Homa: The End of TCP for AI Clusters — John Ousterhout, Stanford
John Ousterhout from Stanford presents Homa, a new transport protocol for AI datacenters.
The problem is that modern AI workloads consist of many small, latency-sensitive messages mixed with large data transfers — when small messages queue behind large ones, GPUs become idle and wasteful. TCP and RDMA react slowly to congestion and cannot prioritize short messages. Homa solves this by understanding message boundaries, letting the receiver control flow with grants to senders, and leveraging priority queues in modern switches. Benchmarks show Homa can reduce tail latency for small messages over 10 times compared to TCP.
Vibekollen prepared this summary with AI from the original publication. The content belongs to AI Engineer.
More to read
Lumaril, which develops AI software for optimizing industrial production by connecting to existing control systems, raised a $66M Series A at a $400M valuation
Techmeme 3 h ago
DeepSeek open sources Super-Agent as a research prototype, a proactive coding agent it says can monitor group chats for coding issues and suggest solutions
Techmeme 3 h ago
Insurance software maker Guidewire agrees to acquire Hyperexponential, which makes AI-powered insurance underwriting and pricing software, for $1.7B in cash
Techmeme 3 h ago
One word for AI agents in 2027
Nate Herk 7 h ago