Re: [MMUSIC] BUNDLE Weekly Summary: Assumptions

[Christer Holmberg <[e-mail]>]

Hi,

>>>> Q4:
>>>>
>>>> We need to keep in mind that there is always a chance that the "top-most" m- line within a bundle group is a port zero one.
>>>
>>> What is the question?
>>
>> Well, there are basically two "warm-up" questions:
>>
>> Q4_1: When the first offer, with DIFFERENT port numbers is sent, and the receiver accepts bundle, is the receiver expected to be able to send media/data according to that?
>
> I hope so.

Ok, so let's stick to that assumption for now.

>> Q4_2: IF Q4_1 THEN which m- line does the receiver use to send media/data (when the second offer comes, with identical port numbers, it is no longer an issue).
>
> Surely the bundle draft must specify this.
>
> My first thought is that the address/port of the *first* m-line in the bundle with a non-zero port will be used for the entire bundle.
>
> (This means that when you initially offer a bundle with a bunch of separate ports, you not only must be prepared to use each of those for an unbundled m-line, you must be prepared to use any of them for the complete bundle.)

Yes, that is what I have been thinking also: until the second offer has been sent and accepted, the offerer has to be able to receive media on any of the offered ports.

Because, the offerer cannot assume that media will come on the *first* non-zero m-line in the offer, because the answerer could  reject it. 

OR, the answerer could accept the m- line but exclude it from a bundle, in which case it will only send media associated to that m- line to the offered port - and the bundled media will be sent to some other port of the offered bundle group.

These are issues that I've had from the beginning with offering different ports in a bundle group, so I would really like the people (Cullen, Richard E etc) that pushed for this to explain me how they intended to solve it  :)

Regards,

Christer