Re: [AVTCORE] RFC5285bis open issues

[Roni Even <[e-mail]>]

Hi Magnus,
Thanks,
I will add the text about the id=0 len>0 using option b

As for support for one-byte only I think that since section 4.1.2 says " the one-byte header form is preferred and MUST be supported by
   all receivers" it provides guideline to use one-byte saying that one-byte will work but two-bytes may not. Do you have another proposed text to address this point?

Roni

> -----Original Message-----
> From: Magnus Westerlund [mailto:[e-mail]]
> Sent: יום ב 20 פברואר 2017 11:47
> To: Roni Even; [e-mail]
> Subject: Re: [AVTCORE] RFC5285bis open issues
> 
> Hi,
> 
> Personal opinions, not responding as WG chair and doc shepherd.
> 
> Den 2017-02-19 kl. 16:06, skrev Roni Even:
> > Hi,
> >
> > Based on the email discussion I see some issues that need conclusion
> >
> >
> >
> > 1.       For the one-byte RTP header extension if id=0 and len>0 what
> > action to take since there is no current text
> >
> > a.       This is also padding
> >
> > b.      This is an error so stop interpreting following header
> > extensions – similar to id=15
> >
> 
>  From my perspective, RFC 5285 is very clear that padding is a byte with value
> 0, i.e. all 8  bits are zero. Also, sending for example 0x04 is very undefined
> and a sender could expect b) option response from the receiver. Without
> agreed interpretation I rather send a clear message, don't use this. Thus, I
> would favour option b).
> 
> 
> > 2.       Can we conclude from RFC5285 that if only IDs between 1-14 are
> > offered it means that only one-byte RTP header extension are supported?
> >
> 
> I think I have been involved in discussion about this previously. The
> conclusion I remember is that it is not possible. It is allowed to negotiate IDs
> below 16 and still use the two-byte header format for a particular RTP
> stream. So, it is not a conclusion we can draw.
> 
> However, the reality was that until we created the SDES header extension
> we didn't have any header extension that had the possibility to be more than
> 16 bytes.
> 
> At least if I remember correctly how the definition of the various registered
> header extensions.
> http://www.iana.org/assignments/rtp-parameters/rtp-
> parameters.xhtml#rtp-parameters-10
> 
> But, there could exist proprietary header extensions that are longer.
> 
> So, not guaranteed, but a fairly safe bet assuming no header extensions that
> are known to allow longer than 16 bytes. But, I think one has to make it clear
> that there are no guarantees, but one could write a guideline for a case that
> is likely to work.
> 
> Cheers
> 
> Magnus Westerlund
> 
> ----------------------------------------------------------------------
> Services, Media and Network features, Ericsson Research EAB/TXM
> ----------------------------------------------------------------------
> Ericsson AB                 | Phone  [phone]
> Färögatan 6                 | Mobile [phone]
> SE-164 80 Stockholm, Sweden | mailto: [e-mail]
> ----------------------------------------------------------------------