Re: [MMUSIC] [Technical Errata Reported] RFC5576 (7544)

[Christer Holmberg <[e-mail]>]

Hi,



            I am not sure I understand your comment, nor the comment from Philipp.
            They seem to be related to some implementation, because I have no idea
            what you refer to when you say "members", "set", "count" and "field"
            :)

            Are there implementations that add multiple instance of the same
            ssrc-id value to the sscr-group attribute?

         There was one implementation that crashed when someone sent it an ssrc-group
         with multiple occurences of the same SSRC, yes. Whether this was an attack
         or a mistake, I don't know.

      It would be nice to know WHY it was sent. Perhaps they think it is a valid
      case, which may also have to be dealt with.

      >I came up with the idea of trying "what happens if" while debugging unrelated SRTP issues where

      >"is this trying to send RTX on the wrong SSRC" was the hypothesis.

      >
      >I do not think it is being used in production, in particular since for FID

      >https://www.rfc-editor.org/rfc/rfc4588#section-4

      >already says two different SSRC values MUST be used for the original stream and the retransmission stream

      >I don't see the same explicit text in RFC 5956 (for FEC-FR) sadly.



      RFC 5956 does, however, say that all SSRC values within an RTP session must be unique. So, that implicitly means you cannot use identical SSRC values for different flows, which means you would not have multiple instance of the same SSRC value within an ssrc-group.



      Regards,



      Christer











      cheers



      Philipp

      (with better quoting hopefully)