On 25 July 2015 at 05:26, Ben Laurie <[email protected]> wrote:
> I'm pretty sure I didn't say that. Not that I am particularly opposed to it,
> my position is simply that I am not sure what value it adds.

Sorry, I don't know where it came from then.  And agreed.

> BTW, I snipped all your comments about SCT gossip, but I do agree with your
> characterisation of the recipient as some kind of Trusted Auditor. I'm not
> sure I saw gossip between Trusted Auditors explicitly addressed - would that
> be allowed, or dangerous?

As long as the Trusted Auditor didn't act in a predictable way with
regards to data fed by a client, I believe it is safe for them do SCT
Feedback to servers, STH pollination to servers, or send their results
directly into a whole different Trusted Auditor.  It's only dangerous
if I can observe a Client <-> Trusted Auditor connection, and make
strong inferences about what happened on that connection.  So, clients
inputs should be delayed/mixed before being acted on, and Trusted
Auditors should be _encouraged_ to both collect as many users as they
can (including other Trusted Auditors), and to do SCT Feedback and STH
Pollination as much as they can.

-tom

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

Reply via email to