Various previous discussions:<p>It's time to replace TCP in the datacenter position paper <a href="https://news.ycombinator.com/item?id=33401480">https://news.ycombinator.com/item?id=33401480</a> <a href="https://news.ycombinator.com/item?id=42168997">https://news.ycombinator.com/item?id=42168997</a><p>Review of Homa protocol <a href="https://news.ycombinator.com/item?id=28204808">https://news.ycombinator.com/item?id=28204808</a><p>Review of Linux implementation of Homa <a href="https://news.ycombinator.com/item?id=28440542">https://news.ycombinator.com/item?id=28440542</a><p>TCP vs. RPC part 1 <a href="https://news.ycombinator.com/item?id=34871670">https://news.ycombinator.com/item?id=34871670</a> part 2 <a href="https://news.ycombinator.com/item?id=34871710">https://news.ycombinator.com/item?id=34871710</a> part 3 <a href="https://news.ycombinator.com/item?id=35228716">https://news.ycombinator.com/item?id=35228716</a>
Without regard to the detailed background, design or analysis, the approach of plucking a concern out of the Network/Data Link layer (prioritization / QoS) and moving it up to the Transport is a remarkably simple / clever start.<p>At least that's how I picture the start of the study of such a design.
What is going on in a datacenter that motivates a protocol like this? I admit I am ignorarant. Is it for internal allocation, say traffic between Kubernetes nodes, this is envisioned?
Is anyone actually using Homa? I have heard that it has a few fundamental issues as described in the paper, and many people who want what Homa offers are using their own thing.