This came up on the fediverse ~18 months ago; here’s what I said at the time:

“Depends on the workload, of course, but it feels like a very comfortable fit. 
Certainly the vast majority of 9p *messages* on a given network are a small 
fraction of their “small” threshold. I need to read the paper more carefully; 
the RPC oriented design sounds promising, but the “at least once” model and no 
explicit ACKs raises some questions for me.

it would be a fun experiment, for sure! Got a datacenter I can try it on? ;-)”

Reading the paper more didn’t really answer the questions, and I still haven’t 
found a datacenter to try it on.


> On Sep 21, 2026, at 22:48, Skip Tavakkolian <[email protected]> 
> wrote:
> 
> I will preface this by acknowledging that I know very little about the
> details of the protocol and I could be totally off the mark.
> 
> I only recently found out about Ousterhout's latest project from this
> recent talk. Homa protocol is based on the paper referenced below. My
> first impression is that the 9P (RPC) on top of IL seem very similar
> to one of Homa's features -- the receiver-driven congestion control.
> I'm also wondering if a variation of IL could implement these
> features, or if 9P could benefit from use of Homa.  Has anyone done
> any deep thinking?
> 
> Ousterhout's talk
> https://www.youtube.com/watch?v=eZ8WWZzoaR0
> 
> Homa: A Receiver-Driven Low-Latency Transport Protocol Using Network 
> Priorities
> https://people.csail.mit.edu/alizadeh/papers/homa-sigcomm18.pdf

------------------------------------------
9fans: 9fans
Permalink: 
https://9fans.topicbox.com/groups/9fans/Td0379c45387a908c-Mbc3538eb37a2395acaa2711c
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

Reply via email to