User-initiated QoS in Tor?

[Björn Påhlsson]

(I was looking over the archive and couldn't find this topic being
discussed before.)

Could user-initiated QoS be used to increase performance of Tor, or are
there any serious security issues preventing it?

I can see that normal QoS at the exit node is bad (network neutrality,
deep packet inspection, etc.), but if the user simply hinted or advised
that a circuit could use a latency boost for a short time, then Tor
routers could use that information to optimize traffic to give better
performance to data that's small and has need for low latency (IM, ssh,
web text). The decision on what need low latency is decided by the user,
and routers have the option to use this information.  If this
optimization is capped to only affect, say, max 1-5% of all data sent
through a node, then data with no priority should not be notably
affected.

/Björn Påhlsson
***********************************************************************
To unsubscribe, send an e-mail to majordomo at torproject.org with
unsubscribe or-talk    in the body. http://archives.seul.org/or/talk/