[tor-project] network team meeting notes, 26 June

Nick Mathewson nickm at torproject.org
Mon Jun 26 18:21:53 UTC 2017

Hi, all!

You can find the transcript from our weekly meeting at

Here are our notes:

Network team pad, for 26 June Meeting (or 27 June, for those in UTC+5 or later)

Notes from last week's meeting:
    * https://lists.torproject.org/pipermail/tor-project/2017-June/001208.html
      (Did we do what we had planned?)

Announcements and reminders:
    * review-group-18 is still not cleared.  We have 3 tickets left in
needs_review there which need to get reviews before nickm makes a
    * review-group-19 will come once the last 3 tickets in
review-group-18 are done.
    * There are still 34 tickets in 0.3.1.x, including 9 in status
"new" with no owner. We need to get 0.3.1.x done!
    * Employees: remember harvest. [also, if you have pending expense
reports, try to submit them before end of june - you can do that via
harvest as well if you like]
    * Isabela will ping people to start collecting information for nsf reports

Things we should talk about:
    * We need a PT person (or at least point of contact)
    * How important is noise for padding statistics? Mike is thinking
"Not very" compared to the type of attack described vs related guard
discovery attacks I want to fix first.

Tasks for after the meeting:
    * July team rotation schedule is available at:

    Last week:
        * Fixed a nasty security bug in PrivCount, released 1.0.2, and
upgraded all the things
        * Deployed PrivCount HSDir upload stats on a relay for testing
        * Looked at all the privcount-tagged trac tickets and updated
the PrivCount TODO lists
        * Tried to help out with the PaddingStatistics implementation
#17857 etc.
        * Stop relays from knowing HS client and service addresses
#22688, #22689
        * Added a hs-network-min network to chutney: #22667
        * Logged a few bugs in Tor and Tor Browser
    This week:
        * Document the PrivCount HSDir upload stats test outcomes
        * Prioritise the remaining 6 hidden service events
          (HSDir client, HSDir cache evict, Intro service, Intro
client, Rend service, Rend client)

    Last Week:
        * Worked on #17857 and #22212.
        * Discussed padding statistics - I am not seeing this as urgent.
        * Reviewed #13605 (reduced exit policy option)
    This Week:
        * Mozilla all hands, all week...

Last week:
   - Caught up with backlog that accumulated during meeting
   - Work on prop247 stuff. Started the simulator work: It can currently do a
     basic prop247 simulation and print out some useful output.
   - Worked with David on some prop224 stuff
   - Reviewed some prop224 code and other tickets
This week:
   - Review prop224 groundwork #21979
   - Continue prop247 simulation. This week I'll spend debugging, make the
     compromise adversary models and create some nice graphs.
   - Update prop224 spec wrt SRV / desc overlap behavior based on discussions
     with david and the current implementation.
   - Rewrite the overlap logic of prop224 to use the slot system instead of
     absolute human times.
   - Revisit #22006 based on nick/isis review.
   - Hopefully Write wilmington blog post if I get the time.

    Last week:
    - Finished reviewing asn's client-side descriptor fetch code. He
replied and I also did reply back :)
    - Read some lattice-based crypto (as a hobby)
    This week:
    - Review dgoulet's load and configure service code #21979 (with asn)
    - Maybe*, take over Fuzz v3 hidden services #21509 (if it's not done yet)

  Last week:
    * DRL meeting on Mon, Tue and Wed. (Nick asks: any takeaway lessons?)
    * Code review on a couple of hidden service tickets.
    * We do have #21859 (prop224 e2e code) in needs_review now on Gitlab which
      is the first ticket in review for prop224 inclusion in 032.
    * Discussed some action items with asn on prop224 about some intro
      points engineering issues and overlap period logic that is out of sync
      in the proposal.
  This week:
    * Full throttle prop224 work with splitting the service branch in smaller
      ones for review and upstream merge.
    * Continuing and hopefully finishing the client implementation.
    * July is really our target for a minimally working prop224 for both our
      sponsors and our Defcon talk (early August). Thus, expect lots of work
      on that on my part.

