Re: [splices] draft-loreto-splices-disaggregated-media-00

[Salvatore Loreto <[e-mail]>]

Hi John,

On 11/26/10 1:10 PM, Elwell, John wrote:
> I still don't like the fact that there is impact on Bob's UA.
supporting new features most likely implies an impact on the UA(s) that 
wants support them.
anyway the features are optional!
> Even though that impact might only be session correlation, this could be quite complex.
if we are able to standardize a good and hopefully general session-id 
mechanism,
then the impact coming from SPLICE is really minimal.

/Sal

> It sounds to me like Bob's UA, on deciding that two SIP dialogs relate to the same session, must somehow merge the two session descriptions (SDP). In the simple case, where one session is audio only and the other is video only, it might be relatively simple, e.g., an assumption that the two need to be lip-synched could be made. But we surely would need to deal with more complex cases, e.g., an additional video stream for presentations, floor control of the presentation stream or other streams, multiple audio and/or video channels for telepresence, grouped media spanning more than one contributing session description, etc.. Multimedia SIP UAs will have no incentive to implement extensions to cover session correlation and merging, so this will constrain interoperability.
>
> John