Re: [tram] FW: New Version Notification for draft-reddy-tram-turn-third-party-authz-00.txt

[Oleg Moskalenko <[e-mail]>]

Hi Tiru

regarding the authentication mechanisms: yes, I am talking about section
10.1 of STUN RFC 5389:

http://tools.ietf.org/search/rfc5389#section-10.1

STUN defines both short-term mechanism and long-term mechanism. They differ
in the ways how the INTEGRITY field is calculated, and in some other minor
details. The names are somewhat unfortunate and confusing; the "short-term
mechanism" is NOT about short-lived ephemeral identities. The proposed
token-based authorization is equally applicable to both short-term and
long-term mechanisms - taking into account different INTEGRITY calculation
algorithms.

>From the pure technical point of view, I think that the token authorization
is equally applicable to the short-term mechanism than to the long-term
mechanism.

Of course the short-term mechanism is not very important in this context
because this proposed standard is mostly about WebRTC. But it would be
"cleaner" if the token authorization will be applied to the whole STUN/TURN
specs, not to a particular case of WebRTC. And may be some day WebRTC will
decide to adopt the short-term mechanism...

Regards,
Oleg