I do not see any major OPS related issues. While reading the document, I found a number of things the authors should look into: - Consider to expand SDP and perhaps CLUE in the abstract - Having both CaptureId and CaptureID in 5.1 is a bit confusing (since the two identifiers only differ by the capitalization of the last character) - Both nXML mode in emacs and xmlint indicate that the xml in section 6 is invalid. Please check. (It could also be an issue with my tools and the namespaces but then also the indentation looks at least somewhat surprising. - Is the RFC editor expected to replace XX in the drawing in section 5.1 with the value assigned for TBA? If so, I think this needs to be documented somewhere. - Is 'roni.even@mail01.huawei.com' is a long term stable identifier for the 'Contact' field of the RTP SDES Compact Header Extensions subregistry? - Security considerations, last paragraph: What is 'a lot of trust'? Why is the SHOULD not a MUST? - s/CaptureIDis/CaptureID is/g - According to idnits, there are RFCs listed in the references that are not cited in the text; please pay attention to idnits reports