Re: [dhcwg] draft-despres-intarea-4rd comments

[Tomasz Mrugalski <[e-mail]>]

On 11-04-01 09:41, Rémi Després wrote:
>> First, I'm sorry if my comments may seem a bit harsh, but the currently
>> proposed option format is catastrophic. It is basically several
>> suboptions stuffed together with suboption-like type and length coded on
>> nibbles, rather than usual 2 byte fields, with some suboptions being
>> optional. These parameters should clearly be defined as separate options
>> (possibly grouped together as suboptions within an option for clarity).
> 
> Thanks for expressing a view, this was the goal of this presentation.
> No catastrophy ahead, there is enough time make proposal that reach consensus in the WG.
I would like to withdraw my comment and apologize for using such strong
words. What I really meant is that such complicated format may easily be
misimplemented, that could even lead to client crashes. While it has
benefits of being very compact, we have agreed that following standard
DHCPv6 format would be better way to move forward as long as all
currently defined parameters are delivered to clients.

>> There is no mention of DHCPv6 related problems in Security
>> Considerations section whatsoever. This should be analysed a bit and
>> commented in the text (possibly briefly if there are no new issues here,
>> with just pointers to existing RFCs).
> 
> No 4rd specific security concern about dhcpv6 ahd been identified, and RFC 5969 on 6rd has no such mention.
Thank you for clarifying that. I'll add appropriate comment to that effect.

>> Also, I have a procedural recommendation. I happen to have worked on
>> DHCPv6 option for another IPv4-IPv6 coexistence architecture. The actual
>> architecture and the DHCPv6 operation were specified in separate, but
>> closely related documents. It worked really well, because discussed
>> problems and review groups are mostly orthogonal. They could also
>> possibly follow separate adoption paths (or perhaps not, but that is
>> something that should be possibly considered).
> 
> 
> Doing it earlier might have been more complex than needed, but doing it for a later version does make sense to me.
> A volunteer for te DHCPv6 draft will be welcome (I am clearly not an expert of this WG).
I have volunteered to be an editor of this draft and discussed this with
Remi, who accepted my proposal. If others are interested in this, please
contact Remi and myself.

Tomek Mrugalski
ISC