Derek Atkins <[email protected]> 写于 2012-09-10 21:23:12:

> Hannes Tschofenig <[email protected]> writes:
> 
> > I am sure that we can come up with many different protocols; the 
> area of key agreement protocols isn't necessarily a new one. 
> >
> > (What by the way is "H(R)" standing for?) 
> 
> I'm pretty sure he means Hash of R.  E.g. you send the SHA-1 Hash of R
> as a commitment of R, and then later you send R.  But see my previous
> message, because this *requires* the AS be involved in EVERY request.  I
> don't think that's a reasonable architecture.  Tokens should have a
> validity period and the client should be free to continually use the
> token without going back to the AS during that period.  Otherwise the AS
> becomes a single point of failure and a bottleneck.

Concerning access token is time limited from a few minutes to 1 hour, 
and R is transported in TLS, it is not necessarily used only once. 


> 
> -derek
> 
> -- 
>        Derek Atkins                 617-623-3745
>        [email protected]             www.ihtfp.com
>        Computer and Internet Security Consultant
> 

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

Reply via email to