Re: [dhcwg] draft-chowdhury-dhc-bcmcv[46]-option-01.txt

On Wed, Nov 24, 2004 at 04:46:56PM -0500, Ralph Droms wrote:
> There is a deployment issue, as some service providers already have
> DHCP servers in place that must be updated for any new options.  Is it
> the case that the options defined in the current drafts can be
> supported without code changes to existing servers?

Speaking for ISC DHCP;

Of draft-chowdury-dhc-bcmcv4-option-1, section 4.2 can be supported
easily today by our server.  Section 4.1 could be made to perform as
documented in the draft, without modifying code, but the configuration
syntax would be somewhat awkward.

> Use of a VIVSO
> sub-option would require code changes to existing servers.  How long
> would it take to deploy DHCP server that can support VIVSO.

Versions of ISC DHCP Server currently in wide deployment could
support VIVCO/VIVSO today, with some trickier than usual
configuration (oddly similar to what would have to be done for
section 4.1).

With one disclaimer:  It would not work well at all for multiple
vendor ID's.

In order to support multiple vendors, you would have to statically
configure responses for every possible combination of VIVCO contents
that you expect to receive (either the VIVCO contents would have to be
consistent on every boot, which when I glanced I think is not mandated
by the RFC, or you would have to interpolate every possible combination
and every subsequent response and provide them all in configuration).

So, I think it could be made to work, but the level of difficulty does
present a significant barrier to adoption, not to mention the lack of
scalability to multiple enterprise id's, which makes this approach,
at least for ISC DHCP, less and less appealing the more vendors get
on board.  This makes it seem unfair to ask 3GPP2 to go this route
when it would be less likely that we would ask anyone else to follow
them.

Beyond this, ISC has not, today, laid down the schedule framework for
the next feature release of ISC DHCP which may or may not contain
formal support for RFC3925.  So, I cannot accurately answer the question
of how long it would take to reach deployment, which suggests we should
estimate my answer to be an infinite value.

So I would promote the approach of the current v4 draft.

-- 
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