[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [tor-bugs] #1854 [Analysis]: Investigate raising the minimum bandwidth for getting the Fast flag
#1854: Investigate raising the minimum bandwidth for getting the Fast flag
---------------------------------------+------------------------------------
Reporter: arma | Owner: arma
Type: task | Status: needs_information
Priority: normal | Milestone:
Component: Analysis | Version:
Keywords: performance loadbalancing | Parent:
Points: | Actualpoints:
---------------------------------------+------------------------------------
Comment(by cypherpunks):
Post by Paul, who still doesn't have a proper account:
Been talking to Ian about this today here at Dagstuhl. I don't think all
the effects of significantly shrinking the set of nodes that are ever
chosen has been considered. If the network shrinks to c. 1/4 it's current
size, this has the potential for tremendous psychological impact on users,
relay volunteers, some adversaries, funders, etc. There is thus a big
difference between switching a lot of nodes to be never chosen vs.
changing the distributions to make them more rarely chosen. Instead of
changing the fast flag, it would then make more sense to alter the
bandwidth weighting. And the more gradual the change in probability of
being chosen, the less any nodes will naturally count as the group that
has been simply excluded. If performance is best served by more of a step
function, then perhaps something in between will still significantly
improve performance statistics without, e.g., resulting in graphs showing
a 75% drop in the number of nodes with the fast flag when the change is
rolled out.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/1854#comment:54>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs