Mysterious missing DHCPv6 feature, was Re: How does one obtain an IPv6 DNS server when VPNing to an ASA?
Michael Loftis
mloftis at wgops.com
Wed May 19 04:03:55 CEST 2010
--On Tuesday, May 18, 2010 5:35 PM -0700 Doug Barton <dougb at dougbarton.us>
wrote:
> On 5/18/2010 3:57 AM, Benedikt Stockebrand wrote:
>>
>> ... but showing up after ten or more years complaining that one's
>> existing business model isn't protected is not.
>
> I actually agreed with a lot of what you wrote, but here is where I take
> exception. I actually DID say "No one is going to deploy IPv6 in an
> enterprise network without DHCP that looks substantially like it does in
> IPv4, and supports the same options." I was shouted down LOUDLY by the
> autoconf religious zealots, so I went away. I had a lot of company in
> both regards (what I said, and not bothering to keep saying it since no
> one was listening).
I also have to chime in with a "Me Too" on Doug's statements. autoconf/RA
doesn't even come close to solving the problems that DHCP does. They're in
entirely different problem spaces, and in some orgs, they're in different
administrative domains. Not having a fully featured/feature parity with
DHCP(v4) DHCPv6 is absolutely going to hurt IPv6 deployment. Only the most
simple scenarios work with the mish-mash of autoconf, RA, and ND that is
being pushed for IPv6. And it's for the exact reasons mentioned, DHCP goes
*beyond* the connectivity bare minimums, which is VERY necessary in many
situations, and DHCP has been found to be a reliable, extendable, scalable,
and well understood method for doing this dynamic configuration.
Surprise....DHCP does what it's supposed to. Having/not having full DHCPv6
should not be left to the (core) networking people, because by and large
they're NOT the ones that care about DHCP, it's those involved with
deployment, management, and maintaining of workstations.
I think that's the root of the issue there, too many of the people who will
actually be using IPv6 haven't been involved in certain aspects of the
process that they really needed to be heard and involved in. And partly
because they don't see the problem with continuing to use IPv4(+NAT/PAT).
More information about the ipv6-ops
mailing list