Homa: A Receiver-Driven Low-Latency Transport Protocol Using Network Priorities (Complete Version)
Not on the map. Why?
Abstract (arXiv)
Homa is a new transport protocol for datacenter networks. It provides exceptionally low latency, especially for workloads with a high volume of very short messages, and it also supports large messages and high network utilization. Homa uses in-network priority queues to ensure low latency for short messages; priority allocation is managed dynamically by each receiver and integrated with a receiver-driven flow control mechanism. Homa also uses controlled overcommitment of receiver downlinks to ensure efficient bandwidth utilization at high load. Our implementation of Homa delivers 99th percentile round-trip times less than 15μs for short messages on a 10 Gbps network running at 80% load. These latencies are almost 100x lower than the best published measurements of an implementation. In simulations, Homa's latency is roughly equal to pFabric and significantly better than pHost, PIAS, and NDP for almost all message sizes and workloads. Homa can also sustain higher network loads than pFabric, pHost, or PIAS.
Key facts. Cited by 9 papers in the map (data through 7 Oct 2026). First posted on arXiv 26 Mar 2018.
Builds on
No reference list yet.
Built on by
Citing papers per year: 2022: 1; 2023: 2; 2024: 3; 2025: 0; 2026: 3.
9 papers.
- 2022A new hope for network model generalizationcited by 11
- 2023Boosting Distributed Machine Learning Training Through Loss-tolerant Transmission Protocolcited by 3
- 2024I've Got 99 Problems But FLOPS Ain't Onecited by 3
- 2023GraphCC: A Practical Graph Learning-based Approach to Congestion Control in Datacenterscited by 0
- 2024MLCC: A Congestion Control Technique to Accelerate ML Trainingcited by 0
- 2024SwiftQueue: Optimizing Low-Latency Applications with Swift Packet Queuingcited by 0
- 2026Assistants, Not Architects: The Role of LLMs in Networked Systems Designcited by 0
- 2026Resilient AI Supercomputer Networking using MRC and SRv6cited by 0
- 2026Tuning Collective Patterns to Alleviate Congestion in Shared AI Clusterscited by 0
- 2026Tuning Collective Patterns to Alleviate Congestion in Shared AI Clusters3 Sep 2026
- 2026Resilient AI Supercomputer Networking using MRC and SRv65 May 2026
- 2026Assistants, Not Architects: The Role of LLMs in Networked Systems Design28 Apr 2026
- 2024SwiftQueue: Optimizing Low-Latency Applications with Swift Packet Queuing8 Oct 2024
- 2024I've Got 99 Problems But FLOPS Ain't One1 Jul 2024
- 2024MLCC: A Congestion Control Technique to Accelerate ML Training14 Feb 2024
- 2023GraphCC: A Practical Graph Learning-based Approach to Congestion Control in Datacenters9 Aug 2023
- 2023Boosting Distributed Machine Learning Training Through Loss-tolerant Transmission Protocol7 May 2023
- 2022A new hope for network model generalization12 Jul 2022