Re: iSCSI: current UNH Plugfest: Reserved bits

[Paul Koning <pkoning@equallogic.com>]

Excerpt of message (sent 2 November 2001) by Ron Cohen:
> The problem with accepting values in reserved bits is that when the reserved
> bit is no longer reserved (evolution in the standard) it gives the
> impression to the initiator that a target implements the new feature when it
> may actually only be ignoring it.

Yes, if you don't do some other things that are needed.

You MUST have protocol version numbers in the initial messages.
Otherwise you are utterly lost.

With that, you know what version the other end has.

The key point is that requiring reserved fields to be ignored is that
you can rely on the fact that new things can be asked for in reserved
fields with the knowledge that old implementations will ignore those
things.

Quite often this is what you want.  Occasionally it is not; for those
cases you make use of the version number.

In any case, what I was describing is long established practice in
many protocols that have successfully navigated the version number
migration path.

	  paul