Re: [v6ops] PCP server in draft-ietf-v6ops-6204bis

[james woodyatt <[e-mail]>]

On Jan 30, 2012, at 12:59 , STARK, BARBARA H wrote:

> The reference to RFC 6092 is in RFC 6204, which was published prior to homenet’s creation.
>  
> Unless the reference to 6092 is truly an error, then we don’t need to be discussing changes to its recommendation (removal, updates, enhancements) in 6204bis. It continues as it was. 6204bis has tried to leave LAN elements of 6204 untouched. They were fully and appropriately discussed with consensus agreement to include them, at the time of 6204. And so they remain, unless there is a problem.
>  
> If you believe that inclusion of 6092 as a “SHOULD” is truly an error, then I would like to understand why you think that. If there exists proof that it’s harmful (I’ve heard some people say that having it enabled by default has caused some problems, so perhaps there is such proof), then maybe we should reconsider its inclusion. If you’re suggesting its removal because the topic is no longer under the charter of v6ops, then I disagree. We didn’t say that we wanted to pull out LAN elements when doing 6204bis. We said we didn’t want to do anything with them, at all (unless they have proven to be a true mistake). Removing the reference to 6092 would be inconsistent with an attempt to make no “LAN functionality” changes to 6204, when creating 6204bis. I would only support it if it were shown to be truly harmful.

If I were to make a technical argument about the harmfulness of RFC 6092, I would start with a lengthy review of its Security Considerations, in section 6.

I think it's a debatable question of procedure, whether it was an error at the time for V6OPS to publish RFC 6204 with its S-1 "requirement" to recommend implementation of RFC 6092.  There was a clamor for the timely publication of a document, and neither HOMENET nor PCP were remotely close to usable when RFC 6204 was in WGLC.

I do think it shouldn't be controversial *now* to say that HOMENET is the venue where any further IETF recommendations about the implementation of RFC 6092 in IPv6 CE routers should originate.  If you disagree, then once again, I have to ask for an explanation of your reasoning.  I'm very interested to know.

If V6OPS wishes to continue stepping on HOMENET's charter, on the grounds that it already has its footprints there from before HOMENET was launched, then I suppose that's fine— but, if so, then I don't see why V6OPS shouldn't also take this opportunity to revise RFC 6204 to recommend a PCP server to meet REC-48 in RFC 6092.  Again, as a purely technical matter— I know, technical matters are so boring— IETF has a standards-track protocol for ""host applications to solicit inbound traffic without advance knowledge of the addresses of exterior nodes with which they expect to communicate" as RFC 6092 recommends.  We did not have this protocol when either RFC 6092 or RFC 6204 were ready for publication, but we do now.  I fail to see any technical basis for leaving this unspecified in RFC 6204bis.  Is there one?

Alternatively, if this is primarily a procedural dispute, then we could open RFC6092bis in order to update REC-48 accordingly, then cite the RFC6092bis in RFC6204bis with a publication in lockstep.  Which would you prefer?


--
j h woodyatt <[e-mail]>