAPBLK_MEM_ALLC command field). However, the user must predict the total expected number of CIDs across all sourced AAL2 VCs before any connection configuration is done. Since AAL2 VCs may be sourced from both SIF and NIF, a prediction must be made for both ingress and egress. Note that the AAL2 allocation within the Connection Table is a contiguous range of entries: one per CID. MEMI-SM can fit one Connection Table per memory word. For example, in the default initialization of resource bases shown in Table 30, the ISIA Connection Table is given a range of 0x800 words, which provides an allocation of 2048 flows sourced to ISIA. Recall, the provisioning constraints are that NifVcPlusChanCnt + SifVcPlusChanCnt may not exceed the overall limit of 4064 (4096—32) flows. There are 32 entries reserved for internal use. The total number of entries in the connection table is divided evenly between channels and non-AAL2 VCs since every bidirectional AAL2 channel—CPSAAL0 flow will require a channel entry and connection entry in the connection table. So a Connection Table entry is needed for each channel and each non-AAL2 VC at a given interface. 128 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.5.4.5 Level 0 Queue Descriptor An overview of the SAR queueing structure is presented in Section 18.4.6 on page 109. Level 0 queueing is the first queueing step, whereby data is obtained from an interface (SIF, NIF, host, or loopback), and sent to the SQASE for buffering in the SQASE shared memory (SQASE-SM). An L0Q is active when the SAR is in the process of accumulating a data unit (for instance, accumulating an PATM packet for segmentation, reassembling a packet from a sequence of ATM-SDUs, etc.). A data unit is the lowest level packet of information associated with a flow/service (a 48-octet cell for AAL0; a CPS packet for AAL2; a PATM packet for PATM; etc.). An L0Q is active during accumulation of a data unit; once the data unit is queued into the SQASE-SM, the L0Q is freed up to be used to accumulate another data unit. L0Qs are a global resource but must also be divided between ingress and egress. The user must specify how many L0Qs are required for the SIF (ingress) and NIF (egress). The general rule that follows is that, for a given interface, the user needs to allocate an L0Q for each simultaneously accumulating service. CPS packets within a sourced AAL2 VC arrive one at a time, so an L0Q is required for each sourced AAL2 VC. The same rule applies for sourced AAL5 VCs. Sourced AAL0 and CPS-AAL0 packets are contained within an ATM cell, so just a single L0Q is required for all such services per direction. If the source user data type (UDT) is PATM or HPF, an L0Q is required for each MPHY. The TAAD08JU2 firmware allocates a static L0Q for each sourced AAL2 VC, and the balance of the L0Qs are dynamically allocated by the SAR to the remaining services. The default allocation of L0Qs is as follows: ! Total L0Qs = 164. ! Reserved L0Qs = 2. ! Egress static L0Qs = 32. ! Ingress static L0Qs = 32. ! Egress dynamic L0Qs = 49. ! Ingress dynamic L0Qs = 49. Dynamic L0Qing can be used to overallocate. The user may configure more non-AAL2 services at the source interface than there are dynamic L0Qs. If the SAR gets sourced data for which there is no L0Q, the data will be discarded. 18.5.4.6 ICID Table Once queued, each user data flow is identified within the SAR with an ICID. The ICID is used to look up destination parameters VPI, VCI, CID, PHY. The ICID table is shared across the destination interfaces (SIF, NIF, host, adaptation loopback). NPT_ADAPTBLK_MEM_ALLC(SifVcPlusChanCnt) + NPT_ADAPTBLK_MEM_ALLC(NifVcPlusChanCnt) must be <= 4064. 18.5.5 SQASE Shared Memory SQASE-SM is an 18K deep (0x4800 entries) memory. Each entry is 128 bits wide. SQASE-SM contains resources as listed in Table 31. Resources have different widths and are packed into SQASE-SM as efficiently as possible. This is illustrated by the width indicator in Table 31. The impact of packing on the user is explained on a case-bycase basis. Each resource in SQASE-SM occupies contiguously incrementing address space starting at a resource base address. Resource base addresses are expressed in memory pages where a SQASE-SM memory page is 64 words, and each word is 128 bits. This allocation is fixed in SQASE_SM. Agere Systems Inc. Agere Systems - Proprietary 129 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Table 31. SQASE-SM Resources Width Indicator1 Default Base2 0x00 SAR Subblock Resource Name SQASE 4 Per-connection queue lengths. SQASE Connection Queue Length Policing Table Policing Table 4 SQASE L0Q Descriptor 1 Maximum queue lengths, referenced for queue length policing. Input queueing stage descriptors. 0x420 SQASE IL1Q Descriptor 0.5 Output queueing stage descriptors. 0xC20 SQASE L1Q Descriptor 0.5 SQASE Subpacket Data Buffer 1 Description 0x400 Output queueing stage descriptors. 0xD20 Internal data storage. 0x1560 1. Indicates, for each resource, the number of resource entries that fit into each SQASE-SM entry. 0.5 indicates that 2 SQASE-SM entries are required for each resource entry. 2. Base address is expressed in SQASE-SM memory pages, single memory page is 64 words, each word is 128 bits. 18.6 Configuration 18.6.1 Connection and Channel Setup The SQASE queueing structure consists of input and output queues. Input queues are used to hold partially constructed data packets. Output queues are used to hold complete data packets pending dequeue. The queueing structure was initially described in Section 18.4.6 and illustrated in Figure 43. This section builds upon this overview to show how connections are set up within this queueing structure. The input queueing stage consists of L0Qs. Each L0Q may hold a single partially constructed IDU. When an IDU is complete, it is transferred from the SQASE input queues to the SQASE output queues where it becomes eligible for dequeue scheduling. After a completed IDU is transferred to the output queues, the previously occupied L0Q is released and may be reused. The selection of L0Q usage and subsequent transferal to output queueing hierarchy is under the control of the adaptation enqueue blocks (ISIA, ECA). The SQASE output queues consist of three stages of hierarchy. The three stages are IL1Qs, L1Qs, and IL2Qs. Each L1Q contains four IL1Qs, which are permanently associated. Each L1Q may be routed via user configuration to a single IL2Q during connection setup. Each IL2Q defines an MPHY/adaptation pairing. Figure 55 illustrates the SQASE queueing hierarchy. The vertical columns indicate levels within the queueing hierarchy; the horizontal flows indicate data paths across the SQASE. The queuing hierarchy is symmetrical between data directions; however, the specific usage of the queue resources varies according to the adaptation type of the transported data. The figure illustrates data flows for the supported adaptation types. Consider the NIF UT2 flows described by the diagram. These are depicted at the lower end of the illustration as horizontal flows from left to right across the queueing structure. All the NIF UT2 flows terminate in the IL2Qs labelled NIF UT2 - MPHY 0x0. All of these data flows will be dequeued from the SAR through the network interface (NIF) on UTOPIA MPHY zero. The text descriptions of these flows (adjacent to the IL2Q column) indicate that there are multiple adaptation types being merged into this single UTOPIA port. 130 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) L0Q IL1Q L1Q AAL2 VCI = 0x0 AAL2 VCI 0 (CID 0x08) AAL2 VCI 1 (CID 0x08) CID = 0x08 AAL2 VCI = 0x1 CPS-AAL0 (VCI 0x20 TO 0x7F) VCI = 0x20 TO 0x3F CPS AAL0 VCI = 0x40 TO 0x7F VCI = 0x80 TO 0xBF AAL0/5 (VCI 0x80 TO 0xFF) SIF UT2—MPHY 0XC IL2Q CID = 0x08 AAL 0/5 CID = 0x08 TO 0x20 CID = 0x08 TO 0x20 VCI = 0x20 TO 0x3F AAL2 VCI = 0x0 AAL2 VCI = 0x1 CPS-AAL0 (VCI 0x20 TO 0x7F) CPS AAL0 AAL0/5 (VCI 0x80 TO 0xFF) VCI = 0x40 TO 0x7F VCI = 0x80 TO 0xBF AAL2 VCI 0 (CID 0x08 TO 0x20) AAL2 VCI 1 (CID 0x08 TO 0x20) SIF UT2—MPHY 0XD VCI = 0xC0 TO 0xFF AAL 0/5 HOST HPF PACKETS HOST HPF PACKETS HPF PACKETS HOST-INGRESS HPF PACKETS HOST-EGRESS VCI = 0xC0 TO 0xFF VCI = 0x80 TO 0xBF VCI = 0xC0 TO 0xFF CPS AAL0 VCI = 0x20 TO 0x3F VCI = 0x40 TO 0x7F AAL 0/5 CID = 0x08 AAL2 VCI = 0x0 CID = 0x08 TO 0x0F CPS-AAL0 VC 0x80 TO 0xFF AAL2 VCI = 0x2 CID = 0x08 TO 0x27 CID = 0x28 TO 0x47 AAL2 VCI = 0x3 CID = 0x08 TO 0x27 CID = 0xF8 TO 0xFF INPUT QUEUEING AAL2 VCI = 0x4 AAL0/5 VCI 0x20 TO 0x7F AAL2 VCI 0 (CID 0x08) AAL2 VCI 1 (CID 0x08 TO 0x0F) AAL2 VCI 2 (CID 0x08 TO 0x0F) AAL2 VCI 3 (CID 0x08 TO 0x47) AAL2 VCI 4 (CID 0x08 TO 0x27 AND 0xF8 TO 0xFF) NIF UT2—MPHY 0x0 AAL2 VCI = 0x1 CID = 0x08 TO 0x0F OUTPUT QUEUEING = INGRESS TRAFFIC = EGRESS TRAFFIC = IDLE RESOURCE 1652 (F) Figure 55. SQASE Queueing Structure Agere Systems Inc. Agere Systems - Proprietary 131 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) These flows are individually explained as follows: ! CPS-AAL0 VC 0x80 to 0xFF. All data carried over these flows is emitted from the SAR as CPS-AAL0-formatted ATM cells. The diagrammed flow shows data destined for VCs in the range 0x80 to 0xFF being split into one of two IL1Qs according to the indicated VCI. There are four IL1Q groups available, but in the indicated example only two are utilized. This is done to permit QoS scheduling between the two VCI groups where each VCI group represents a traffic class such as video, voice, or data. ! AAL0/5 VCI 0x20 to 0x7F. All data carried over these flows is emitted from the SAR as AAL0/5-formatted ATM cells. The diagrammatic flow shows data destined for VCs in the range 0x20 to 0x7F being split into one of two