1 / 26

Performance Adhoc Group IEEE 802.17 Summary of Progress

This document provides a summary of the progress made by the Performance Adhoc Committee within the IEEE 802.17 Working Group in analyzing and comparing architectural ideas related to performance issues. It outlines the objectives, participation guidelines, and decisions made during the plenary meeting.

olgavasquez
Télécharger la présentation

Performance Adhoc Group IEEE 802.17 Summary of Progress

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Performance Adhoc GroupIEEE 802.17Summary of Progress Khaled Amer IEEE 802.17 Plenary Meeting March 2001

  2. Formation of the Perf Adhoc Group Formed in Jan 2001 as a separate Adhoc committee within 802.17 WG to look into performance issues Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  3. Perf Adhoc Objectives Objectives: • Set parameters, metrics, scenarios to help provide a consistent way of comparing architectural ideas • Analyze simulations results presented in the 802.17 WG meetings • Not chartered to run simulations for the working group Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  4. Perf Adhoc Objectives ... • Agree on common/consistent perf simulation scenarios and metrics: • Traffic Models • Performance Metrics • Test Scenarios • Presentation format/style • Other? Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  5. Perf Adhoc Objectives ... • These would be used to: • Compare the performance characteristics of various proposals • Compare performance characteristics of RPR solutions vs. competing technologies Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  6. Mailing List • Currently using the 802.17 reflector • Put ‘RPR Perf:’ in subject field • May decide to have a separate mailing list for perf discussions in the future (?) Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  7. Participation in Perf Adhoc Group • Anyone welcome to participate • People who can contribute to the perf analysis and perf modeling efforts • People just interested in these topics • People concerned about performance related issues and comparison process • And then … anyone is welcome! Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  8. Goals for this week • Address unresolved and open issues to facilitate running simulations adhering to the methodology developed by the perf adhoc committee • Finalize plans for Phase I of simulations and have participants start simulations Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  9. Progress and Status Report • Presentations and discussions held Monday morning, Monday evening and Wed evening • Closed on many of the general performance metrics and scenarios • Arrived to agreement on details of initial simulation scenarios (Phase I) Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  10. Decisions made on Monday • Document describing methodology and framework (David James + volunteers) • Phase I of simulations • Starts now • Results in May and July • No CoS • No TCP Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  11. Decisions made on Monday ... • Phase II of simulations • Start discussions of Phase II details in May 2001 • Start simulations in July 2001 • Results in Sept and Nov • Add CoS • Add TCP Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  12. Resolved Issues • Use of different modeling tools • Not considered to be a big problem • May even be an advantage (?) Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  13. Resolved Issues • Define the RPR MAC/PHY interface model • Architectural/behavioral abstractions needed for each RPR proposal • Understand the effect of various architectural aspects instead of various vendor implementations • Develop a reference model - Harry Peng Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  14. Resolved Issues for Phase I • Ring overload: continuous supply of packets • Hubbing: • All nodes send and receive to/from a common node on the ring • Random: • Source/dest pairs: uniformly distributed • No need to run the ring underloaded (?) Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  15. Open Issues • Availability of models from various vendors • ETE delay: From where to where? Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  16. Suggestions Made • Comparison of RPR with SONET • Should we model/check the ordering of packets? • In normal mode • In protection mode • Ingress/Egress buffer size: 100 ms • Should we make it 2*RTT? Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  17. Phase I Simulation Scenarios • No upper layer protocol • No staggering of inputs • Test Basic Ring Parameters: • Ring Performance • Congestion Control • Fairness Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  18. Ring Performance • Metrics: • Link utilization under heavy loads • Flow control overhead • Global throughput Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  19. Congestion Control • Metrics: • Throughput in the presence of congestion • Per class • Per node • Per conversation (or flow) Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  20. Fairness • Metrics: • Throughput and end-to-end packet delay and jitter: • Per class • Per node • Per conversation (or flow) • Need scenarios that demonstrate fairness in overload conditions Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  21. Suggested Starting Configuration • Dual Ring • 16 nodes • Ring running under capacity and well as over capacity (overload) • Ring circumference (100Km, 1000Km)? • Ring rate: 10G Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  22. Suggested Starting Applications • Hub application • 50% of the traffic is generated by all nodes and flows to the hub node (let’s say node #15) • 50% of the traffic is generated by the hub node and flows to all the other nodes Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  23. Suggested Starting Applications ... • Random source/destination pairs • Would demonstrate spatial reuse effect better than hub application Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  24. Suggested Traffic Characteristics • Packet size distributions (probabilistic): • Trimodal • (60% 64B, 20% 512B, 20% 1518B) • Quadmodal (?) • (50% 64B, 15% 512B, 15% 1518B, 20% 9K) • Committed rate per node • 30% of ring capacity / # nodes • 60% of ring capacity / # nodes Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  25. Suggested Simulation output results • Throughput • ETE delay • Jitter (99.9th percentile of delays) • Should we use CDF’s? • For all output results: • Show curves and numbers • Per node, per class, per conversation Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

  26. Next steps and Discussions Performance Adhoc Committee IEEE 802.17 - Plenary Meeting - Hilton Head

More Related