[Crisp] minutes from Paris

I'm SO sorry--totally spaced on getting these out and Dave gave them to us 
like instantly! I was going to massage a bit but I think I'm late already. 
very sorry...

CRISP WG Minutes
Paris IETF

April: areg-12 was (at last minute) submitted to AD.  Got a lot of
feedback from Scott & Ted so far.

Technically this was the last milestone.  And yet, still have some
drafts.  New drafts still basically fit under charter.

Andy: [gives presentation on his 4! drafts] Update on LWZ, XPC and DCHK.

LWZ, XPC are transport drafts + common-transport.

XPC changes: bit order clarifications, other clarifications
LWZ changes: bit order clarifications
DCHK changes: added date/time elements that were missing from DCHK but in DREG

GGM: where are you in implementation?

Andy: haven't implemented these last changes but everything else.
Other's have done some implementation.

Frederico: we have initial implement of lwz, xpc and dreg.  We have
added some comments. Some more will be forthcoming

Andy asks Fred if they would have implemented if BEEP was required?
Fred: probably not.  Andy: so we did the right thing!

April: are these docs are ready for last call?
Andy: yes.

Tomoya Yoshidia: Gives presentation on routing registry. (not a wg item)

Notes that IRR data is different in that it is duplicated across
different organizations.  Motivation is to improve the accuracy of
the routing registries by using CRISP. Changes from -00: handles
routing policy, has no natural root so defined rreg.nro.org(?).
Gives example for resolving via RREG.

Root resolution requires more discussion.  Would like this draft to be
adopted.

Ted: worried about "building half a bridge".  Basically, concerned
about update mechanisms.  Obviously they are related.  Wonder if
community should take this on without taking on the update side as
well?  How is Yoshida-san keeping his data up to date?

April: maybe we should  collect questions now, and ask on the list.

Andy: have noticed increased provreg traffic.  Does AD think that an
EPP extension is appropriate?  Scott: could be.

Ted: not suggesting that this group takes on update side.

Ed: looks like this data will be sitting with the RIRs & NIRs, if the
operators are eager to register their routes with RIRs, then we should
have a protocol.  But are they?

Scott: re: provreg -- what has been happening is about moving from PS
to Draft Standard.  Trying to get feedback on missing features, etc.
Haven't gotten any yet.  Haven't seen anything that justifies a new
WG, but if there really is something, then maybe we should handle both
new f

Larry: should there be a separate root for rreg?  Seems like a
duplicate effort.  Draft still seems very rough.

William L: routing registries are run in an ad hoc way, we need to
consult ISPs to change.

Kengo: We understand that update is a serious problem, but we were
concern about the lack of structure between the routing registries.

Andre: There are problems injecting new data into routing registries,
and IRIS doesn't help with that. routing registries are intrinsically
non-hierarchical.  At least, have a number of different hierarchies
you could follow.

Cathy: Did you have a change to go the to the operator community?
Yoshida-san: we can do that.

April: next steps.  One, we accept rreg as a working group item, which
requires a charter change, so can't do this right now.

GGM: a corridor conversation with one of the RPSL authors indicated
that an XML schema for Routing info would be "cool".

Andy: had a similar conversation.  An XML version of RPSL would be
nice, and can be independent.  Ted: is this a prerequisite?  Andy: it
could be.  That would be a good idea for analysis reasons.

GGM: if it is to be useful, then there should be an xslt that emits RPSL.
Andy: that is more of a nice to have.  If people are still consuming
RPSL, what is the point?

Shane: think rreg might be a waste of time.  IRIS works well for areg
for us, but RPSL doesn't have the same sort of problems.  If there is
a structural problem, then that is an administrative problem, not
protocol work.  We can do this, but don't see it as falling under this
group.

Morishita-san: does deployment questionnaire.  Not a CRISP expert, but
we have a deployment questionnaire:

   O. who here is a RIR or TLD? [ about 12 people ]

   1. what is current deployment status of CRISP and EPP?

      CRISP [ deploying/testing: 2 people: verisign(tld), .br(tld) plan
      to introduce: 10]

      EPP [ deployed: vrsn, neustar, .au devel/testing: 2 .se, .ch/.li
      both for TLD and ENUM, planning 3, no plan: 3? 4? ]

Andy: dreg2... didn't cover everything that we thought, outright
oversights.  If we get enough maybe we can start effort of dreg2.

GGM: is this a major rewrite or minor surgery?

Andy: mostly just mucking with status codes and result objects, not
major surgery.

Marcos: IRIS versioning means that backwards compatibility is not a
great concern.

Marcos: what does it take to move from PS to draft? open-source?

Ted: two independent implementations, you do create an interop report
that list each feature, and things that don't get interop are probably
removed.

GGM: one is java, one is C++, but between ccTLD and gTLD, there might
be different policy decisions that change feature set.

Ted: also, 6 months have to pass.  This group does not have to do the
interop and leave it for a future WG.  But the interop tests do tell
you a lot.

_______________________________________________
Crisp mailing list
Crisp@ietf.org
https://www1.ietf.org/mailman/listinfo/crisp