IL1Qs according to the indicated VCI. There are four IL1Q groups available, but in the indicated example only two are utilized. As in the CPS-AAL0 flows described above, the grouping is performed to permit QoS scheduling between the two VCI groups where each VCI group represents a traffic class. ! AAL2 VCI 0, 1, 2. All data carried over these flows is emitted from the SAR as AAL2-formatted ATM cells. This description refers to a set of three AAL2 flows that all route into the same IL2Q. Each of the three VCIs is allocated its own L1Q and will be scheduled with equal fairness. Within each of the three VCIs, data is grouped according to its CPS-layer addressing (channel index (CID)). In the example, each VC carries a single channel and thus requires a single IL1Q. ! AAL2 VCI 3, 4. All data carried over these flows is emitted from the SAR as AAL2-formatted ATM cells. This description refers to a set of two AAL2 flows that both route into the same IL2Q. Like the previous description, each of the two VCIs is allocated its own L1Q and will be scheduled with equal fairness. However, the SAR is able to allocate scheduler bandwidth between the IL2Qs so that a user can control the relative service given to the VCI group 0, 1, 2 vs. the VCI group 3, 4. This is done to permit QoS scheduling between the groups of VCIs where each VCI group represents a traffic class. Each individual flow within each of these VCI’s flows is routed into IL1Qs according to the CID. By grouping the packets according to CID within IL1Qs, the SAR is able to perform QoS scheduling between groups of CIDs within a VC. All of these flow groups terminate within a single IL2Q. The SAR is able to configure the scheduler service given to each IL2Q so that the profile of traffic sourced from each IL2Q is controllable within this single UTOPIA MPHY. The diagram shows additional flow groups for egress and host traffic. These additional examples use the same principles described above with varied configurations of traffic sources. The following sections describe in more detail the capabilities of the SQASE for each adaptation type. 132 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.6.1.1 AAL2 Data Flow (CPS/SSSAR/SSTED) AAL2 flows utilize all queueing levels within the SQASE queueing hierarchy. For an AAL2 flow, the L1Q is analogous to the ATM layer VCI. All AAL2 data queued through each L1Q will be emitted on a single (programmable) VCI. The simplest implementation of an AAL2 datapath routes all incoming CPS-layer data connections into a single IL1Q within the L1Q. All CPS-layer data will then be packed (in order of arrival) into the ATM cell with packet straddling and TimerCU support where required. If a configurable quality of service (QoS) is required at the CPS layer, the incoming CPS-layer connections may be grouped, according to required QoS, into four groups (classes) of connections. These are then mapped to the four IL1Qs permanently associated with the L1Q. This allows the scheduler to perform AAL layer scheduling within the VC by extraction of data packets in parallel from the four IL1Qs. Within each of the IL1Qs, the data will still be presented in the order of arrival. Each L1Q (VCI) is then routed into an IL2Q. In the example illustrated in Figure 55, the egress AAL2 VCIs are routed into either MPHY 0xC or MPHY 0xD. This indicates that, within the MPHY, the AAL2 sources will receive equal ATM layer bandwidth. Where the user requires configurable ATM layer bandwidth, a rudimentary ATM layer scheduling function is provided. L1Qs (ATM VCIs) may be grouped and routed into different IL2Qs per group. This is illustrated in Figure 55 for ingress AAL2 data flow where the 5 AAL2 VCIs are split into two groups that are then routed separately to IL2Qs. By manipulation of the IL2Q data scheduling, the user is able to control the proportion of data sourced from each of the groups of VCs. The restriction is that there are only 124 (128 – 4) L1Qs. Four L1Qs are reserved for internal usage. Each AAL2 VC must have a dedicated L1Q. Thus, the maximum AAL2 VCs for both directions is 124, and there are 62 bidirectional AAL2 VCs. In the AAL2 VC table, either egress or ingress interface cannot have more than 64 AAL2 VCs. So, for example, if one interface has 64 AAL2 VCs, the other interface can only have a maximum of 60 AAL2 VCs. 18.6.1.2 CPS-AAL0 Data Flow CPS-AAL0 flows pack a single CPS packet into each ATM cell. While the outgoing VC and CID within the produced cell may be individually configured, there is only one effective addressing level since there is only one CID carried over each ATM layer VC. This translates to a single layer of queueing required within the SQASE. As previously defined for an AAL2 flow, incoming CPS-layer connections are grouped into four classes and stored in the corresponding IL1Qs. However, since there is no further addressing required, only a single L1Q is utilized within an MPHY. Additional MPHYs carrying CPS-AAL0 will each require an additional L1Q. The L1Q is routed directly into a single IL2Q for each MPHY. 18.6.1.3 AAL0/AAL5 Data Flow Similarly to CPS-AAL0, AAL0/5 requires a single (ATM) layer of addressing, which translates to a single layer of queueing required within the SQASE. As previously defined for an AAL2 flow, incoming AAL0/5 connections are grouped into four classes and stored in the corresponding IL1Qs. However, since there is no further addressing required, only a single L1Q is utilized within an MPHY. Additional MPHYs carrying AAL0/5 will each require an additional L1Q. The L1Q is routed directly into a single IL2Q for each MPHY. 18.6.1.4 HPF Data Flow HPF packets are routed to and from the host processor. These require a single (ATM) layer of addressing, which translates to a single layer of queueing within the SQASE. All data destined for the host is queued into one of four IL1Qs according to its incoming address. These four groups are then scheduled into the host. Four IL1Qs provide support for provisioning QoS onto the data sent to the host, where data sent to the host represents multiple traffic classes. Agere Systems Inc. Agere Systems - Proprietary 133 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.6.2 Configuration for QoS 18.6.2.1 Packet Scheduling Packet scheduling is executed within the SQASE. There are two scheduler functions: the IL1Q scheduler and the IL2Q scheduler. The application of these schedulers varies according to transported adaptation type and is listed in Table 32. 18.6.2.2 IL1Q Scheduler Algorithm The IL1Q scheduler allows configuration of the relative service given to each of the four IL1Qs within each L1Q. The scheduler does not support rate guarantees or explicit bandwidth controls. The operation of the scheduler is as follows. The four IL1Qs within the L1Q are assessed for eligibility. Eligibility is based on the following factors: ! Queue credit counter ! Queue occupancy The queue credit counter maintains a history that reflects the service previously given to each IL1Q. The queue occupancy indicates whether an IL1Q contains data or is empty. The most eligible queue is indicated by the queue that currently shows the lowest queue credit value and also contains data. A packet is extracted from the queue selected as most eligible and dequeued from the SAR. The queue credit counter is updated to reflect the length of the packet that was dequeued, and the scheduling process is repeated. The user may configure a credit weight parameter that is used to moderate the update of the credit counter. This parameter directly controls the proportion of data emitted from each of the four IL1Qs within each L1Q. For example, a set of four IL1Qs with credit weights of 64, 64, 64, 32, respectively, for queues 0, 1, 2, 3 will see an equal quantity of data emitted from each of queues 0, 1, and 2 and twice as much data emitted from queue 3. Note the following: ! The relation between credit weight and service is reciprocal. ! Empty queues do not accrue credit within the scheduler. 18.6.2.3 IL2Q Scheduler Algorithm The IL2Q scheduler allows configuration of the relative services given to each IL2Q (i.e., each adaptation type/PHY combination) within the device. The scheduler does not support rate guarantees or explicit bandwidth controls. The operation of the scheduler is as follows. Each IL2Q within the device is assessed for eligibility. Eligibility is based on the following factors: ! Queue credit counter ! Queue occupancy ! Flow control. 