Re: [dhcwg] DHCPv6 router option

[John Jason Brzozowski <john_brzozowski at cable.comcast.com>]

Hemant,

I am certain I am qualified to reply here given my role leading Comcast's
IPv6 deployment.  The deployment of IPv6 in cable can proceed without a
default router option, I mentioned this in the INT area meeting and I will
repeat it here now.

I cannot speak for other MSOs.  I will likely investigate this further and
encourage others to speak up as needed.

I like yourself have spent an extraordinary amount of time over the past
several years ensuring that cable IPv6 implementations are deployable.  This
work has ranged from work on cable modems to CMTSs and provisioning systems
and then some.

Finally, I will also state that I have no issue with the this option being
specified as long as it done with proper context and attention to detail.

John
=========================================
John Jason Brzozowski
Comcast Corporation
e) mailto:john_brzozowski at cable.comcast.com
m) 609-377-6594
=========================================


> From: Hemant Singh <shemant at cisco.com>
> Date: Mon, 23 Mar 2009 19:18:02 -0400
> To: Bernie Volz <volz at cisco.com>, Thomas Narten <narten at us.ibm.com>
> Cc: dhc WG <dhcwg at ietf.org>, Ted Lemon <Ted.Lemon at nominum.com>, Ralph Droms
> <rdroms at cisco.com>
> Subject: Re: [dhcwg] DHCPv6 router option
> 
> Ralph,
> 
> Besides Alain Durand who has been asking for this potentially broken
> idea to add a router option in DHCPv6, who else is a SP who asked for
> this work? Alain also does not represent other cable vendors like
> Time-Warner, or Cablevision etc.  For that matter, Alain himself says "I
> don't even represent Comcast", so who else from Comcast can speak up and
> say this work is needed?
> 
> Let me also bring folks up to speed on Cable IPv6 standards from the
> CableLabs.  We finished our Cable IPv6 standards more than three years
> back and consumer devices like cable modems already follow the RA model
> to get default rtr info and use DHCPv6 as well for ipv6 address
> acquisition and other parameters like DNS etc.  A consumer device like a
> cable modem has a huge inertia to changing any feature.   Likewise a
> CMTS router that terminates the cable modems, being a Service Provide
> box, has its own inertia to changing core features.  I spent two years
> of my life between 2005=2007 nursing cable modem vendors to fix very
> simple and core IPv6 bugs in their chips and I do consider this draft to
> be rocking that boat.  The Provisioning system also changes. Then I and
> Wes spent time finding show stopper bugs in hosts behind cable modems
> and discovered that most operating systems were broken in some core data
> forwarding behavior for IPv6 for a specific RA.  I do not like any IPv6
> boat rocked even an iota unless there is compelling justification.
> 
> So let's not go backward.  Alain and Ralph, please explain what is the
> problem in the cable broadband network that this draft is needed?
> Please be super specific as to what the problem is.  Sorry, if the good
> reasons have been given to the mailer - I have not caught up to mails on
> this thread yet.  One year back when we discussed this issue with Alain,
> I didn't hear any compelling reason.  What has changed since then?
> 
> Thanks,
> 
> Hemant
> 
> 
> _______________________________________________
> dhcwg mailing list
> dhcwg at ietf.org
> https://www.ietf.org/mailman/listinfo/dhcwg