Hi Gorry, This is a vastly improved version. Minus the sections at the end where there are editor's notes and to-be-completed sections, I think you've covered the subject completely and well. Some minor issues: I'd like to consider moving the new service definitions to another draft. This draft is already long enough and it'll be easier to digest without that section. Also, creating those services raises some security issues, and I don't want to drag this draft down with those issues. Use of sender/receiver and client/server -- you tend to vary between sender/client for one side and receiver/server for the other. Since the client can be a receiver too (albeit for other messages than you're usually discussing) I feel this gets confusing. I'd prefer to see it stick to client/server; I think those terms tend to be less ambiguous. For example, I think the title of section 3.3 should be "Using Service Codes at the Server" rather than "at the Receiver". Intro, 3rd paragraph: "The use of Service Codes can assist in identifying the intended service when the server by a Middleboxes" Huh? Same paragraph: I think the middleboxes that should be most interested in service codes are firewalls. I think it's less likely to be used by NATs. So maybe reword that list to "primarily Firewalls, but possibly NATs, NAPTs, etc." If you think saying "primarily" is too strong maybe just put firewalls first. Tom P.