134 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 18 ATM Adaptation Layer (AAL) Block (continued) The queue credit counter maintains a history that reflects the service previously given to each IL2Q. The queue occupancy indicates whether an IL2Q contains data or is empty. The flow control indicates the ability of the destination PHY to accept data. UTOPIA PHYs may flow control the dequeue of data via deassertion of DUAV clav responses, host PHY and adaptation loopback flow control is managed internally. The most eligible queue is indicated by the queue that currently shows the lowest queue credit value and also contains data and is not currently flow controlled by the destination PHY. A unit of service (equivalent size to an ATM PDU) is extracted from the most eligible queue and dequeued from the SAR. The queue credit counter is updated to reflect the data extraction. The user may configure a credit weight parameter that is used to moderate the update of the credit counter. This parameter directly controls the proportion of data emitted from each IL2Q. Note the following: ! The relationship between credit weight and service is reciprocal. ! Empty queues do not accrue credit within the scheduler. ! Specification of the credit weight used exponent mantissa representation to counter loss of accuracy at extremes of range caused by reciprocal relationship. This causes requested values to be aligned to the closest available value within the constraints of the number representation scheme. ! Queues that are eligible for service but flow-controlled by the destination PHY are deemed to surrender that service slot and will experience a lower server rate than requests by the credit weight configuration. ! IL2Q service is logically divided into ingress and egress service groups that will receive a fair split of the overall service. .I Table 32. L1Q and IL2Q Scheduling Adaptation Type IL1Q Scheduler CPS-AAL0 Provision QoS between groups of CPS-AAL0 connections sharing a single common destination PHY. Outgoing CPS-layer and ATMlayer addressing are not affected. AAL0 Provision QoS between groups of AAL0 connections sharing a single common destination PHY. AAL2 Provision QoS between groups of AAL2 connections sharing a single common destination PHY and a single common destination ATM layer address. AAL5 HPF Agere Systems Inc. IL2Q Scheduler Provision service of CPS-AAL0 traffic within a PHY carrying multiple adaptation types. Provision service between multiple MPHYs. Provision service of AAL0 traffic within a PHY carrying multiple adaptation types. Provision service between multiple MPHYs. Provision service of AAL2 traffic within a PHY carrying multiple adaptation types. Provision service between multiple MPHYS. Additional IL2Qs may be added to support provision of QoS between groups of AAL2 formatted ATM layer connections sharing a single common destination PHY. Provision QoS between groups of AAL5 con- Provision service of AAL5 traffic within a PHY nections sharing a single common destinacarrying multiple adaptation types. Provision tion PHY. service between multiple MPHYs. Provision QoS between groups of connections destined for the host PHY. Provision service of HPF traffic to the host. Agere Systems - Proprietary 135 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.6.2.4 Latency Policing TAAD08JU2 provides a mechanism to remove cells that have been stored in the queues longer than a programmed amount of time. This is called latency policing, and it has the following features within TAAD08JU2: ! Two latency policing periods. ! Configurable latency period from 10 µs to 10 s in 10 µs steps. ! Hierarchical selection mechanism to allow per L1Q enable/disable with minimum performance overhead. ! Per IL1Q enable/disable within an enabled L1Q. The TAAD08JU2 SAR supports a traffic discard method that derives eligibility for discard from the age of the transported data (latency policing). No external references, either in-band or out-of-band, are required or supported by this function. The latency policing method is implemented using latency timers. A latency timer may be configured at any time. Two latency timers are instanced within the SQASE subblock. The two timers are identical and interchangeable but are intended to be used to support the following two functional variants of latency policing. These two variants are called latency-sensitive data discard and internal queue housekeeping and are described in the following document sections. 18.6.2.5 Latency-Sensitive Data Discard The SAR transports traffic classes that are sensitive to transmission latency (e.g., transmission of voice packets). User-configurable QoS provisioning supports relative bandwidth controls per traffic class. This mechanism does not include specific delivery latency guarantees, and during periods of high network congestion traffic may experience transmission delays above the maximum permissible for the class. In this circumstance, the correct behavior is to discard aged data that retains no usefulness to the end application. This alleviates congestion within the network. The SAR supports this feature in both egress and ingress data directions. 18.6.2.6 Internal Queue Housekeeping The SAR operation depends on efficient use of its fixed on-chip buffers. Under exceptional circumstances, such as downstream device failure, it is possible for the SAR to be left with a residual fill that is never dequeued. Internal queue housekeeping is supported to allow the SAR to discard this traffic after a configurable period of time. 18.6.2.7 Reference Clock Generation The latency monitor action is triggered by a reference clock signal. The reference clock signal is generated by a twostage divider chain. Stage One Divider Stage one of the timing reference is a clock-compensation circuit that is used to modulate the incoming system clock frequency into a 10 µs reference signal period. The DividerRatio is set during initialization to a value that yields a ReferencePeriod of 10 µs. 1 ReferencePeriod = ---------------------------------------------------------------------- x DividerRatio SystemClockFrequency 136 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Stage Two Divider The clock-compensated 10 µs reference signal period is divided down into the required latency period for the use within the latency timers. This parameter defines the latency policing period irrespective of the number of enabled queues, as follows: LatencyPeriod = StageOneReferencePeriod x StageTwoDividerRatio Table 33. Example Stage-Two Divider Settings ReferencePeriod DividerRatio npAalTimerConfig (nExpirePeriodMultiple) LatencyPeriod 10.000 µs 10:1 10 100 µs 10.000 µs 1000:1 1000 10 ms 10.000 µs 100000:1 100000 1s Note that each of the stage-two dividers has an independent enable. This allows a single latency timer to be operational while the second timer is globally disabled by removing its reference clock input. Further configuration of a disabled latency timer is not required. 18.6.2.8 Latency Timer Enable/Disable Functions Latency policing is performed on the IL1Qs within the SQASE. A three-stage enable hierarchy is used to control which queues are policed. Enabling latency policing on a single IL1Q requires that all of the enable stages within the enable/disable hierarchy above that IL1Q are configured correctly. The enable hierarchy for latency policing is as follows: ! Global Latency Timer Enable. Each of the two latency-timer reference signals is independently enabled. If neither latency-timer reference signal is enabled, no latency policing will be performed within the SAR. ! L1Q Range Configuration. Each of the two latency timers performs latency policing over a defined range of L1Qs. The range must be contiguous and is fully defined by a start and end pointer. Note that inclusion in the defined L1Q range does not necessarily force latency policing onto all constituent IL1Qs because a further stage of the enable hierarchy is still applicable. ! IL1Q Latency Monitor Enable Mask. Each of the L1Qs is configured using a descriptor word stored in the main SQASE memory. A 4-bit field defines a bit mask of enables for each of the IL1Q w.r.t latency policing. Any combination of this mask is legal, including all disabled. (All disabled is applicable where the configuration of the SAR forces inclusion of nonlatency policed queues within the start/end L1Q defined in the second stage of the enabling hierarchy.) 18.6.2.9 Queue Length Policing All queue levels within the SQASE queueing hierarchy are independently policed against programmable maximum length parameters. A policing table housed in the SQASE-SM is used to hold a set of maximum queue lengths that are then referenced by policing operations. Length policing can be programmed on each of the queues referencing the table. This allows configuration of policing for the connection queues IL1Q, L1Q, and IL2Q. Each of these is described in the following document sections. Agere Systems Inc. Agere Systems - Proprietary 137 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.6.2.10 Connection Queue Length Policing Each internal connection within the SAR, identifiable by an internal connection identifier (ICID), may be optionally policed to a user-definable length. The policed length is expressed in octets of subpacket data buffer fill. Attempted violation of the defined maximum length will result in data discard and exception generation. The connection queue length policing table and policing table SQASE-SM resources are used as indicated in Figure 56. The incoming ICID is used to reference into a connection queue descriptor. The referenced connection queue descriptor contains a pointer into the policing table and a current queue length. The pointer into the policing table is used to look up the allowed maximum queue length. The maximum queue length and the actual queue length are compared to evaluate pass or fail for any attempted enqueue operation. If the policing check indicates failure, the enqueue attempt is rejected and the enqueueing data is discarded. The current contents of the queue are unaffected. {A < B}? PASS: FAIL SQASE-SM RESOURCE CONNECTION QUEUE DESCRIPTORS QUEUE LENGTH A THRESHOLD B MAX LENGTH POINTER SQASE-SM RESOURCE POLICING TABLE ICID + #00 #00 + #00 #00 1653 (F) Figure 56. Connection Queue Length Policing 138 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.6.2.11 IL1Q and L1Q Length Policing Each L1Q and its constituent IL1Qs is policed against user-definable lengths. The policed length is expressed in octets of subpacket data buffer fill. Attempted violation of the defined maximum length will result in data discard and exception generation. The IL1Q descriptor and level 1 descriptor SQASE-SM resources are used as indicated in Figure 57. ! IL1Q Policing. The incoming L1Q