[tor-reports] George's status report: August 2014

[George Kadianakis]

# Activities of August 2014

- During August I mainly worked on the guardiness part of prop 236 (section 1.3).
  That's ticket #9321 and it's the major blocker that needs to be solved
  if we ever want to increase the guard lifetime period (more on this later).

  To this end, I have written a Python script that calculates the
  guard fraction of guards given a set of consensuses. I have
  published it here and it's ready for review:
  https://gitweb.torproject.org/user/asn/hax.git/shortlog/refs/heads/guardiness
  That script is supposed to be run by the directority authorities, as
  described here:
  https://trac.torproject.org/projects/tor/ticket/9321#comment:22

  The deployment process involves a cron'ed script that summarizes
  consensuses to summary files, and another cron'ed script that reads
  those summary files to calculate the guard fraction of guards. Then
  little-t-tor reads the results of the second script. Rather iffy, eh?
  I suspect deployment is going to be trickier than it seems, and I
  need to think of a few ways to make it easier and more failsafe.

  On the little-t-tor side, I have written some PoC code that
  implements GuardFraction voting by authorities, and I have also
  implemented the consensus GuardFraction parsing by clients. I have
  been testing voting/parsing in a local test network and it seems to
  work.

  Next step is to make clients consider GuardFraction during path
  selection. And then unittests. And then making sure that our
  deployment strategy is actually doable. Sebastian and Nick has
  commented on the deployment strategy and they found it workable...

  I also wrote a torspec patch for proposal 236 to further specify how
  authorities should vote on GuardFraction:
  https://lists.torproject.org/pipermail/tor-dev/2014-August/007439.html

  Another existential question here, is whether we actually want to
  increase the guard lifetime period. Originally, proposal 236
  suggested extending guard lifetime to 9 months. It's still unclear
  whether this is a good idea. If we don't plan to extend the lifetime
  period of guards any time soon is it still worth working on #9321?

  This project is roadmapped for November and I think we are in a good
  shape so far. 

- My unittests (#12207) for Tor's entry guards got merged in
  little-t-tor :) Same goes for a smaller refactoring change (#12202)
  that improves entry code readability.

- While playing with consensuses and guard nodes, I noticed that many
  unmeasured relays have the Fast/Stable flag; afaik this is not
  supposed to happen. I filed a report at
  https://trac.torproject.org/projects/tor/ticket/12877 but I haven't
  had the chance to investigate further.

- Wrote two research-y [tor-dev] posts.

  The first is about improving Tor's entry guard data structures (#12595)
  and specifically about gracefully handling network offline events:
  https://lists.torproject.org/pipermail/tor-dev/2014-August/007346.html

  The second is about Introduction Point security in the upcoming
  rend-spec-ng.txt:
  https://lists.torproject.org/pipermail/tor-dev/2014-August/007335.html

  Both threads received helpful feedback. I still need to find
  something smart to respond in the first thread.

- I reviewed Yawning's obfs4proxy (#12606) and it seems to be in a
  very good shape. I have also been running it in my bridge (it
  performs obfs3/obfs4) and it's serving many thousands obfs3 users
  every day like a boss.

  Yawning tagged the first release (0.0.1) yesterday, and next steps
  is server-side packages and client-side TBBs :)

- After mentioning that the webpage needs an update with new PT
  information, some folks sent me an email asking me what changes are
  needed. I sent an email to [tor-dev] with some ideas:
  https://lists.torproject.org/pipermail/tor-dev/2014-August/007294.html
  Hopefully someone will pick it up :)

- Tested OONI release ooni-1.1.0 and QA'ed the bridge_reachability
  test. We also found an FTE bug:
  https://github.com/kpdyer/fteproxy/issues/159

  The bridge_reachability test now works quite well and the OONIs are
  taking good steps towards deployment.

- Finished GSoC with ahmia.fi successfully :)

  Now that GSoC is over, Juha from ahmia.fi is in big need of
  developers and help.  Shoot him an email at juha.nurmi at ahmia.fi if
  you are interested, or send him patches in github:
  https://github.com/juhanurmi/ahmia

  Also, enjoy a picture with ahmia's infrastructure:
  https://raw.githubusercontent.com/juhanurmi/ahmia/1ae0b6d255e39592029de933349f605bd6e723dc/technical_architecture.png
  
# Activities for September 2014 

- Work on #9321, the guardines stuff.

  During September, I hope to bring the little-t-tor patch closer to
  completion, write unittests and give the whole system some good
  testing using Chutney. I also plan to talk more with dirauth
  operators to get a better understanding on how to make deployment
  more practical.

- Continue helping with obfs4 deployment and PTs. 

- Help more with the bridge_reachability OONI test.