[tor-dev] New Proposal 306: A Tor Implementation of IPv6 Happy Eyeballs

[teor]

Hi,

On July 11, 2019 12:37:03 AM UTC, neel at neelc.org wrote:
>I'm really sorry about the delay in responding to your review. I was 
>busy with an internship (unrelated to Tor, but still related to 
>security) and was out a lot in my "free time".
>
>I have implemented your requested changes and the GitHub PR is here: 
>https://github.com/torproject/torspec/pull/87
>
>Hopefully I have not missed anything.
>
>Most of these changes you (Iain and Teor) suggested sound good. I'm not
>
>a huge fan of preferring IPv4 in the case of tunneled IPv6 connections 
>(reason: we stay with IPv4 longer than we should), but understand why 
>you have it (reason: better network performance) and have added this 
>change anyways.

Thanks for these revisions.

I have some overall comments on the proposal:

1. The proposal is very large now. Let's add an intro section that splits the changes into:
  * initial feasability testing (for initial developer review)
  * minimum viable product (for testing in Tor Browser Alpha)
  * parameter tuning (for performance, load, and user experience)
  * relay statistics (for ongoing monitoring)
  * optional features, and how we will know if we need them (if users experience particular bugs)

2. Bridges can't be configured with an IPv4 and an IPv6 address in tor, because tor only accepts one IP address per bridge. Let's be clear that bridges are out of scope. (Tor already attempts to connect to all? its configured bridges.)

3. Each revision of this proposal has added text. Is there any text that is redundant or not essential? Can we make it shorter?

After these revisions, I will do a final review. I hope we can get another tor developer to also do a final review.

T

--
teor
----------------------------------------------------------------------