identifier is used to reference into an IL1Q descriptor. The referenced IL1Q descriptor contains a pointer into the policing table and a current queue length. The pointer into the policing table is used to lookup the allowed maximum queue length. The maximum queue length and the actual queue length are compared to evaluate pass or fail for any attempted enqueue operation. ! L1Q Policing. The incoming L1Q and IL1Q identifiers are used to reference into an L1Q descriptor. The referenced L1Q descriptor contains a pointer into the policing table and a current queue length. The pointer into the policing table is used to look up the allowed maximum queue length. The maximum queue length and the actual queue length are compared to evaluate pass or fail for any attempted enqueue operation. If either of these policing checks indicates failure, the enqueue attempt is rejected and the enqueueing data is discarded. The current contents of the queue are unaffected. {(A < B) AND (C < D)}? PASS: FAIL QUEUE LENGTH A C B SQASE-SM RESOURCE IL1Q DESCRIPTORS QUEUE LENGTH D THRESHOLD THRESHOLD MAX LENGTH POINTER MAX LENGTH POINTER SQASE-SM RESOURCE L1Q DESCRIPTORS IL1QID + L1QID + + #00 #00 + SQASE-SM #00 RESOURCE POLICING TABLE 4xL1QID + #00 1654 (F) Figure 57. IL1Q/L1Q Length Policing Agere Systems Inc. Agere Systems - Proprietary 139 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.6.2.12 IL2Q Length Policing Each IL2Q is policed against user-definable lengths. The policed length is expressed in octets of subpacket data buffer fill. Attempted violation of the defined maximum length will result in data discard and exception generation. Unlike all other policing operations within the SQASE, IL2Q policing does not use the SQASE-SM policing table resource. The maximum values against which the queue lengths are policed are held in the shared IL2Q descriptors. The destination IL2Q (read from the L1Q descriptor) is used to select an IL2Q descriptor from the IL2Q private descriptors. This IL2Q descriptor contains a queue length and a reference to a shared Il2Q descriptor. The reference to the shared Il2Q descriptor is used to read a maximum queue length from an entry in the IL2Q shared descriptor table. This referenced maximum length and the actual queue length are compared to assess pass/fail for the enqueue attempt. If the policing check indicates failure, the enqueue attempt is rejected and the enqueueing data is discarded. The current contents of the queue are unaffected. SQASE-SM RESOURCE IL2Q PRIVATE DESCRIPTORS QUEUE LENGTH A THRESHOLD {A < B}? PASS: FAIL B SHARED POINTER DEST IL2Q + #00 #00 + #00 SQASE-SM RESOURCE IL2Q SHARED DESCRIPTORS #00 1655 (F) Figure 58. IL2Q Length Policing 18.6.3 Configuration for Exceptions Subblocks issue exceptions when an error or exceptional condition is encountered. A 5-bit exception code (EXCCODE) indicates the nature of the exception, and a 15-bit exception parameter (EXCPARAM) identifies the connection. SM prepends a 2-bit exception ID (EXCID) to the EXCCODE to distinguish identical EXCCODEs from different subblocks (an EXCCODE carries a unique meaning, depending on which subblock has issued it). Table 34 shows the subblock and EXCID, along with its EXCCODEs and EXCPARAMs, plus a name and description. Note that the entries given for EXCPARAM in Table 34 show only valid bits; all other bits are reserved and are always 0. Note that enqueue block exceptions, as well as SIF and NIF exceptions, are produced per data unit (SDU or IDU). SQASE exceptions are produced per subpacket. 140 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Table 34. Exceptions Subblock EXCCODE (EXCID) ISIA (01) and ECA (10) EXCPARAM Name 6:0 = MPHY ENQ_IF_PROTOCOL_VIOLATION Interface protocol violation. An unexpected PSI indication has been received. 201 ENQ_DU_RX_DISABLED_PORT Received data unit on disabled port. 211 ENQ_DU_VPI_DISPARITY 19 22 1 23 1 Description Disparate VPI received. ENQ_DU_VCI_OUT_OF_RANGE VC out of range. ENQ_DU_RX_DISABLED_VCI Received data unit on disabled VC. ENQ_AAL2 MAAL-ERROR_0 AAL2 MAAL-ERROR(0). ENQ_AAL2_MAAL_ERROR_1 AAL2 MAAL-ERROR(1). 3 ENQ_AAL2_MAAL_ERROR_2 AAL2 MAAL-ERROR(2). 4 ENQ_AAL2_MAAL_ERROR_3 AAL2 MAAL-ERROR(3). 1 2 5:0 = AAL2_VC_INDEX ENQ_AAL2_MAAL_ERROR_4 AAL2 MAAL-ERROR(4). 6 11:0 = ICID ENQ_AAL2_MAAL_ERROR_5 AAL2 MAAL-ERROR(5). 7 5:0 = AAL2_VC_NDEX ENQ_AAL2_MAAL_ERROR_6 AAL2 MAAL-ERROR(6). ENQ_AAL2_MAAL_ERROR_7 AAL2 MAAL-ERROR(7). ENQ_AAL2_MAAL_ERROR_8 AAL2 MAAL-ERROR(8)/Received data unit on codepoint (UUI/PTI according to service type) flagged for discard. ENQ_AAL2_MAAL_ERROR_9 AAL2 MAAL-ERROR(9). Data received on a disabled connection. 5 8 9 10 11:0 = ICID 29 Reserved 25 ENQ_0_LEN_MUP 261 ENQ_PKT_ENQ_FAILURE 271 ENQ_PP_EPH_FAILURE Partial packet EPH enqueue failure. 281 ENQ_PP_ENQ_FAILURE Enqueue partial packet failure. 30 ENQ_PD_ENQ_FAILURE Enqueue PD failure. 11 ENQ_AAL2_MAAL_ERROR_10 AAL2 MAAL-ERROR(5)/ SSSAR MAAL-ERROR(10)/ SDU length policing violation. 17 ENQ_AAL5_ERR_C Received 0 length AAL5 frame (Err_C). 18 ENQ_AAL5_ERR_D Received illegal pad length in AAL5 frame (Err_D). 16 ENQ_AAL5_ERR_B Received illegal CPI in AAL5 frame (Err_B). 13 ENQ_AAL2_MAAL_ERROR_20 AAL2 MAAL-ERROR(20). 14 ENQ_AAL2_MAAL_ERROR_21 AAL2 MAAL-ERROR(21). Reserved. Received 0 length map unit data primitive (abort frame). Packet enqueue failure. 1. For silicon revisions prior to TAAD08JU2 V2.0 (revision register 0000 0011), these exceptions are subject to errata 6.22 and should not be programmed without masking. Agere Systems Inc. Agere Systems - Proprietary 141 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) Table 34. Exceptions (continued) Subblock EXCCODE (EXCID) ISIA 15 (01) and ECA 12 (10) EXCPARAM 11:0 = ICID SQASE (00) 00X1X2 9:8 = IL1QID 7:0 = L0QID 001XX2 9 11:0 = ICID 10 11 12 24 SIF (11) NIF (11) 25 8:2 = L1QID 1:0 = IL1QID 8:2 = L1QID 23 4:0 = MPHY ENQ_AAL2_MAAL_ERROR_22_A AAL5/SSTED CRC verification error (AAL5: Err_A; SSTED: AL5_ERR_A ENQ_L0Q_LL_EXHAUSTED SQASE_FLERR SQASE_FLDIRERR SQASE_FLQOSERR SQASE_VCONNQ SQASE_VIL1Q SQASE_VL1Q SQASE_VIL2Q SQASE_LATMON SQASE_TCU SIF_TXPKTERR 13 SIF_TXUT2ERR 03 SIF_RXUT2ERR 243 25 263 273 Description MAAL-ERROR(22). ENQ_AAL2_MAAL_ERROR_11_A RAS timer expiry (AAL5 Err_G; AAL2 MAAL-ERROR(11). AL5_ERR_G 24 00XX12 Name 4:0 = MPHY NIF_PRTYERR NIF_CRCERR NIF_TXUT2ERR NIF_RXUT2ERR L0Q linked list exhaustion. Free list global error. Free list direction error. Free list quality of service error. Connection queue length violation IL1Q length violation. L1Q length violation. Intra-Level 2 queue (IL2Q) length violation. Latency monitor discard. TimerCU scheduler action. Packet signaling error on SAR Tx (egress). Protocol error on UTOPIA Tx (egress). Protocol error on UTOPIA Rx (ingress). Parity error on UT2 Tx (egress). ATM HEC error on UT2 Tx (egress). Protocol error on UT2 Tx (egress). Protocol error on UT2 Rx (ingress). 1. For silicon revisions prior to TAAD08JU2 V2.0 (revision register 0000 0011), these exceptions are subject to errata 6.22 and should not be programmed without masking. 2. Free list exceptions are identified by 00 in the leading 2 bits of the exception code. The remaining 3 bits are used to interpret the exception type, which is defined according to the number of bits set in the code. One bit set: the corresponding error (FLERR, FLDIRERR, FLQOSERR) has occurred. Two bits set: the exception code is now viewed as a mask. Each of the set bits corresponds to an exception that has occurred. For example: code 00011 indicates that a direction error has occurred at the same time as a free list QoS error. Three bits set: When all three bits are set in the exception mask, this can mean one of two things: (a) All three free list errors have occurred simultaneously. (b) The enqueue primitive was rejected due to a previous free list exception on the same L0Q. This is an optimization within the SQASE that chooses not to incur cycles enqueueing a subpacket into a queue that is already corrupt due to a prior failure to enqueue a subpacket. Under these circumstances, the exception code indicates that one or more of the three errors occurred but the SQASE is unable to define which. The cause should be assumed to be identical to the previously received exception code on this L0Q. This discarding state is reset on reception of an EPM primitive on this L0Q. 3. Not supported for silicon revisions prior to Rev 3. 142 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.7 Interface Timing Diagrams 18.7.1 SIF UT2/UT2+ Interface The SIF UT2 interface complies with the ATM Forum standard AF-PHY-0039.000. The UT2+ is an extension of the UT2 interface. The interfaces is a master and can operate both in 8-bit and 16-bit modes. Data on this interface is always accompanied by a header as shown in Figure 59. Note that this header is identical to the ATM cell header at the user-network interface. 