Re: [dispatch] Session recording in SIP

Hi, 

> If the original media is sRTP, does this (or not) mean sRTP 
> needs to be used to the media sink? That takes more 
> negotiation (and time). Does the sink always trust what comes 
> at it from any UA? It can be configured much more easily to 
> trust certain (limited number of) servers that can send it 
> media to record.
That is one point we adressed in
http://tools.ietf.org/html/draft-wing-sipping-srtp-key-04#page-15 by
using a setup, were the call recording is done on the media way and the
client just discloses the session key. This saves an extra encryption
between the client and the recorder.

Steffen

> 
> I could be wet behind the ears, but I don't believe the UA 
> model is low hanging fruit if the server model is specified. 
> I (think I) see each as completely separate solutions, 
> requiring however much work to do regardless of whether the 
> other is done too (other than having the recording sink 
> specified/developed -- that part is free once it is done for 
> either model).
> 
> This is an interesting problem space, and I think we should 
> look into it. I think there will be more than we expect once 
> we peel away the layers.
> 
> James
> 
> At 10:54 PM 6/11/2009, Hadriel Kaplan wrote:
> 
> >Sure there are advantages/disadvantages to each model, but 
> why bother 
> >arguing over it?  I think we probably all agree we need to 
> support at 
> >least a B2BUA model, so once you have to standardize that 
> doesn't the 
> >UA model come for virtually free?  (i.e., I would presume 
> the details 
> >of a B2BUA model are either a superset or the same as a single UA 
> >model, not less)
> >
> >-hadriel
> 
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
>