    Last week:
        * Thanks to ahf, helped diagnose big regression from prop140
(consensus diff) work: we weren't checking if-modified-since correctly
any more!  Ahf has a fix; testing again now.
        * Worked on compression issues that toralf found.
        * Reviewed lots of code. Merged a bunch. Continued bug triage.
        * Chased through some contacts to find clarity about copyright
status for patches on old ticket from Access Labs. All's well.
    This week:
        * Pending isabela, send out notes on Who Works On What Between
Now And September.
        * Circulate "supported platforms list" current draft from
https://pad.riseup.net/p/QWpY4UYfAsaK , turn it into a wiki page,
pending comment. Any comments?
        * Review more code
        * Fix pending 0.3.1.x bugs
        * Release another 0.3.1.x alpha
        * Write some stuff for 0.3.2.x
        * Work on or defer remaining 0.3.1.x tickets
        * More ticket triage on "unspecified"
        * Prepare for 0.3.2.x triage.


Last week:

    Sponsor 4:
        - Been working on Sponsor 4 measurements with Nick:
            Looks good for clients so far, but we have regressed badly for
            relays. We think we know what the primary cause is (see #22702).
            - Been working on a fix for bug #22702 (Never send a consensus
              which the downloader already has).
            - Wrote a script to collect directory measurement from Tor
logs. Can be
              found in ahf/tor-sponsor4-compression on GL.com
            - Working with Nick on the initial results: data in
        - Code review: #22719, #22502 (and thus #22626 + #22628), #22629
        - Code review: #3056

This week (ordered):

    - Land fix(es) for #22702 after we have confirmed they work.
    - Land fix for the issue where a relay gets killed by the sandbox because
      we allocate a little bit over 16 MB memory.
    - Actually get around and finish the merge of prop #278 that I've kept
    - Help out with 0.3.1 issues
    - If time allows for it: write down a short text that tells the progress
      we have made with the sponsor 4 work (?)

    Last week:
        * Reviewed a lot of review-group-18 and 0.3.1 tickets
        * Progress on reviewing #7869
        * Registered for oniongit
    This week:
        * Check #7869 test coverage
        * Investigate #20532
        * Investigate more test-stem issues
        * Bug triage
        * Ping komlo about followup from Wilmington testing session
        * Follow up and review support person hiring stuff
        * Experiment more with oniongit

    last week:
        - passed on next tasks to invoice for sponsorR to Brad
        - finally sent sponsor4 may report
        - working with tommy on modularization SOI
        - synced with Brad to get the info to build the spreadsheet
for network team people's time allocation
    this week:
        - still waiting on one spreadsheet from brad to get back with
nick on people's time allocation - hope to get that done this week
        - looking forward to finish a couple of pending tasks: update
the brief for tor launcher automation and get an update on user growth
strategy out to tor-internal
        - off on july 4th (national holiday) and july 5th-7th to get
some personal stuff done
        - catch up with hiro on what we discussed in wilmington
related to gitlab and trac

    Last weekish:
        - I hope you like documentation. I wrote a bunch
        - I hope you like tests. Did those too
    Time weekish:
         - nickm asks: "how is the permission-to-release coming?"

          - pastly says: no news. Was told it would take ~4 weeks
about ~2.5 weeks ago

- but plans on sharing with dgoulet and anyone that asks nicely ASAP

        - Finish fleshing out KIST tests. Main one remaining is for
the scheduling loop
        - Once done fleshing out, will push to oniongit and ping
dgoulet to make time for meeeeee

    Last week:
        * code review on #22006 and #13703
        * small patch for #4019
        * did employment/insurance/tax paperwork
        * answering questions and comments for the internship post
    This week:
        * Expenses for amsterdam (!), seattle, and wilminton
        * Lay out tickets for tasks in the next couple months and
figure out sponsors and timelines
        * Some meetings for Mozilla All-Hands that I should attend
        * BBC Radio interview
        * Start setting up backend stuff for moat since I believe the
TB Team needs it soon(anything else that is higher priority?) [isa: I
think for now this is it. like we chated TB team in august:
        * Review the internship applications and narrow down the
choices and pick one and get them started
        * Email to Rust team about where we're at
        * Email to oxidators about steps moving forward and tasks we
need to assign ourselves

More information about the tor-project mailing list