[tor-dev] Next version of the algorithm

[George Kadianakis]

Ola Bini <obini at thoughtworks.com> writes:

> Hi George,
>
> Sorry I've been slow to answer and send more info - was spending all
> day Saturday and Sunday, and this week I'll be in all day meetings.
>
> Anyway, just to avoid confusion - After getting the algorithm to the
> latest stage you saw below, we started work on implementing it to
> simulate it. Iván and Tania have been working on that.
>
> I also wrote it in draft form, so this:
> https://github.com/twstrike/torspec/blob/review/proposals/259-guard-selection.txt
> Is completely rewritten more or less from scratch.
>
> When it comes to receive new consensus, it is in the latest version of
> the written proposal above.
>
> The link Chelsea sent is a link to the original protocol, not the new
> one.
>
> So to summarize where we are:
>
> - We have fixed a bunch of issues in the existing guardsim code
> - We have added a lot of simulation cases to the existing guardsim
> - code
> - We have implemented most of the original implementation in the
> - guardsim code
> - We are implementing the new algorithm, described in the spec above
> - in the guardsim code, and expect that to be done today or tomorrow
> - The spec above is polished enough, that if you agree we could send
> - it to a wider distribution and get thoughts.
> - We might be able to start implementing this in tor proper tomorrow
> - or wednesday hopefully.
>

Sounds great.

I would be interested in learning the results of those simulations!

In any case, starting the tor implementation will be useful anyhow.  Even if
the very final algorithm is not 100% the same as the new prop259, it will
definitely be closer to that than the old mess.

Maybe we can have another IRC meetup this week to speak about the simulation
results and discuss how the little-t-tor changes could work? I will be
available tomorrow and on Wednesday from 14:00 to 18:00 UTC. Let me know if
that works for you.