> On Feb 5, 2017, at 12:26 AM, Ron Garret <[email protected]> wrote:
> 
> 
> On Feb 4, 2017, at 1:53 PM, Nadim Kobeissi <[email protected]> wrote:
> 
>> Forward secrecy relies much more on SPKs than OTPKs. Rather, OTPKs are there 
>> to provide some notion of “freshness” to a authenticated key 
>> exchange/agreement, so that two successive sessions between two people 
>> aren’t more stale on the shared secret front due to SPK and identity key 
>> re-use.
> 
> I thought that was what Alice’s ephemeral key was for?
> 
> Actually, I think I’ve figured it out.  I had this idea in my head that the 
> OTPKs were like nonces, that re-using them would compromise security, that is 
> why the server had to be careful to only hand them out once (hence the OT 
> part of OTPK).  But that is not the case at all.  The OTPKs are OT because 
> Bob is going to destroy the corresponding secret key as soon as he receives a 
> message encrypted using a particular OTPK in order to close the window of 
> opportunity for an adversary as quickly as possible.  So it actually does 
> make sense to look at them as an add-on to X3DH and not as a completely 
> separate protocol.

Yes, that’s correct.

> 
> This still leave a couple of operational concerns (OTPKs must be refreshed 
> fairly often, and I think they need an expiration date).  But it not the case 
> that a server compromise leads to a security breach.  (The PK in OTPK should 
> have been a clue!)

Yup!

> 
> Thanks for the help,

You’re welcome!

> rg
> 

_______________________________________________
Messaging mailing list
[email protected]
https://moderncrypto.org/mailman/listinfo/messaging

Reply via email to