Re: [mif] draft-ietf-mif-dhcpv6-route-option-04 published

[Lorenzo Colitti <[e-mail]>]

On Fri, Feb 24, 2012 at 19:38, Tomek Mrugalski wrote:
>
> >       Title           : DHCPv6 Route Options
> >       Author(s)       : Wojciech Dec
> >                         Tomasz Mrugalski
> >                         Tao Sun
> >                         Behcet Sarikaya
> >       Filename        : draft-ietf-mif-dhcpv6-route-option-04.txt
> >       Pages           : 22
> >       Date            : 2012-02-24
>

[CCing a few of the folks we had a chat with about this in Taipei]

Hi,

I believe this document is not appropriate to be discussed in this working
group, for two reasons:

1. It has almost nothing to do with multiple interfaces. Section 2 contains
text on why this document applies to "multi-homed scenarios". However,
there is nothing in this option that makes it any more applicable to a
multi-homed scenario than to a single-homed scenario.

At least 11 out of the 14 use cases cited (#1, #2, #3, #5, #6, #7, #10,#11,
#12, #13, and #14) are equally applicable to hosts with only one interface.

Further: a host needs to obtain routing information in both single and
multi-homed scenarios, and how the information is obtained (DHCPv6, RA,
HTTP, carrier pigeon) is completely orthogonal to whether the host has one
or more interfaces.

2. This option provides a complete replacement for the ND functionality
specified by RFC 4862 and RFC 4191, including default routes, more-specific
routes, and on-link determination. This is all in 6man's domain. So this
option should be taken to 6man.

I have a number other issues with the document as well. The use cases do
not seem particularly convincing; it seems to me that some of them are
niche use cases, some are circular arguments, and some are specious. The
long-term implications of configuring routing via DHCP, which is a static
protocol mediated by servers and suited to making promises, instead of
using dynamic information originated by routers, does not seem to be worth
it.

But most importantly this document should be moved to 6man so that the
people who are responsible for the architecture can see it, comment on it,
and possibly reach consensus on it.

Regards,
Lorenzo