[Bloat] Adding congestion control to the Tor network

[Bloat] Adding congestion control to the Tor network

[Toke Høiland-Jørgensen]

I figured that this lengthy overview of past and future attempts at
adding congestion control to the Tor network might be interesting to
people here:

https://lists.torproject.org/pipermail/tor-dev/2020-January/014140.html

-Toke

Re: [Bloat] Adding congestion control to the Tor network

[Dave Taht]

I'm looking forward to their upcoming netdevconf talk:

https://netdevconf.info/0x14/session.html?talk-reducing-kernel-queuing-delays-with-TCP-window-space-events

There are a metric ton of network and congestion control related talks at
this conference in this go-round. Some ideas seem brilliant, others,
brain damaged.

Toke Høiland-Jørgensen <toke@toke.dk> writes:

> I figured that this lengthy overview of past and future attempts at
> adding congestion control to the Tor network might be interesting to
> people here:
>
> https://lists.torproject.org/pipermail/tor-dev/2020-January/014140.html

That was reallly good, thank you. Heavy on ECN derived
ideas... fq_codel/pie not in there... 

Once again FQ gets a bit of short shrift, given that only a few fat
flows need congestion control to kick in relative to the others. Of
course I have no idea from the regularity of classic FQ how that yields
to traffic analysis in one way or another, but I do certainly hope they
are mixing up traffic as much as possible within a cell?

I've generally found tor to be pretty unusable for my impatient self. 
The last time I paid attention was in this: http://www.cypherpunks.ca/~iang/pubs/defenestrator.pdf
which is also referenced in the above discussion.



>
> -Toke
> _______________________________________________
> Bloat mailing list
> Bloat@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/bloat