Re: [dhcwg] Comments on draft-ietf-dhc-3315id-for-v4-03

On Wed, Aug 11, 2004 at 06:34:13PM -0700, Ted Lemon wrote:
> different IP addresses.   There is no harm in this, unless you are 
> concerned about excessive IP address consumption.   If you are 

I'm really confused by this statement, because I thought I documented
the harm.

Go back and read what I said about Win XPe please?  I never answerd
your other mail because your recommendations about alternative actions
equally surprised me...they won't work.  You also seem fixated upon
having the PXE client send 'a different client identifier' when in fact
the problem is that it doesn't send that option at all, and the problem
would be made worse if it ever did.

So there's quite a lot of confusion floating around already.

To briefly restate, an undocumented behaviour which fixes this problem
without user intervention or new PXE clients has arisen in the majority
of DHCP servers.  At least one client has arisen (Win XPe) that relies
upon this behaviour in order to boot...at all.

So it is no longer about over-consumption of addresses.  That issue, and
the way that those servers choose to solve it for their paying customers,
is what has led us now to this impasse.

> concerned about that, there are stopgap solutions to get you over the 
> hump until the last of those PXE cards dies a well-deserved death.

That will not happen.

I don't think that any iteration of 'dumb bootstrap device' will
ever learn anything more complicated than it absolutely has to, for
darn good reasons.  And I really doubt that Microsoft will ever make
something like a client identifier a user-configurable option (like
Apple does).

-- 
David W. Hankins		"If you don't do it right the first time,
Operations Engineer			you'll just have to do it again."
Internet Systems Consortium, Inc.		-- Jack T. Hankins