←back to thread

It's time to replace TCP in the datacenter (2023)

(arxiv.org)
188 points ilove_banh_mi | 1 comments | 18 Nov 24 01:42 UTC | HN request time: 0.199s | source
Show context
akira2501 ◴[18 Nov 24 04:53 UTC] No.42169811[source]▶
> If Homa becomes widely deployed, I hypothesize that core congestion will cease to exist as a significant networking problem, as long as the core is not systemically overloaded.

Yep. Sure; but, what happens when it becomes overloaded?

> Homa manages congestion from the receiver, not the sender. [...] but the remaining scheduled packets may only be sent in response to grants from the receiver

I hypothesize it will not be a great day when you do become "systemically" overloaded.

replies(2): >>42170099 #>>42182735 #
1. bewo001 ◴[19 Nov 24 12:33 UTC] No.42182735[source]▶
I don't understand the difference to TCP here. If the path is not congested but the receiving endpoint is, the receiver can control the bandwidth by reducing the window size. Ultimately, it is always the sender that has to react to congestion by reducing the amount of traffic sent.

RPC is something of a red flag as well. RPCs will never behave like local procedure calls, so the abstraction will always leak (the pendulum of popularity keeps swinging back and forth between RPC and special purpose protocols every few years, though).