[avtext] Proposed RTP Taxonomy updates

[Bo Burman <[e-mail]>]

Dear all,

Based on my interpretation of the mailing list discussions, these are the edits that are planned for the next version:

*         Re-structure the entire document such that alternate usages, describing existing terms using Taxonomy wording, are separated out into their own (alphabetically sorted) sub-sections, making it possible to find them through the table of contents

*         Remove any sub-section level 4 numbering possibly remaining after the above change

*         Change the term "Packet Stream" to "RTP Stream" throughout the document, since "Packet Stream" is too unspecific

*         Remove all topologies discussion in section 4 to avoid any conflict or multiple description of topologies, even if informational.

*         In the above, keep one informational figure that depicts one example of the terms Communication Session, Multimedia Session, RTP Session, and End Point, and move that figure, as an example, to the Concepts sub-section on Communication Entities.

*         Clarify in Synchronization Context section that CNAME is not an End Point identifier

*         In that same section, point out that although RTP uses NTP-format timestamps, they don't need to be NTP synchronized to be useful. Also reference draft-ietf-avtcore-clksrc.

*         Add a new relation section with an explicit explanation of RTP duplication as of RFC 7198, preferably with a figure.

*         Remove the statement that a CLUE Media Capture is identified via indexed notation, since it is incorrect.

*         Make it clear in the characteristics of RTP Stream that the SSRC identifier of a single RTP Stream can change, but only as a result of SSRC collision or a clock rate change (referring RFC 7160)

*         Clarify that multiple 6190 packet streams from a single Media Source can be present in a single RTP Session, despite what is said in 6190 section 1.2.2

*         Remove the current discussion on and definition of MST-SS/MS and SST-SS/MS

*         Mention that an "End Point" is typically a single "host"

Unless there are objections, I plan to submit an updated draft including the above changes during next week.

Comments and questions are welcome!

Cheers,
Bo