8-bit MODE 16-bit MODE GFC VPI(7:4) VPI(3:0) VCI(15:12) GFC VCI(11:0) VCI(11:4) VCI(3:0) PTI CLP HEC 6 5 4 3 2 1 0 CLP PTI HEC UNUSED PAYLOAD 1 PAYLOAD 2 PAYLOAD 3 PAYLOAD 4 48 OCTET (24 WORD) CELL IN CELL MODE OR VARIABLE LENGTH PACKET IN PACKET MODE. 48 OCTET CELL IN CELL MODE OR VARIABLE LENGTH PACKET IN PACKET MODE. 7 VCI(15:12) VPI 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 Figure 59. UT2/UT2+ Header at the SIF Interface The interface is initially configured to work in either UT2 or UT2+ mode. The selection of 8-bit or 16-bit data at the interface is also programmable. In the ingress direction, the SIF polls the PHYs to check for data availability. Once it detects that a particular PHY has data (based on the SRXCLAV/SRXPA signal), it selects it for data reception. While it is receiving data on a particular PHY, it continues to poll the other PHYs to check if they have data. The SIF interface is capable of accepting back-to-back cells and packets. A start-of-packet (SRXSOP) signal identifies the start of a cell or a packet. The end of a packet is identified by a end-of-packet (SRXEOP) signal. This signal is valid only in UT2+ mode. The data is also accompanied by a parity signal that can be programmed to be either odd or even. When the interface is in 8-bit UT2+ mode, a size signal identifies whether the end of the packet is contained in bits 7—0 or 15—8. The valid signal is used by the PHY to stall in the middle of data transmission. This signal is applicable only in UT2+ mode. In UT2 mode, the interface expects to receive an entire cell from the selected PHY. There is also an error (SRXERR) signal that is applicable only in UT2+ mode. When this signal is asserted along with the SRXEOP signal, it implies that the received packet is corrupted and needs to be aborted. In egress direction, the SIF polls the PHYs to check if they can accept data. Once it detects that a particular PHY can accept data, based on the STXCLAV/STXPA signal, it selects that PHY and starts to transmit data. The size, STXSOP, STXEOP, STXERR, and parity signals have the same definitions as in the ingress direction. In UT2 mode, the SIF transmits an entire cell without stalling. In UT2+ mode, the SIF can stall in the middle of a packet if it runs out of data. The PHY can stall the SIF too, using the STXSPA signal if it has no room to accept more data. In UT2+ mode, the SIF interface does not permit multiplexing of packets in both ingress and egress directions. Once a packet is started in either direction, it has to be completed before starting a new packet. Therefore, STXSPA and SRXSPA signals should reflect the status of the current selected PHY once the packet transmission/ reception has started. These signals should be driven independent of the STXENB/SRXENB signals once the PHY has been selected. Agere Systems Inc. Agere Systems - Proprietary 143 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) SUCLK STXADDR 1F 00 1F 00 1F 01 1F 02 1F 0000 0001 0002 0003 0004 STXPA STXENB STXDATA STXPRTY STXSOP 1603 (F) Figure 60. Cell Transmission on the SIF Interface 144 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) SUCLK TXADDR 1F 00 1F 00 1F 01 1F 02 1F 0000 0001 0002 0003 0004 STXPA STXENB STXDATA STXPRTY STXSOP STXSPA STXEOP STXERR STXSIZ 1604 (F) Figure 61. Packet Transmission on the SIF Interface with No Stalls Agere Systems Inc. Agere Systems - Proprietary 145 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) SUCLK STXADDR 1F 00 1F 00 1F 01 1F 01 1F 02 0000 0001 0002 0003 0004 0005 1F 03 1F 04 0006 0007 STXPA STXENB STXDATA STXPRTY STXSOP STXSPA STXEOP STXERR STXSIZ 1605 (F) Figure 62. Packet Transmission on the SIF Interface with the PHY Stalling 146 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) SUCLK STXADDR 1F 00 1F 00 1F 01 1F 02 1F 0000 0001 0002 0003 0004 03 1F 00 1F 04 0005 0006 STXPA STXENB STXDATA STXPRTY STXSOP STXSPA STXEOP STXERR STXSIZ 1606 (F) Figure 63. Packet Transmission on the SIF Interface with the Master Stalling Agere Systems Inc. Agere Systems - Proprietary 147 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) SUCLK SRXADDR 1F 00 1F 00 1F 01 1F 02 1F SRXCLAV SRXENB SRXDATA 0000 0001 0002 0003 SRXPRTY SRXSOP 1607 (F) Figure 64. Reception of a Cell on the SIF Interface 148 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) SUCLK SRXADDR 1F 00 1F 00 1F 01 1F 02 1F SRXPA SRXENB SRXDATA 0000 0001 0002 0003 SRXPRTY SRXSOP SRXVAL SRXEOP SRXERR SRXSIZ 1608 (F) Figure 65. Reception of a Packet on the SIF Interface with No Stalls Agere Systems Inc. Agere Systems - Proprietary 149 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) SUCLK SRXADDR 1F 00 1F 00 1F 01 1F 02 1F 03 1F 04 1F 05 SRXPA SRXENB SRXDATA 0000 0001 0002 0003 0004 0005 0006 SRXPRTY SRXSOP SRXVAL SRXEOP SRXERR SRXSIZ 1639 (F) Figure 66. Reception of a Packet on the SIF Interface with the PHY Stalling 150 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) SUCLK SRXADDR 1F 00 1F 00 1F 01 1F 02 1F 00 1F 03 1F 04 SRXPA SRXENB 0000 SRXDATA 0001 0002 0003 0004 0005 SRXPRTY SRXSOP SRXVAL SRXEOP SRXERR SRXSIZ 1640 (F) Figure 67. Reception of a Packet on the SIF Interface with the Master Stalling Agere Systems Inc. Agere Systems - Proprietary 151 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.7.2 Polling Algorithms for UTOPIA 2 and UT2+ Modes 18.7.2.1 Receive Interface Polling The polling sequence is governed by the following rules: ! It starts at the MPHY address, the minimum address configured after being enabled. ! It increments the PHY address by 1 for each successive poll. ! After reaching the maximum configured address, it wraps back to the minimum. These first three rules define the basic PHY address polling sequence. When the cell/packet available signal is detected in response to a poll, the SIF polls one more PHY address and suspends polling. The response to the second polled address is ignored. Immediately after the first PHY address is selected, polling resumes. The next poll address will be the PHY address that was ignored. ! In UTOPIA 2 mode, it skips the active PHY address without delay during the first 27 cycles (53 cycles in 8-bit mode) of a cell transfer (SRXENB asserted). In UT2+ mode, it skips the active PHY address without delay until SRXEOP is received from the active PHY device. ! If, after the 27th cycle (53rd cycle in 8-bit mode) of a cell transfer in UTOPIA 2 mode or after SRXEOP is detected in UT2+ mode, the round-robin PHY address is not the (previously) active PHY address and polling is not stopped, it preempts the round-robin sequence for one poll slot, and polls the (previously) active PHY address. After preemption, the next poll address will be the PHY address that was preempted. Each device should indicate its status by asserting the SRXPA/SRXCLAV signal low (no packets/cells to transfer) or high (packets/cells available) after sampling its address on SRXADRR. A receive cell transfer occurs when the following conditions are all met: ! A physical device is detected with a packet/cell available. ! A transfer is not currently active. ! There is space in the internal FIFO for a new cell. Cell transfers occur as soon as the transfer conditions are met. The SIF selects the device by placing the device address on the address bus SRXADRR, and subsequently bringing down the receive enable signal SRXENB. Selection preempts the polling sequence if it is active. In UTOPIA 2 mode, the SIF then reads 27 words (53 words in 8-bit mode) without pausing, starting from when the receive start of cell/packet signal is asserted. The transfer proceeds in open-loop fashion until all the words are read. In UT2+ mode, the SIF reads words from the selected PHY, pausing if the slave deasserts SRXVAL, until SRXEOP is received from the slave. 152 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 18 ATM Adaptation Layer (AAL) Block (continued) 18.7.2.2 Transmit Interface Polling The polling sequence is governed by the following rules: ! It starts at the minimum configured PHY address after being enabled. ! It increments the PHY address by 1 for each successive poll. ! After reaching the maximum configured PHY address, it wraps back to the minimum. These first three rules define the basic PHY address polling sequence. ! In UTOPIA 2 mode, it skips the active PHY address (without delay) during the first N – 4 cycles (N is 27 and 53 for 16-bit mode and 8-bit mode, respectively) of a cell transfer (STXENB asserted). In UT2+ mode, it skips the active PHY address (without delay) until it sends out an STXEOP. ! If, after the N – 3 cycle of a cell transfer in UTOPIA 2 mode or STXEOP assertion in UT2+ mode, the round-robin PHY address is not the active PHY address, it preempts the round-robin sequence for one poll slot, and polls the active PHY address. After preemption, the next poll address will be the PHY address that was preempted. Each device should indicate its status by asserting the STXPA/STXCLAV signal low (no packets/cells to transfer) or high (packets/cells available) after sampling its address on the transmit address bus STXADDR. If the device responds with packet/cell available asserted, the SIF then transmits to it 27 or 53 words (corresponding to 16-bit mode and 8-bit mode), without pausing in UTOPIA 2 mode. In UT2+ mode, it transmits words, pausing for the slave’s assertion of STXPA/STXCLAV, until STXEOP assertion. The SIF asserts the transmit start-of-cell signal only for the duration of the first word being transmitted. If an MPHY unexpectedly changes status after being polled (for instance, if it responds it is available when polled but unavailable when selected), the SIF sets a status bit indicating a protocol error but continues to send a complete packet/cell. 18.7.3 NIF The NIF UTOPIA 2 interface is fully compliant with ATM Forum Technical Committee, UTOPIA Level 2, Version 1.0, AF-PHY-039.000. NIF implements up to four Tx slave PHYs and a single Rx slave PHY. NIF is a 16-bit device only; 8-bit mode is not supported. NIF implements cell-level handshaking only; octet-level handshaking is not supported. 