Pyrrha: Congestion-Root-Based Flow Control to Eliminate Head-of-Line Blocking in Datacenter
Kexin Liu, Zhaochen Zhang, Chang Liu, Yizhi Wang, Vamsi Addanki, Stefan Schmid, Qingyue Wang, Wei Chen, Xiaoliang Wang, Jiaqi Zheng, Wenhao Sun, Tao Wu
Abstract
In modern datacenters, the effectiveness of end-to-end congestion control (CC) is quickly diminishing with the rapid bandwidth evolution. Per-hop flow control (FC) can react to congestion more promptly. However, a coarse-grained FC can result in Head-Of-Line (HOL) blocking. A fine-grained, per-flow FC can eliminate HOL blocking caused by flow control, however, it does not scale well.
This paper presents Pyrrha, a scalable flow control approach that provably eliminates HOL blocking while using a minimum number of queues. In Pyrrha, flow control first takes effect on the root of the congestion, i.e., the port where congestion occurs. And then flows are controlled according to their contributed congestion roots. A prototype of Pyrrha is implemented on Tofino2 switches. Compared with state-of-the-art approaches, the average FCT of uncongested flows is reduced by 42%-98%, and 99th-tail latency can be 1.6×-215× lower, without compromising the performance of congested flows.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Your agent calls
Luneget_paper_fulltext
Free to start. No credit card required.
Terminal
Install the CLIlune papers fulltext bd79b4f0-843d-469f-b48a-f070b5d564b5Builds on20
- MegaScale: Scaling Large Language Model Training to More Than 10, 000 GPUsZiheng Jiang, Haibin Lin, Yinmin Zhong, Qi Huang et al.NSDI 2024 · 415 citations
- PINT: Probabilistic In-band Network TelemetryRan Ben Basat, Sivaramakrishnan Ramanathan, Yuliang Li, Gianni Antichi et al.SIGCOMM 2020 · 268 citations
- Alibaba HPN: A Data Center Network for Large Language Model TrainingKun Qian, Yongqing Xi, Jiamin Cao, Jiaqi Gao et al.SIGCOMM 2024 · 173 citations
- CASSINI: Network-Aware Job Scheduling in Machine Learning ClustersSudarsanan Rajasekaran, Manya Ghobadi, Aditya AkellaNSDI 2024 · 144 citations
- SP-PIFO: Approximating Push-In First-Out Behaviors using Strict-Priority QueuesAlbert Gran Alcoz, Alexander Dietmüller, Laurent VanbeverNSDI 2020 · 140 citations
Related papers
- Backpressure Flow ControlPrateesh Goyal, Preey Shah, Kevin Zhao, Georgios Nikolaidis et al.NSDI 2022
- InfiniFlow: Decoupling Virtual Channel Scalability from Buffer Requirements in Lossless Datacenter NetworksZerui Tian, Sen Liu, Minkun Xue, Hao Shangguan et al.SIGCOMM 2026
- FLB: Fine-grained Load Balancing for Lossless Datacenter NetworksJinbin Hu, Wenxue Li, Xiangzhou Liu, Junfeng Wang et al.USENIX ATC 2025 · 10 citations
- Towards timeout-less transport in commodity datacenter networksHwijoon Lim, Wei Bai, Yibo Zhu, Youngmok Jung et al.EuroSys 2021 · 28 citations
- BCC: Re-architecting Congestion Control in DCNsQingkai Meng, Shan Zhang, Zhiyuan Wang, Tao Tong et al.INFOCOM 2024 · 11 citations
