Re: [AVTCORE] I-D Action: draft-ietf-avtcore-multiplex-guidelines-12.txt

[Magnus Westerlund <[e-mail]>]

Hi Gunnar, 

 

Thanks for the feedback. 

 

From: Gunnar Hellström <[e-mail]> 
Sent: den 20 juni 2020 11:38
To: Magnus Westerlund <[e-mail]>; [e-mail]
Subject: Re: [AVTCORE] I-D Action:
draft-ietf-avtcore-multiplex-guidelines-12.txt

 

Hi,

Thanks for an inspiring draft.

I have two issues:

1.) In section 3.4.2 it is said:

"RTP Considerations for Endpoints Sending Multiple Media Streams" [RFC8108
<https://tools.ietf.org/html/rfc8108> ].

I first got the impression that this would refer to a section in RFC 8108,
but there is none with that title.

And it is also not the title of RFC 8108. I suggest that you update the
string to what it was intended to refer to.

[MW] Yes, the title should really be changed. I am not certain I understand
the first part. This whole document do discuss and clarify how to handle the
aspects that are raised. So I tried to clarify that “These” really refer to
aspects and not anything else. Updated paragraph below. 

Using multiple SSRCs at one endpoint in an RTP session requires resolving
some unclear aspects of the RTP specification. These aspects could
potentially lead to some interoperability issues as well as some  potential
significant inefficiencies, as further discussed in “Sending Multiple RTP
Streams in a Single RTP Session" [RFC8108]. An RTP application designer
should consider these issues and the possible application impact from lack
of            appropriate RTP handling or optimization in the peer
endpoints.

 

 

2.) I see a contradiction between two statements:

----------------------------------------------------------------------------
----

I.) In 3.4 you say:    "Using multiple RTP streams is a well-supported
feature of RTP."

II.) In 5.2, 4.b you mention a disadvantage with using multiple streams for
the same media type in one RTP session:

"b. There is some potential for concern with legacy implementations that
don't support the RTP specification fully when it comes to handling multiple
SSRC per endpoint."

----------------------------------------------------------------------------
---------------

Can you in some way tell for which cases there is concern and for which it
is a well-supported feature. 

My interest is for the multi-party-rtt drafts, where we have so far avoided
the solution with multiple media streams of one media type in a single RTP
session saying that there may be lack of support for it in current RTP
stacks. Our current choice to instead have a mixer combine all streams into
one stream has some disadvantages and does not seem to be your
recommendation.

[MW] So the 3.4 is related to the RTP protocols capabilities as
specification. The second aspect is related to system employing RTP and what
capabilities they have required and can be expected to support. 

 

Cheers

Magnus Westerlund