18.7.4 ESI Figure 68 shows the functional timing diagram for message reporting on the ESI. The beginning of the message is marked by a synchronization pulse RESYNC. Message data REDATA[16:0] arrives on three consecutive clock cycles. Output RECLK is produced so that RESYNC and REDATA may be captured by an off-chip message gatherer on the positive edge of RECLK with plenty of setup time. RECLK runs at the same frequency as the GCLK pin, but is not phase-aligned with GLCK. RECLK RESYNC REDATA MESSAGE-A WORD 0 MESSAGE-A WORD 1 MESSAGE-A WORD 2 MESSAGE-B WORD 0 MESSAGE-B WORD 1 MESSAGE-B WORD 2 MESSAGE-C WORD 0 MESSAGE-C WORD 1 MESSAGE-C WORD 2 1656 (F) Figure 68. ESI Functional Timing Diagram Agere Systems Inc. Agere Systems - Proprietary 153 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 19 Absolute Maximum Ratings Stresses in excess of the absolute maximum ratings can cause permanent damage to the device. These are absolute stress ratings only. Functional operation of the device is not implied at these or any other conditions in excess of those given in the operational sections of the data sheet. Exposure to absolute maximum ratings for extended periods can adversely affect device reliability. Table 35. Absolute Maximum Ratings Parameter Supply Voltage Range Symbol Min Max Unit V VDD33 –0.5 3.6 Supply Voltage Range Maximum Voltage (digital pins) Minimum Voltage (digital pins) with Respect to GND Storage Temperature Range Ambient Temperature VDD15 — — Tstg TA –0.5 1.8 — 5.5* –0.5 — –65 125 –40 85 V V V °C °C * This maximum rating only applies when the device is powered up with VDD. 20 Power Requirements The power used by TAAD08JU2 varies according to the operating mode; therefore, a number of power figures are supplied here for reference. . Table 36. Power Requirements Operating Mode* Internal PHY Mode External PHY Mode Internal PHY Mode External PHY Mode SAR-only Mode GCLK 50 MHz 50 MHz 25 MHz 25 MHz 50 MHz VDD33 Power 1.0 1.0 0.7 0.5 1.0 VDD15 Power 2.0 1.5 1.5 1.0 1.2 Total Power 3.0 2.5 2.2 1.5 2.2 Unit W W W W W * The power numbers listed in this table are worst-case estimates based on characterization in a typical system environment. These values are not 100% production tested. Table 37. Operating Conditions Parameter Power Supply Power Supply 154 Symbol VDD33 VDD15 Min 3.14 1.4 Typ 3.3 1.5 Max 3.47 1.6 Agere Systems - Proprietary Unit V V Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 21 Handling Precautions Although protection circuitry has been designed into this device, proper precautions should be taken to avoid exposure to electrostatic discharge (ESD) during handling and mounting. Agere employs a human-body model (HBM) and charged-device model (CDM) for ESD-susceptibility testing and protection design evaluation. ESD voltage thresholds are dependent on the circuit parameters used in the defined model. No industry-wide standard has been adopted for the CDM. However, a standard HBM (resistance = 1500 Ω, capacitance = 100 pF) is widely used and, therefore, can be used for comparison purposes. The HBM ESD threshold presented here was obtained by using these circuit parameters. Table 38. Handling Precautions Device TAAD08JU2 Voltage 2000 22 Electrical Characteristics The following specifications apply to both versions 2.1 and 3.1 unless otherwise indicated. 22.1 Logical Interface Electrical Characteristics, Version 2.1 TA = 0 °C to 70 °C; VDD = 3.3 V ± 10%; VSS = 0 V. Table 39. Version 2.1 Logic Interface Characteristics Parameter Input Leakage Current: Nonpull-up Pins Pull-up Pins Nonpull-up I/O Pins Pull-down Pins Output Voltage: Low High Output Voltage: Low High AARXCLKN, AARXCLKP, ABRXCLKN, ABRXCLKP Common-mode Input Voltage AARXCLKN, AARXCLKP, ABRXCLKN, ABRXCLKP Differential Input Voltage Input Capacitance Bidirectional/Output Capacitance Load Capacitance High-level Input Voltage Low-level Input Voltage Symbol Test Conditions Min Typ Max Unit IIL IIL IIL IIL VSS < VIN < VDD ± 10% VIN = VSS VSS < VIN < VDD ± 10% VIN = VDD ± 10% — — — — — — — — 10 60 70 300 µA µA µA µA VOL VOH IOL = –10 mA IOH = 10 mA1 — 2.4 — — 0.4 — V V VOL VOH VIC IOL = –6 mA IOH = 6 mA2 — — — 2.4 — 1.4 VDD33/2 0.4 — 1.9 V V V VDIFF — 0.4 0.5 VDD33–VIC V Cin Cin CL VIH VIL — — — — — — — — 2.0 0 2.5 5.0 — — — — — 70 VDD 0.8 pF pF pF V V 1. All outputs or bidirectional pins except switch fabric interface outputs. 2. Switch fabric outputs: AATXDATA, AATXPARITY, AATXSOC, AATXCLKP, AATXCLKN, ABTXDATA, ABTXPARITY, ABTXSOC, ABTXCLKP, ABTXCLKN. Agere Systems Inc. Agere Systems - Proprietary 155 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 22 Electrical Characteristics (continued) 22.2 Logical Interface Electrical Characteristics, Version 3.1 TA = 0 °C to 70 °C; VDD33 = 3.14 V(min) –3.47 V(max); VDD15 = 1.5 V ± 100 mV; VSS = 0 V. Table 40. Version 3.1 Logic Interface Characteristics Parameter Symbol Test Conditions Min Typ Max Unit IIL IIL IIL IIL VSS < VIN < VDD33 ± 10% VIN = VSS VSS < VIN < VDD33 ± 10% VIN = VDD ± 10% — — — — — — — — 10 60 70 300 µA µA µA µA Output Voltage: Low High VOL VOH IOL = –10 mA IOH = 10 mA1 — 2.4 — — 0.4 — V V Output Voltage: Low High VOL VOH IOL = –6 mA IOH = 6 mA2 — 2.4 — — 0.4 — V V Input Capacitance CIN — — 2.5 — pF Bidirectional/Output Capacitance CIN — — 5.0 — pF Load Capacitance CL — — — 70 pF High-level Input Voltage VIH — VDD1 – 0.5 — VDD1 V Low-level Input Voltage VIL — 0 — 1.0 V VIC — 1.4 VDD33/2 1.9 V VDIFF — 0.4 0.5 VDD33 – VIC V Input Leakage Current: Nonpull-up Pins Pull-up Pins Nonpull-up I/O Pins Pull-down Pins Common-mode Input Voltage Differential Input Voltage 1. All outputs or bidirectional pins except switch fabric interface outputs. 2. Switch fabric outputs: AATXDATA, AATXPARITY, AATXSOC, AATXCLKP, AATXCLKN, ABTXDATA, ABTXPARITY, ABTXSOC, ABTXCLKP, ABTXCLKN. 156 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics 23.1 Input Clocks, Versions 2.1 and 3.1 Table 41. Versions 2.1 and 3.1 Main System Clock (GCLK) Timing Specifications Signal Name GCLK Description Min Max Unit GCLK Frequency (nominal) 25 52 MHz GCLK Duty Cycle 40 60 % GCLK Frequency Tolerance — 0.05 % GCLK Rise/Fall Time 0.4 2 ns Table 42. Version 2.1 UTOPIA Input Clocks (UCLK) Timing Specifications Signal Name UCLK Description UCLK Frequency (nominal) UCLK Duty Cycle UCLK Frequency Tolerance UCLK Rise/Fall Time Min 10 40 — 0.4 Max 52 60 0.05 2 Unit MHz % % ns Table 43. Version 3.1 UTOPIA Input Clocks (UCLK_A[B]) Timing Specifications Signal Name UCLK_A[B] Agere Systems Inc. Description Min Max Unit UCLK_A[B] Frequency (nominal) 10 52 MHz UCLK_A[B] Duty Cycle 40 60 % UCLK_A[B] Frequency Tolerance — 0.05 % UCLK_A[B] Rise/Fall Time 0.4 2 ns Agere Systems - Proprietary 157 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) 23.2 Host Interface Timing tCYC HCLK tASET tRHD tAHD HA HCEN VALID HWEN tRDEL tRSET HDATA OUT DATA VALID 0506 (F) Figure 69. Data Read from TAAD08JU21 Table 44. Host Read Timing Characteristics Symbol tASET tAHD tRDEL tRSET tRHD HCLK — tHR tHF 158 Parameter Address Setup Time Address Hold Time Read Data Delay Read Data Setup Time Read Data Hold Time Host Interface Clock Frequency Host Interface Clock Duty Cycle Host Interface Clock Rise Time Host Interface Clock Fall Time Min 3 1 — tCYC – 10 1 — 40 — — Max — — 10 — — 66 60 3 3 Unit ns ns ns ns ns MHz % ns ns Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) tCYC HCLK tASET tWHD tWHD HA HCEN VALID HWEN tWSET HDATA IN DATA VALID 0507 (F) Figure 70. Data Written to TAAD08JU2 Table 45. Host Write Timing Characteristics Symbol tASET tAHD tWSET tWHD HCLK — tHR tHF Parameter Address Setup Time Address Hold Time Write Data Setup Time Write Data Hold Time Host Interface Clock Frequency Host Interface Clock Duty Cycle Host Interface Clock Rise Time Host Interface Clock Fall Time Agere Systems Inc. Min 3 1 3 1 — 40 — — Max — — — — 66 60 3 3 Agere Systems - Proprietary Unit ns ns ns ns MHz % ns ns 159 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) 23.3 Reset Timing . HCLK CLOCKS UNSTABLE GCLK 250 µs GRESET 1671 (F) Figure 71. Power-On Reset HCLK GCLK AT LEAST EIGHT HCLK AND GCLK CYCLES GRESET 1672 (F) Figure 72. Stable Reset 160 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) 23.4 Concentration Highway (CHI) Timing, Versions 2.1 and 3.1 Table 46 and Table 48, with Figure 73 and Figure 74, respectively, illustrate the detailed CHI timing for clock, data, and frame synchronization. Table 46. Version 2.1 CHI Transmit Timing Characteristics Symbol fCK tCK tR tF tS tH tPD * Parameter Clock Frequency* Clock Period Clock Rise Time Clock Fall Time Frame Sync Setup Time Frame Sync Hold Time Clock to CHI Data Delay Min 2.048 61.04 0 0 35 0 — Max 16.384 488.2 3 3 — — 25 Unit MHz ns ns ns ns ns ns Min Max Unit 2.048 61.04 0 0 35 5 — 16.384 488.2 3 3 — — 25 MHz ns ns ns ns ns ns fCK can be either 2.048 MHz, 4.096 MHz, 8.192 MHz, or 16.384 MHz. Table 47. Version 3.1 CHI Transmit Timing Characteristics Symbol fCK tCK tR tF tS tH tPD Parameter Clock Frequency* Clock Period Clock Rise Time Clock Fall Time Frame Sync Setup Time Frame Sync Hold Time Clock to CHI Data Delay * fCK can be either 2.048 MHz, 4.096 MHz, 8.192 MHz, or 16.384 MHz. tCK CTXCLK CLOCK tH tS CTXFS FRAME SYNC tPD CTXDATA DATA 5-9080(F) Figure 73. CHI Transmit I/O Timing Agere Systems Inc. Agere Systems - Proprietary 161 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) Table 48. CHI Receive Timing Characteristics Symbol fCK tCK tR tF tSSYNC tHSYNC tSDATA tHDATA * Parameter Clock Frequency* Clock Period Clock Rise Time Clock Fall Time Frame Sync Setup Time Frame Sync Hold Time CHI Data Setup Time CHI Data Hold Time Min 2.048 61.04 0 0 30 5 25 5 Max 16.384 488.2 3 3 — — — — Unit MHz ns ns ns ns ns ns ns fCK can be either 2.048 MHz, 4.096 MHz, 8.192 MHz, or 16.384 MHz. tCK CRXCLK CLOCK tSSYNC tHSYNC CRXFS FRAME SYNC tSDATA tHDATA CRXDATA DATA 5-9081(F) Figure 74. CHI Receive I/O Timing 23.5 Fabric Interface—Ports A and B, Versions 2.1 and 3.1 Table 49. Version 2.1 Fabric Interface Timing Specifications (Transmit Interface) Signal Name Description