[Model-t] draft-thomson-tmi

[Martin Thomson <[e-mail]>]

There have been a lot of discussions recently about the threat model, but I thought that I would contribute something concrete.

This is an attempt to capture what I believe to be the principles behind some of the more prominent recent protocol design efforts.  In some ways this is a broader take than that in RFC 8558, which was a discussion of transport protocols.  QUIC very much embodies these principles, but I could point to a number of other efforts across the IETF.  The principles I recommend are:

   1.  Prefer designs without intermediaries

   2.  Failing that, control which entities can intermediate the protocol, and

   3.  Limit actions and information that are available to intermediaries

The draft is here: https://datatracker.ietf.org/doc/html/draft-thomson-tmi-00
A live HTML version is here: https://martinthomson.github.io/tmi/draft-thomson-tmi.html
And if you feel motivated to contribute, on GitHub: https://github.com/martinthomson/tmi

This might not be the right venue to discuss something like this, but as this came out of discussions in Singapore, I thought I would share it here.  This is less about the threat model than a specific response to some of the threats we've discussed; to that end, it probably isn't in the charter of this group, but I hope that it informs discussion.