[TWN team] Recent changes to the wiki pages
Lunar
lunar at torproject.org
Mon Sep 22 19:40:06 UTC 2014
===========================================================================
=== https://trac.torproject.org/projects/tor/wiki/TorWeeklyNews/2014/38 ===
===========================================================================
version 25
Author: harmony
Date: 2014-09-22T19:04:17+00:00
remove covered item
--- version 24
+++ version 25
@@ -207,6 +207,5 @@
Possible items:
- * Guardiness: Yet another external dirauth script https://lists.torproject.org/pipermail/tor-dev/2014-September/007489.html
* (event) https://mailman.boum.org/pipermail/tails-project/2014-September/000037.html next Tails contributors meeting
* Attack linking Tor circuits https://lists.torproject.org/pipermail/tor-dev/2014-September/thread.html#7489
version 24
Author: harmony
Date: 2014-09-22T18:57:31+00:00
write guardiness item
--- version 23
+++ version 24
@@ -44,6 +44,49 @@
[XXX]: https://www.eff.org/deeplinks/2014/09/tor-challenge-inspires-1635-tor-relays
[XXX]: https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
[XXX]: https://www.torproject.org/docs/tor-doc-relay
+
+Guardiness and Tor’s directory authorities
+------------------------------------------
+
+When a Tor relay is first assigned the Guard flag by the directory
+authorities (or “dirauths”) it sees a dip in the amount of traffic
+passing through it, because Guard capacity is a scarce resource on the
+Tor network and, as Roger Dingledine explained last year [XXX], “all the
+rest of the clients back off from using you for their middle hops,
+because when they see the Guard flag, they assume that you have plenty
+of load already from clients using you as their first hop”, an
+assumption which is only correct after clients have had enough
+opportunity to select the new guard. With the recent move to single
+entry guards, an even longer period of time may pass before a young
+guard can be selected as a first hop by old clients.
+
+“Guardiness”, or GuardFraction, is a proposed measurement to let
+dirauths, and therefore clients, work out how much of a relay’s capacity
+is being used for first hops by clients, and how much for second and
+third hops, by finding the fraction of recent consensuses in which
+the relay has been given the Guard flag [XXX]. George Kadianakis
+published [XXX] an analysis of ways in which dirauths’ votes could be
+extended to include this guardiness measurement, taking into account the
+time and effort required to parse large numbers of Tor consensuses very
+quickly. The initial proposal was to ask dirauths to run a script each
+hour that would extract the data required for parsing into “summary
+files”, while Peter Palfrader suggested [XXX] that loading every
+consensus into a database for later querying might be more efficient,
+and Sebastian Hahn asked [XXX] how this measure might fail in different
+situations.
+
+“This feature is by far the trickiest part of prop236 (guard node
+security) and I wanted to inform all dirauths of our plan and ask for
+feedback on the deployment procedure”, wrote George. If you have any
+comments to add to the discussion so far, please send them to the tor-dev
+mailing list [XXX].
+
+ [XXX]: https://blog.torproject.org/blog/lifecycle-of-a-new-relay
+ [XXX]: https://gitweb.torproject.org/torspec.git/blob/HEAD:/proposals/236-single-guard-node.txt#l101
+ [XXX]: https://lists.torproject.org/pipermail/tor-dev/2014-September/007489.html
+ [XXX]: https://lists.torproject.org/pipermail/tor-dev/2014-September/007526.html
+ [XXX]: https://lists.torproject.org/pipermail/tor-dev/2014-September/007490.html
+ [XXX]: https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-dev
Miscellaneous news
------------------
--
Your friendly TWN monitoring script
In case of malfunction, please reach out for lunar at torproject.org
or for the worst cases, tell weasel at torproject.org to kill me.
More information about the news-team
mailing list