SP-PIFO: Approximating Push-In First-Out Behaviors using Strict-Priority Queues
summary
Approximation of PIFO behavior uses a mapping that adapts packet rank→priority-queue thresholds at packet time and runs on real switches via a P4 implementation with one register per queue and occasional packet resubmission. SP-PIFO claims line-rate operation, arbitrary scaling to many flows and ranks, and near-ideal results with as few as 8 priority queues (evaluated in a 144-switch leaf‑spine sim with 1/4 Gbps links). Two local mechanisms push-up and push-down minimize scheduling errors; FCT and fairness metrics match ideal PIFO closely.
AI-written summary · about
chapters (9)
- 0:00The problem of scheduling
- 1:32Introducing SP-PIFO
- 2:36PIFO abstraction explained
- 3:52SP-PIFO mapping challenges
- 5:15Mapping strategies
- 7:23Dynamic adaptation mechanisms
- 10:57Hardware implementation
- 12:07Performance evaluation
- 14:04Q&A and conclusion
related talks
15mEverything Matters in Programmable Packet Scheduling
13mSifter: An Inversion-Free and Large-Capacity Programmable Packet Scheduler
25mApproximating Fair Queueing on Reconfigurable Switches
13mClubHeap: A High-Speed and Scalable Priority Queue for Programmable Packet Scheduling
16mFlow Scheduling with Imprecise Knowledge
22mEiffel: Efficient and Flexible Software Packet Scheduling
16mBackpressure Flow Control
17mFormal Abstractions for Packet Scheduling