QoS Solution?

Michal Kolinski mkol at poczta.neostrada.pl
Tue Mar 29 12:31:46 UTC 2005


> I don't know of any magic answers here. No solution will work perfectly
> for all situations. So we need to consider how much benefit we get from
> each approach compared to how much complexity we add to the Tor code. I
> think it might be worth trying this one out (especially as Mike keeps
> volunteering to write code for us :)

If you decide to add this stupid QoS please add an option to disable it.
Why do you think that downloading 100 chunks 200 KB each should be faster 
than downloading one 20000KB file? Actually 100x200KB only would cause 
higher overhead than 1x20000KB.
Tor must have either perfect QoS (each USER, not connection gets the same 
amount of bandwidth, no matter what service - filesharing or WWW) or no QoS 
at all. No QoS is better than QoS per transferred bytes or per service.
QoS per service works properly only in single-user environment.

When PNG graphics format was being developed (many years ago), I was always 
saying that W3C should either add perfect gamma correction or no correction 
at all. They completely ignored me and added gamma correction that doesn't 
fully work. Effects of this: http://hsivonen.iki.fi/png-gamma/. It made PNG 
worse than GIF.



More information about the tor-talk mailing list