AATXCLKP, AATXCLKN, ABTXCLKP, Frequency ABTXCLKN Duty Cycle Frequency Tolerance Rise/Fall Time AATXDATA[7:0], AATXPRTY, AATX- Valid from the Crossing Point of SOC, ABTXDATA[7:0], ABTXPRTY, AA[B]TXCLKP Rising Edge and ABTXSOC AA[B]TXCLKN Falling Edge 162 Agere Systems - Proprietary Min Max Unit 0 40 — 0.2 –1 104 60 1 3 2 MHz % % ns ns Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) Table 50. Version 3.1 Fabric Interface Timing Specifications (Transmit Interface) Signal Name Description Min Max Unit 0 100 MHz 40 60 % — 1 % Rise/Fall Time 0.2 3 ns Valid from the Crossing Point of AA[B]TXCLKP Rising Edge and AA[B]TXCLKN Falling Edge (10 pF spec) 0.75 3 ns Min Max Unit — 35 — 0.4 4 104 65 1 2 — MHz % % ns ns 0 — ns Min Max Unit AATXCLKP, AATXCLKN, ABTXCLKP, Frequency ABTXCLKN Duty Cycle Frequency Tolerance AATXDATA[7:0], AATXPRTY, AATXSOC, ABTXDATA[7:0], ABTXPRTY, ABTXSOC Table 51. Version 2.1 Fabric Interface Timing Specifications (Receive Interface) Signal Name Description Frequency Duty Cycle Frequency Tolerance Rise/Fall Time AARXDATA[7:0], AARX- Input Setup to the Crossing Point of AA[B]RXCLKP Rising Edge and AA[B]RXCLKN PRTY, AARXSOC, ABRXDATA[7:0], ABRX- Falling Edge PRTY, ABRXSOC Input Hold to the Crossing Point of AA[B]RXCLKP Rising Edge and AA[B]RXCLKN Falling Edge AARXCLKP, AARXCLKN, ABRXCLKP, ABRXCLKN Table 52. Version 3.1 Fabric Interface Timing Specifications (Receive Interface) Signal Name Description AARXCLKP, Frequency AARXCLKN, Duty Cycle ABRXCLKP, ABRXCLKN Frequency Tolerance AARXDATA[7:0], AARXPRTY, AARXSOC, ABRXDATA[7:0], ABRXPRTY, ABRXSOC Agere Systems Inc. — 100 MHz 35 65 % — 1 % Rise/Fall Time 0.4 2 ns Input Setup to the Crossing Point of AA[B]RXCLKP Rising Edge and AA[B]RXCLKN Falling Edge 2.5 — ns 0 — ns Input Hold to the Crossing Point of AA[B]RXCLKP Rising Edge and AA[B]RXCLKN Falling Edge Agere Systems - Proprietary 163 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) 23.6 Expansion UTOPIA2 Interface 23.6.1 Receive Interface Timing Table 53. Expansion UTOPIA2 Receive Interface Timing Specifications: 50 MHz Signal Name Description URXDATA[15:0] Input Setup to UCLK Rising Edge URXPRTY, URXSOC, URXCLAV Input Hold from UCLK Rising Edge URXENB, URXADDR[4:0] Valid from UCLK Rising Edge Min Max Unit 4 1 2 — — 10 ns ns ns Min Max Unit 4 1 2 — — 10 ns ns ns 23.6.2 Transmit Interface Timing Table 54. UTOPIA2 Transmit Interface Timing Specifications: 50 MHz Signal Name Description UTXCLAV UTXDATA[15:0] UTXPRTY, UTXSOC, UTXENB, UTXADDR[4:0] Input Setup to UCLK Rising Edge Input Hold from UCLK Rising Edge Valid from UCLK Rising Edge 23.7 Enhanced Services Interface (ESI), Versions 2.1 and 3.1 Table 55. Version 2.1 ESI Interface Timing Specifications Signal Name AECLK AEDATA[15:0], AESYNC Description Frequency (nominal) Duty Cycle Frequency Tolerance Rise/Fall Time Valid from A[S]ECLK Rising Edge Min Max Unit 25 40 — 0.5 2 52 60 0.5 3 10 MHz % % ns ns Table 56. Version 3.1 ESI Interface Timing Specifications (SAR and APC) Signal Name AECLK/RECLK AEDATA[15:0], AESYNC, REDATA[15:0], RESYNC 164 Description Min Max Unit Frequency (nominal) 25 50 MHz Duty Cycle 40 60 % Frequency Tolerance — 0.5 % Rise/Fall Time 0.5 2 ns Valid from A[R]ECLK Falling Edge — 3 ns Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) 23.8 JTAG Table 57. JTAG Timing Specifications Signal Name TCK TDI TDO TRSTN Description TCK Frequency (nominal) TCK Duty Cycle TCK Rise/Fall Time Input Setup to TCK Rising Edge Input Hold from TCK Rising Edge Valid from TCK Falling Edge Asynchronous Reset Min — 40 — 20 20 — — Max 10 60 5 — — 40 — Unit MHz % ns ns ns ns — 23.9 System Interface, Version 2.1 23.9.1 Receive Interface Timing, Version 2.1 Table 58. Version 2.1 Receive Interface Timing Signal Name Description Min Max Unit SRXDATA[15:0], Input Setup to SCLK Rising Edge Input Hold from SCLK Rising SRXPRTY, Edge SRXSOC, SRXCLAV, SRXEOP, SRXERR, SRXSIZ, SRXVAL 4 1 — — ns ns SRXENB, SRXADDR[4:0] 2 10 ns Valid from UCLK Rising Edge 23.9.2 Transmit Interface Timing, Version 2.1 Table 59. Version 2.1 Transmit Interface Timing Signal Name Description Min Max Unit SRXCLV STXERR STXSPA Input Setup to SCLK Rising Edge Input Hold from SCLK Rising Edge 4 1 — — ns ns 2 10 ns STXDATA[15:0] Valid from UCLK Rising Edge STXPRTY STXSOC STXENB STXADDR[4:0], STXEOP STXSIZ Agere Systems Inc. Agere Systems - Proprietary 165 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 23 Timing Characteristics (continued) 23.10 System Interface, Version 3.1 23.10.1 Receive Interface Timing, Version 3.1 (SUCLK Is an Input.) Table 60. Version 3.1 Receive Interface Timing (SUCLK Input) Signal Name SRXDATA[15:0], SRXPRTY, SRXSOC, SRXCLAV SRXEOP, SRXERR, SRXSIZ, SRXVAL SRXENB SRXADDR[4:0] Description Min Max Unit Input Setup to SCLK Rising Edge Input Hold from SCLK Rising Edge 4 1 — — ns ns Valid from SUCLK Rising Edge 2 10 ns 23.10.2 Transmit Interface Timing, Version 3.1 (SUCLK Is an Input.) Table 61. Version 3.1 Transmit Interface Timing (SUCLK Input) Signal Name Description SRXCLV Input Setup to SCLK Rising Edge STXERR Input Hold from SCLK Rising Edge STXSPA STXDATA[15:0], STXPRTY, STXSOC, Valid from SUCLK Rising Edge STXENB, STXADDR[4:0], STXEOP, STXSIZ 166 Agere Systems - Proprietary Min Max Unit 4 1 — — ns ns 2 10 ns Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 24 Referenced Documents 1. Inverse Multiplexing for ATM (IMA) Specification, Version 1.0, July, 1997, AF-PHY-0086.00 2. ANSI T1.231-1997, Digital Hierarchy—Layer 1 In-Service Digital Transmission Performance Monitoring 3. ANSI T1.403, 1995, Bit-Oriented Messages (BOM) 4. ANSI T1.403, 1995, Network-to-Customer Installation—DS1 Metallic Interface; March 21, 1995 5. ANSI T1.403, 1999, Network and Customer Installation Interfaces—DS1 Electrical Interface 6. APC Version 3e Data Sheet 7. ATM Forum Technical Committee, UTOPIA Level 2, Version 1.0, AF-PHY-039.000 8. ETS 300 233 Integrated Services Digital Network (ISDN); Access Digital Section for ISDN Primary Rate, May 1994 9. ETS 300 417-1-1 Transmission and Multiplexing (TM); Generic Functional Requirement for Synchronous Digital Hierarchy (SDH) Equipment; Part 1-1: Generic Processes and Performance; January 1996 10. ITU-T Recommendation G.703, Physical/Electrical Characteristics of Hierarchical Digital Interfaces, 1991 11. ITU-T Recommendation G.732, Characteristics of Primary PCM Multiplex Equipment Operating at 2048 kbits/s; 1993 12. ITU-T Recommendation G.733, Characteristics of Primary PCM Multiplex Equipment Operating at 1544 kbits/s; 1993 13. ITU-T Recommendation G.775, Loss of Signal (LOS) and Alarm Indication Signal (AIS) Defect Detection and Clearance Criteria; November 1994 14. ITU-T Recommendation G.826, Error Performance Parameters and Objectives for International, Constant Bit Rate Digital Paths at or above the Primary Rate, February 1999 15. ITU-T Recommendation G.963, Access Digital Section for ISDN Primary Rate at 1544 kbits/s, March 1993 16. ITU-T Recommendation I.361, B-ISDN ATM Layer Specification 17. ITU-T Recommendation I.363.1, B-ISDN ATM Adaptation Layer Specification: Type 1 AAL 18. ITU-T Recommendation I.363.2, B-ISDN ATM Adaptation Layer Specification: Type 2 AAL 19. ITU-T Recommendation I.363.5, B-ISDN ATM Adaptation Layer Specification: Type 5 AAL 20. ITU-T Recommendation I.366.1, Segmentation and Reassembly Service Specific Convergence Sublayer for the AAL Type 21 21. ITU-T Recommendation I.366.2, AAL Type 2 Service Specific Convergence Sublayer for Narrow-Band Services 22. ITU-T Recommendation O.151, Error Performance Measuring Equipment Operating at the Primary Rate and Above; October, 1992 23. ITU-T Recommendation O.152, Error Performance Measuring Equipment for Bit Rates of 64 kbits/s 24. ITU-T Recommendation O.153, Basic Parameters for the Measurement of Error Performance at Bit Rates Below the Primary Rate; October, 1992 25. ITU-T Recommendation O.161, In-Service Code Violation Monitors for Digital Systems; 1993 26. ITU-T Recommendation O.162, Equipment to Perform In-Service Monitoring on 2048, 8448, 34,368, and N x 64 kbits/s; October, 1992 27. ITU-T Recommendation O.163, Equipment to Perform In-Service Monitoring on 1544 kbit/s Signals; October, 1992 28. TTC Standard JT-G704, Synchronous Frame Structures used at 1554, 6312, 2048, 8488 and 44736 kbits/s Hierarchical Levels; July 1995 1. TAAD08JU2 does not support the Service Specific Assured Data Transmission portion of the Segmentation and Reassembly Service specific Convergence Sublayer for the AAL type 2 included in the ITU-T Recommendation I.366.1. The AAL engine can provide limited functionality in support of an external host-software-based SSADT solution. Agere Systems Inc. Agere Systems - Proprietary 167 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 24 Referenced Documents (continued) 29. UTOPIA Level 2 Specification (AF-PHY-0039.000): — I.371, Traffic Control and Congestion Control in B-ISDN — I.610, B-ISDN Operation and Maintenance Principles and Functions — ITU-T Recommendation G.804, ATM Cell Mapping into Plesiochronous Digital Hierarchy (PDH) — ITU-T Recommendation G.704, Synchronous Frame Structures used at 1554, 6312, 2048, 8488 and 139,264 kbit/s Signals; October, 1992 — ITU-T Recommendation G.706, Frame Alignment and Cyclic Redundancy Check (CRC) Procedures Relating to Basic Frame Structures defined in Recommendation G.704; 1991 168 Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 25 Glossary AAL ATM Adaptation Layer AFE Analog Front End AHB Advanced High-Performance Bus AIS Alarm Indication Signal API Application Programming Interface ARM Advanced RISC Machine ATMF ATM Forum BAPI Board-Level Application Programming Interface BSC Base