Keith asked me to review the draft; here are a few quick comments:
While the mechanics and namespaces defined seem relatively straightforward, it is hard to evaluate anything beyond the syntax and editorial issues without knowing something about what motivated the design. Why 50 namespaces and not 42 or 99? Is this based on some existing PSTN technology or is this an attempt to have 50 departments in the DoD share some resource, without declaring who is more important than the other fiefdom or assistant vice deputy director? Why priority values that are even only?
Depending on the motivation and background, there could well be other designs, such as 250 levels, where preemption is only within groups of 5. I'm not arguing that this is a better design, but it's hard to say anything even vaguely interesting without knowing the requirements.
The wording is somewhat redundant, with the same information repeated three times, but in slightly different ways. In particular, the notion of non-preemption across namespaces seems to be couched in caveats. It seems that preemption across namespaces is sometimes permitted and sometimes not, which makes it rather difficult to build an implementation for the namespace without either a lot of configuration options or special software versions for each policy. We have seen in other areas that excessive configurability leads to interoperability problems and code complexity. I have a hard time picturing how to build such a configurable system without a monster language and all kinds of strange interactions. What happens if namespace 30 can preempt namespace 28, and vice versa? What about circular preemption chains (30 -> 17 -> 10)? In those cases, do namespaces have absolute priority, i.e., any priority in 30 beats any priority in 28? You'd have to create a matrix with 250 by 250 entries
Paragraph breaks appear to be random, with a single thought split into three paragraphs (introduction).
I can't parse sentences such as "matter of local policy. However, this policy is a known requirement..." What policy? The local policy, undefined?
Henning _______________________________________________ Sip mailing list https://www1.ietf.org/mailman/listinfo/sip This list is for NEW development of the core SIP Protocol Use [EMAIL PROTECTED] for questions on current sip Use [EMAIL PROTECTED] for new developments on the application of sip
