Re: [v6ops] Fwd: New Version Notification for draft-ietf-dhc-pd-exclude-04.txt

[jouni korhonen <[e-mail]>]

Lorenzo,

On Dec 20, 2011, at 9:28 PM, Lorenzo Colitti wrote:

> On Mon, Dec 19, 2011 at 23:49, jouni korhonen <[e-mail]> wrote:
> We added the applicability statement into the Introduction, which now states the pd-exclude is intended for deployments where CEs are on their own layer 2 domain. This should clear your concern on 'multicast RAs".
> 
> As discussed already, I have two objections to this on the grounds that it needlessly complicates implementations.
> First, I object to the text that says "The requesting router must create sink routes for the delegated prefixes minus the excluded prefixes." This is complex to implement properly.

I am not sure I agree. Can you clarify why it is so complex to implement properly? Ole already responded to this in http://www.ietf.org/mail-archive/web/v6ops/current/msg11526.html

> I think it should say: "The requesting router must create sink routes for the entire delegated prefix. If it is desired that the requesting router route the excluded prefix in any other way, this MUST be configured to do so via other means (for example, via a Router Advertisement containing a Prefix Information Option for the excluded prefix and the on-link bit set).

Why not proposing the exact text you want to see in the draft?

> 
> Second, I think the draft should make it clear that only one prefix can be excluded per delegated prefix, and that if 

Section 4.1 says  "There can be at most one OPTION_PD_EXCLUDE option in one OPTION_IAPREFIX option". I say it is clearly stated.

> the RR receives (at any time), another OPTION_IAPREFIX option for the same prefix with different excluded prefixes, then the RR is free to ignore the previous excluded prefix.

Why would a DR do such? I do not see such scenario. No problem to state this though.

- Jouni