Station Controller BTS Base Transmission Station CAC Call Admission Control CBR Constant Bit Rate CD Context (or Connection) Descriptor CID Connection Identifier CHI Concentrated Highway Interface CLP Cell Loss Priority COS Class of Service CPE Customer-Provided Equipment CPCS Common Part Convergence Sublayer CPS Common Part Sublayer CRC Cyclic Redundancy Check DAPI Device Application Programming Interface DMAC Direct Memory Access Controller FE Far-End FEBE Far-End Bit Error FERF Far-End Reporting Function GCRA Generic Cell Rate Algorithm GFC Generic Flow Control HEC Header Error Control HOL Head of Line HSD High-Speed Data ICP IMA Control Protocol IMA Inverse Multiplexing over ATM IWF Interworking Function LAPI Low-Level Application Programming Interface LI Length Indication Agere Systems Inc. Agere Systems - Proprietary 169 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 25 Glossary (continued) LIU Line Interface Unit LLC Logical Link Controller LOC Loss of Cell Delineation LUT Look-Up Table MIB Management Information Base MPHY Multiphysical Layer NE Near End OAM Operations, Administration, and Maintenance OAMP Operations, Administration, Maintenance, and Provisioning OSF Offset Flag PDU Protocol Data Unit PH Packet Header PHY Physical Interface PM Performance Monitoring PMD Physical Medium Dependent POS Packet over SONET PTI Payload Type Indicator PVC Permanent Virtual Circuit QoS Quality of Service RAC Remote Access Concentrator RTOS Real-time Operating System SAP Service Access Point SAR Segmentation and Reassembly SCFQ Self-Clocking Fair Queueing SDU Service Data Unit SN Sequence Number SSCS Service-Specific Convergence Sublayer SEG-SSCS Segmentation and Reassembly Service-Specific Convergence Sublayer SSCOP Service-Specific Connection Oriented Protocol SSSAR Service-Specific Segmentation and Reassembly SSTED Service-Specific Transmission Error Detection SSADT Service-Specific Assured Data Transfer STF Start Flag SVC Switched Virtual Circuit TC Transmission Convergence TDM Time-Division Multiplexed 170 Agere Systems - Proprietary Agere Systems Inc. Data Sheet August 18, 2003 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 25 Glossary (continued) UNI User-to-Network Interface UPC Usage Parameter Control UUI User-to-User Indication UT2+ Enhanced UTOPIA or Packet over SONET (POS) UTOPIA Universal Test and Operations Port Interface for ATM VC Virtual Circuit VCI Virtual Channel Identifier VPI Virtual Path Identifier VCC Virtual Channel Connection VToA Voice Traffic over ATM VPC Virtual Path Connection Agere Systems Inc. Agere Systems - Proprietary 171 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 Appendix A. Revision History This version of the data sheet covers both Versions 2.1 and 3.1. Table A-1 lists revisions to this data sheet made since the August 2001 release. Any change to the document since the last release is marked by change bars; the substantive changes are listed below. Table A-1. Revision History No. Page Date Description 1 page 1, page 11 2 page 13 Description of LTXNDATA pin changed to read “Line Interface Transmit Negative Rail Data. When the TAAD08JU2 device operates in a dual-rail line interface mode, this pin is the negative transmit data sent to external LIU. When the TAAD08JU2 device operates in a single-rail line interface mode, this pin outputs an 8 kHz frame sync pulse.” 3 page 14 CRXCLK pin description revised to indicate that the pin does not require a clock to be connected if that pin is not being used for any framer or CHI clocking modes. 4 page 34 “Packet-over-SONET (UT2+) MPHY master port compatible with Agere TDAT and TADM parts” replaced with “Packet-over-SONET (UT2+) MPHY master port.” 5 page 37 “AAL2/13bb.1 SARing” changed to “AAL2/I.366.1 SARing.” 6 page 89 Spatial multicasting support removed from APC block features. 7 Table 39, page 155 8 page 1 9 page 1 ATM Forum AF-PHY-0029.000 and TR-NWT-000170 removed from standards list. ATM Forum AF-TM-0121.000 replaces ATM Forum Traffic Management. 10 page 19 TMODE = 0 changed to TMODE = 1. 11 page 19, page 157 12 page 48 Reference to firmware section removed. 13 page 50 In ARP register, 0X00E* changed to 0X00E. 14 page 85 Change to Table 22, TAAD08JU2 Exceptions to the IMA PICS Proforma, on page 85. 15 page 85 Reference to NPT_PHY_IMA_CONFIG_GROUP firmware command changed to npIMAGroupParameter. 1/25/02 Change from support of 2047 AAL2 CIDs and 2048 high-speed VCs to 2032 AAL2 CIDs and 2032 high-speed VCs. High-level input voltage minimum changed to 2.0; maximum to VDD. Lowlevel input voltage minimum changed to 0; max to 0.8. 4/8/02 Features list expanded to include setup file utility to provision TAAD08JU2 and API reference manual available for device manager software. GCLK maximum frequency tolerance revised to 0.05%. 16 page 100—page 153 ATM Adaptation Layer block chapter revised throughout. 17 page 100 Rx and Tx labels transposed. 18 page 100 “Flow” changed to “connection.” 19 page 101 Bullet changed from “Segmentation service support for up to 2k layer-2 ingress VCs into as many as 62 destination AAL-2 VCs” to “Segmentation service support for up to 2032 layer-2 bidirectional VCs into as many as 124 destination AAL2 VCs.” 20 page 133 Paragraph added regarding restriction to number of L1Qs and resultant maximum numbers of AAL2 VCs. 21 page 138 SQASE-SM replaces SQASE-SRB in Figure 56. 172 Agere Systems - Proprietary Agere Systems Inc. TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 18, 2003 Appendix A. Revision History (continued) Table A-1. Revision History (continued) No. Page 22 page 143—page 153 Date Description 4/8/02 Signal names changed throughout, e.g., PPA changed to SRXCLAV/SRXPA or STXCLAV/STXPA. SOP changed to SRXSOP or STXSOP; err changed to SRXERR; spa changed to STXSPA. TxSPA changed to STXSPA; RxSPA changed to SRXSPA; TxEnb/RxEnb changed to STXENB/SRXENB. SUCLK added as first signal in Figure 60—Figure 67. 23 page 155 Old Recommended Operating Conditions table incorporated into Table 39. 24 page 157 Change to description column in Table 42. 25 page 159 Figure 70 revised to show tCYC begins halfway up rising ramp of HCLK. 26 page 162 Change to maximum rating for Frame Sync Hold Time and CHI Data Hold Time. 27 — Commands and Indications chapters deleted. 28 — References to the J2 mode removed throughout. 29 page 154 Table 36, Power Requirements, updated. 30 page 158—page 159 HADDR changed to HA and HDATAOUT changed to HD in Figure 69 and Figure 70. 31 page 165 Test load information removed from Table 57. 32 page 154 4/16/02 Change to Table 40. 33 page 162 4/22/02 Change to Frame Sync Hold Time and CHI Data Hold Time ratings in Table 49. 34 — 35 page 28 36 page 55, page 59 5/20/02 Chapters 18, Firmware Flows, 19, Embedded Device Controller, and 20, Device Manager deleted. 5/20/02 Chapter 9, Software Components, added. 5/30/02 References to AIS-CI removed from Framer chapter. 37 page 14—page 17 6/12/02 Changes to Tables 4 and 5. 38 page 40, page 154 6/12/02 Change to designations of operating modes. 39 page 17—page 37, page 87. 6/17/02 References to “ASX” changed to “switch fabric.” 40 page 39—page 43 6/17/02 References to “framer mode” changed to “PHY mode.” References to “SAR slave mode” changed to “SAR-only mode.” 41 page 19 6/26/02 Change to description of HCEN signal. 42 page 54 8/19/02 Changed TND to LTXNDATA, TPD to LTXPDATA, TLCK to LTXCLK, RPD to LRXPDATA, RND to LRXNDATA, and RCLK to LRXCLK. 43 page 68 8/19/02 Changed TCHIDATA TO CTXDATA, RCHIDATA to CRXDATA, TCHIFS to CTXFS, and RCHIFS to CRXFS. 44 page 85 8/19/22 Changed the comment regarding LDD.2, R-75. 45 page 53, page 55 46 page 54 47 page 138, page 172 48 page 33 9/11/02 Textual changes: e.g., changed F1 to FT; deleted HDLC features. 9/11/02 Added Section 14.3.1, Line Interface References/Standards; renumbered following sections. 9/11/02 SAR supports latency-sensitive data discard in both ingress and egress directions 11/19/02 Added a list of actions that a Device Manager user needs to perform under certain conditions. 49 page 155—page 166 11/19/02 Version 3.1 electrical and timing characteristics added. 50 page 85 ff. Agere Systems Inc. 4/28/03 Additional exceptions to IMA PICS Proforma in Table 22. 173 TAAD08JU2 T1/E1/J1 ATM Processor, Versions 2.1 and 3.1 Data Sheet August 2003 Appendix A. Revision History (continued) Table A-1. Revision History (continued) No. Page 51 page 52 52 throughout Date Description 8/11/03 Change to description of version register. 8/11/03 Document type changed from Preliminary Data Sheet to Data Sheet. For additional information, contact your Agere Systems Account Manager or the following: INTERNET: http://www.agere.com E-MAIL: [email protected] N. AMERICA: Agere Systems Inc., Lehigh Valley Central Campus, Room 10A-301C, 1110 American Parkway NE, Allentown, PA 18109-9138 1-800-372-2447, FAX 610-712-4106 (In CANADA: 1-800-553-2448, FAX 610-712-4106) ASIA: Agere Systems Hong Kong Ltd., Suites 3201 & 3210-12, 32/F, Tower 2, The Gateway, Harbour City, Kowloon Tel. (852) 3129-2000, FAX (852) 3129-2020 CHINA: (86) 21-5047-1212 (Shanghai), (86) 755-25881122 (Shenzhen) JAPAN: (81) 3-5421-1600 (Tokyo), KOREA: (82) 2-767-1850 (Seoul), SINGAPORE: (65) 6778-8833, TAIWAN: (886) 2-2725-5858 (Taipei) EUROPE: Tel. (44) 1344 296 400 Agere Systems Inc. reserves the right to make changes to the product(s) or information contained herein without notice. No liability is assumed as a result of their use or application. Agere is a registered trademark of Agere Systems Inc. Agere Systems and the Agere Logo are trademarks of Agere Systems Inc. Copyright © 2003 Agere Systems Inc. All Rights Reserved August 2003 DS03-198ATM