Stateless vs. Stateful

["Hollenbeck, Scott" <shollenbeck@verisign.com>]

As part of the ongoing discussion of our IETF last call comments, Jaap, Ed,
Patrik and I have been talking about the issue of a stateless vs. stateful
protocol.  Right now EPP has some provisions to operate both ways (see the
description of "sessionless" operation in the core spec).  Patrik, our AD,
has suggested that the protocol needs to be either stateful or stateless,
but not both, and after thinking about things some more I agree with him.
Believing that we agreed long ago that a stateful session-oriented protocol
is the more desirable operating mode, the stateless "sessionless" stuff is
the candidate for extraction.

I originally put the "sessionless" stuff in the protocol thinking that it
might be useful for a low bandwidth, high latency transport like email.
However, it would also be possible to keep the core protocol completely
stateful and session-oriented (and thus a bit simpler) while still working
in such an environment if the transport mapping were defined appropriately.
With email, for example, a transport mapping could be written to say that
all of the needed commands must be sent in one message/session.

Anyway, to get back to the point I think we need to remove the "sessionless"
features from the core protocol to get the documents ready for the IESG.
What that would mean is the removal of the optional credentials elements
from the commands, and changing the corresponding text.

Does this cause any issues for anyone?

-Scott-