RRSh> But of course the whole point is that we don't need this.. at least RRSh> not with SCTP
There is a small matter of getting 500 million hosts to convert to SCTP and then to convert all Internet applications over to it.
I think this argument can be taken too far. Yes, there are 500 million hosts out there. So long as we don't break the existing world for them, there is nothing wrong with adding new functionality.
The next question is whether the new functionality is useful until it's implemented in 500 million hosts. I claim the answer is "yes". Like any other advance it can be used in limited deployments for limited purposes until such time as it is generally available.
Imagine an application checkbox that says "try SCTP first". I think part of the SCTP transition, however, should be something more than simply modifying socket options. I would rather see our notion of services and well known ports more comprehensively addressed at the same time, so as to avoid such "checkbox" solutions in the future.
Eliot