[tor] Adding another node

[Ben Kochie]

Thanks Isis.

On Wed, Aug 28, 2013 at 3:38 PM, isis <isis at torproject.org> wrote:
> Ben Kochie transcribed 4.0K bytes:
>> Sounds like something we could take over, but didn't we want to limit our
>> impact on the network?
>>
>> Maybe we can find a non-overlapping crew for the 2nd node.
>
> While "diversity of the Tor network" has been discussed frequently of late,
> within the Tor Project (and also within the broader community) it is my
> understanding that the current view is that having responsible, transparent,
> and friendly organisations running families of exit relays is positive for the
> network, in more ways than merely adding exit bandwidth capacity. People seem
> more apt to trust the relays run by Noisebridge, CCC, and other allies.
>
> I would argue that trust is absolute imperative in a decentralised system (as
> much as the Tor network could be considered decentralised), and that NoiseTor
> running Amunet would be better than some other party running Amunet.
>
> That said, if the sysadmin work has grown to be too much...I understand.

Ok, that's answers my question.  The sysadmin work is pretty minor.

>> On Aug 27, 2013 8:22 PM, "Andy Isaacson" <adi at hexapodia.org> wrote:
>> >
>> > The funders of Amunet have offered to donate to Noisetor if we are
>> > willing to take over running the node.  It's a similar value to our
>> > current exit01, approximately $800/mo for 500 Mbps of throughput, on a
>> > rented colo server.
>> >
>
> Also, there are now opportunities which are soon going to be extended to
> "friendly"/"allied" organisations which run exit nodes. One of the
> deliverables for some Tor sponsor is to increase the overall bandwidth
> capacity of the entire network by at least 3GB/s (roughly a 10% increase)
> within the next few years, and an outside organisation, the folks at
> TorServers.net, have offered to deal with distributing this money for us (we
> really want as little to do with running the network as possible). NoiseTor is
> very likely eligible for these extra funds, of which there is a set ammount to
> be distributed per month, and relays run in countries with a lesser
> probability of a Tor client exiting from are prioritised (and somewhat
> incentivised through extra funding). It's not exactly determined what this
> should look like yet, but the output from Compass [0] shows that a user's
> chance of exiting in the USA is about 22%. Several countries have an exit
> probability of 0%, even though they *do* have relays. I've attached the
> current output of the following commands for reference:
>
> ∃!isisⒶwintermute:(master u=)~/code/torproject/compass ∴ ./compass.py  --by-country -t -1 > ~/`date +'%Y-%m-%d'`-compass.txt
> ∃!isisⒶwintermute:(master u=)~/code/torproject/compass ∴ ./compass.py  --by-country -t -1 --json > ~/`date +'%Y-%m-%d'`-compass.json
>
> And though the factors for computing the incentivisation of lesser probability
> countries are as-yet-undetermined, I did a bit of work on this recently, the
> script is here [1], current output with the above compass files also
> attached. (Mostly because I wanted to fuck around with multidimensional
> mixed-typed arrays in numpy...) The winsorization function needs a bit of
> love, if anyone knowledgable on robust statistics and Python wants to play
> with it.

This sounds like fun.

>> > I think we should take this opportunity to double our capacity.
>> >
>> > Thoughts?
>
> tl;dr: This is my very geeky way of saying, "Please, NoiseTor, it would be
> awesome if you took Amunet. Also, there is likely additional no-strings
> funding regardless of running Amunet or not. And ♥Ⓐ♥ Noisebridge ♥Ⓐ♥."

I vote noisetor accept Amunet.

> --
>  ♥Ⓐ isis agora lovecruft
> _________________________________________________________
> GPG: 4096R/A3ADB67A2CDB8B35
> Current Keys: https://blog.patternsinthevoid.net/isis.txt