What's our roadmap for getting ipv6 working broadly in Tor? How can we break it down, what are intermediate points where the results are useful to users, how much time will various parts take, what are the risky points that could derail our plans, etc?
We should have a solid outline of these pieces by mid March.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information
Child items ...
Show closed items
Linked items 0
Link issues together to show that they're related.
Learn more.
The technical plan in a nutshell will be "first we get bridges working in an ad hoc way on ipv6, then we get bridge address lines and descriptors able to do ipv6, then we let clients ask for ipv6 destinations, which includes changes to exit policies, then we take a step back."
We'll probably also want to point out the increasing value of ipv6 support especially for censored countries.
Sebastian reminds us that we also need to consider the metrics piece of the question, such as "what is the impact on path diversity of having only x% of the relays supporting ipv6?" Along with that question we'll need a design for a "no relays support ipv6 until a switch gets flipped in the consensus" approach so we don't deploy an anonymity-risking bottleneck at the beginning.
Relevant proposals include 117-ipv6-exits.txt and 118-multiple-orports.txt . See also the parts of tor-spec about ipv6 that are not yet implemented. See also also the thread starting at http://archives.seul.org/or/dev/Feb-2011/msg00018.html .
Looked over it and cleaned it up a bit. We'll need to make more precise plans for each of these components when we get around to actually working on them.
In the mean time, before we close this ticket, should we send the doc to or-dev so other people know it exists, and/or give it a proposal number (with status 'informative' or the like)?
Hrm. We have no actual way to declare informative proposals. We could either make one, or declare that there's a new thing like proposals that isn't. (Notes? TRs?)