Scalable QoS Provision Through Buffer Management Roch Guerin - PowerPoint PPT Presentation

Scalable QoS Provision Through Buffer Management Roch Guerin Sanjay Kamat Vinod Peris Raju Rajan IBM Thomas J. Watson Research Labs

Outline • Differentiated Packet Treatment for QoS • Performance Objectives • Design Space Scheduling and Buffer Management Schemes • Comparing FIFO vs WFQ -- Worst case buffer tradeoffs • Examining tradeoffs with strict buffer partitioning • Examining tradeoffs with buffer sharing • Hybrid Schemes • Conclusions

Differentiated Packet Treatment Buffer Routing and Scheduling Management Policing Classification Discard Discard • Flows (Unit of service guarantee -- varying granularity) • QoS Resources -- Link capacity and Buffer Space • Scalability – Processing time per packet versus number of flows – State size versus number of flows

QoS Performance Objectives . Throughput maximization Flow Isolation & Fair allocation of rate guarantees excess resources Implementation cost Memory Processing

QoS Allocation Schemes . No Buffer Management (FIFO) Scheduling •FIFO Buffer •WFQ Partitioning headroom Buffer Management Buffer •No Buffer Management (FIFO) Sharing •Buffer Partitioning •Buffer Sharing •Fair Excess Capacity distribution E mpty buf f er spaces (Choudhury and Hahne) Excess capacity sharing & fairness

Design Space . FIFO Buffer Fair and flexible excess Management capacity distribution Per-Flow Scheduling FIFO Scheduling (WFQ) Rate Based Per Flow FIFO treatment of Buffer Management excess traffic

Benchmarks . WFQ scheduling & Buffer Management FIFO scheduling & WFQ Scheduling & Buffer Management No Buffer Management FIFO scheduling & No Buffer Management

Rate Guarantees through buffer management alone Reservation Buffer Partition Offered Service Size Load Achieved ρ bits/sec B ρ /R bits Conformant Lossless Non-conformant Losses limited by non-conformance ( σ bits , ρ bits/sec ) σ + B ρ /R bits Conformant Lossless Non-conformant Losses limited by non-conformance Partitioned Buffer Buffer Size B bits Link Capacity R bits/sec FIFO Scheduler

Buffer Partitioning w. FIFO vs. WFQ ( Worst case comparison ) R õ Σρ i Bandwidth availability WFQ Admissibility Checks B õ Σσ i Buffer availability R õ Σρ i Bandwidth availability FIFO Admissibility B õ Σσ i + ( Σρ i ) B/R Buffer availability Checks Equivalently B õ Σσ i /(1-utilization)

Experimental Setup Link On-off source Regulator Buffer Manager Scheduler 500KB to 5MB 48Mbps Flow Token Token Mean Peak Number bucket Rate Rate Rate 0 50 kB 2Mbps 2Mbps 16Mbps 1 50 kB 2Mbps 2Mbps 16Mbps 2 50 kB 2Mbps 2Mbps 16Mbps 3 100 kB 8Mbps 8Mbps 40Mbps 4 100 kB 8Mbps 8Mbps 40Mbps 5 100 kB 8Mbps 8Mbps 40Mbps 6 50 kB 0.4Mbps 4Mbps 40Mbps 7 50 kB 0.4Mbps 4Mbps 40Mbps 8 50 kB 2Mbps 16Mbps 40Mbps

Benchmarks . WFQ scheduling & Buffer Management FIFO scheduling & WFQ Scheduling & Buffer Management No Buffer Management FIFO scheduling & No Buffer Management

Aggregate Throughput Buffer Partitioning vs. No Buffer Management Throughput (Mbps Buffer Size (Mb)

Losses for conformant flows Buffer Partitioning vs. No Buffer Management Loss (percent) Buffer Size (Mb)

Aggregate Throughput Buffer Sharing vs. No Buffer Management Throughput (Mbps) Buffer Size (Mb)

Excess Capacity Sharing Throughput (Mbps) Buffer Size (Mb)

The Hybrid Scheme F l ow - - U ni t of buf f er allocation Queue - - uni t of bandwi dth allocation WFQ Scheduler P rotect i ndi vi dual fl ows F ai r E xcess Capaci ty whi l e shari ng buf f ers di stri buti on based on “ hol es” “ A l l ocati on” & “ H eadroom”

Throughput Integrated Scheme Throughput(Mbps) Buffer Size (Mb)

Flow Isolation and Rate Guarantees Integrated Scheme Loss (percent) Buffer Size (Mb)

Excess Capacity Distribution Integrated Scheme Throughput (Mbps) Buffer Size (Mb)

Conclusions The Integrated Scheme •Scheduling on a fixed number of “Queues” and buffer management on the finer granularity of “Flows” Rules of thumb for flow grouping •Buffer sharing & flow isolation within a single use “headroom” to limit impact on conformant flows in buffer limited system •Empty buffers (holes) to regulate the sharing of excess bandwidth flexible notion of fairness •Flexibility to choose tradeoffs depending on operational environment •Some quantitative guidelines on the choice of design and setting of parameters