Tom Ritter <[email protected]> wrote
Mon, 8 May 2017 12:38:45 -0500:

> Anyway, so I still think there's fingerprinting concerns. But even if
> we require deterministic signatures - if the log *wants* to be
> malicious, it can still issue non-deterministic signatures, and
> there's no way to know, because as I said above - detection is hard.

You seem to focus on SCTs, which are indeed hard to share between
privacy concerned clients because they contain sensitive data. But STHs
have signatures too and I think we should limit the ways a log can track
clients asking for STHs.

Catching a log serving different log clients different bits for the same
timestamp, tree_size and root_hash is a matter of clients comparing
retrieved STHs. Keeping the requirement for deterministic signatures and
the limitation on STH issuance frequency (6962bis-24 4.8) makes
gossiping about STHs reasonable even for clients serving end users with
privacy expectations (gossip-04 10.5.4).


> So I think Eran's suggested changes are okay.

I think we should keep requiring deterministic signatures being used but
stop mandating how it is being done.

_______________________________________________
Trans mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/trans

Reply via email to