Re: [arch-d] I-D Action: draft-iab-protocol-maintenance-08.txt

[Martin Thomson <[e-mail]>]

On Wed, Jul 13, 2022, at 14:03, Brian E Carpenter wrote:
> Would a state machine analysis detect possible race conditions? 

Some tools exist for doing this sort of analysis, but not all manifestations of the tool capture the details necessary to be able to detect the attack.  Some of the work from INRIA on TLS included analysis of protocol state machines: https://mitls.org/pages/attacks/SMACK

> These are not mentioned in the draft, and I think they're the sort of corner 
> case that the robustness principle should help with.

This idea that the robustness principle might step in where the protocol design falls short is exactly the sort of thing that needs to stop.  We can do better (see above), but relying on implementers to just get it right was a poor strategy even when we didn't know much better.  Collectively, we are a group experts who are equipped with the best knowledge and tooling, so there really isn't any excuse for shipping specifications that fall over when they encounter something as trivially predictable as a race condition.