Re: [AVT] SRTP with layered encoding

["Jonathan Lennox" <jonathan at vidyo.com>]

Ah, I see, yes.

It still feels somewhat like a layer violation to me (the SRTP stack
needs to know that it's processing a layered encoding), but I see why
it's different than multimedia.

I think this would make sense only for EKT-type keying mechanisms, i.e.
where session keys are bound to SSRCs, not for SDesc or DTLS/SRTP where
a single session key applies across all SSRCs of an RTP session.  Yes?
In the latter case, the fact that SSRCs are shared doesn't make any
difference.

-- 
Jonathan Lennox
Vidyo, Inc
jonathan at vidyo.com


> -----Original Message-----
> From: Marc Petit-Huguenin [mailto:petithug at acm.org]
> Sent: Tuesday, October 20, 2009 12:36 PM
> To: Jonathan Lennox
> Cc: avt at ietf.org
> Subject: Re: [AVT] SRTP with layered encoding
> 
> Jonathan Lennox wrote:
> > What I guess I don't see is why layered encoding should be treated
> any
> > differently than "ordinary" multimedia, e.g. separate audio and
video
> > sessions.
> 
> They share the same SSRC, so they are treated as a whole (i.e. BYE and
> non-CNAME
> SDES only on base layer).  I thought that it made sense to also apply
> SRTP (and
> EKT) to this sessions as a whole.
> 
> --
> Marc Petit-Huguenin
> Personal email: marc at petit-huguenin.org
> Professional email: petithug at acm.org
> Blog: http://blog.marc.petit-huguenin.org