[16:23:12] --- anewton has joined
[16:26:15] --- Hollenbeck has joined
[16:26:27] --- Hollenbeck has left
[16:27:28] --- Hollenbeck has joined
[16:31:32] --- geoff has joined
[16:34:00] --- geoff has left: Disconnected
[16:34:07] --- geoff has joined
[16:36:34] --- dcrocker has joined
[16:37:13] --- amarine has joined
[16:39:07] --- robertml has joined
[16:39:17] <anewton> george: starts the meeting
[16:39:45] --- mjo has joined
[16:39:46] <anewton> george: doc status
[16:40:13] <amarine> did ggm get a minute taker?
[16:40:50] <robertml> Kim Davies does the minutes.
[16:40:52] <anewton> yes. kim davies
[16:41:00] <amarine> merci. i wil not be quiet
[16:41:04] <amarine> uh "now"!!
[16:41:26] --- CathyM has joined
[16:41:28] <anewton> fred: <goes to present areg>
[16:41:31] <robertml> unless if you have comments :)
[16:42:38] <anewton> fred: findOrganizations in 3.1.2, make it similar to 3.1.1.. create new section 3.1.10 organization serach group
[16:42:54] <anewton> fred: 3.1.5 find network by specificy to by handle
[16:43:30] <anewton> fred: drop areg-urires-01. add new URI res section to areg draft. top down res still recommended but top is name lookup
[16:43:44] <CathyM> 3.2.3 added address,city,region,postal code and country - was already in xml scheme, only had to add to text
[16:44:07] <anewton> fred: <reading proposed text for section 7
[16:44:10] <anewton> >
[16:44:15] --- ggm has joined
[16:45:10] <anewton> fred: need anchor because addresses are different than domains
[16:45:39] <anewton> fred: talking about text Andy sent to the mailing list
[16:46:43] <robertml> If a server doesnt know about the address it is asked about, it will refer to the anchor
[16:46:46] <CathyM> think this is it:
[16:46:49] <CathyM> 7.1 Application Service Label The application service label associated with AREG registry type MUST be "AREG1". This is the abbreviated form the URN for this registry type, urn:ietf:params:xml:ns:areg1. 7.2 Operational Considerations Address registries do not have natural links to DNS. Using reverse DNS tree presents problems for IP address delegation (for example, delegations do not fall into byte boundaries, unlike reverse DNS), and DNS does not currently contain any information regarding autonomous system delegation. Therefore, in order for top-down resolution to operate properly, it is requested that the IAB include an AREG1 S-NAPTR record for the the name areg-iris.arpa. 7.3 Top-Down Resolution The top-down alternative resolution method MUST be identified as 'top' in IRIS URIs. The process for this condition is as follows: 1. The IRIS [X] direct-resolution process is tried against areg-iris.arpa. 2. If the direct-resolution process yields no server for which a connection can be made, then a negative response is returned and no further action is taken. 7.4 Discussion It is RECOMMENDED that IRIS clients issuing AREG1 requests use the 'top' resolution method when no resolution method has been explicitly given by a user. IRIS servers accepting AREG1 requests seeking information for which they are not authoritative SHOULD refer clients using the 'top' resolution method.
[16:47:28] <anewton> leslie: text could be a little different, but we can wordsmith it. gives example.
[16:47:44] <anewton> leslie to send text
[16:49:14] <robertml> (there's at least one remote participant, so please use mikes)
[16:49:50] --- levigner has joined
[16:49:51] <robertml> ggm says it all looks like minor fixes to the text.
[16:50:32] <robertml> next 40 minutes are draft-newton-crisp-iris-xpc-00 (20 min) and draft-kengo-crisp-iris-rreg-00 (20 min)
[16:50:58] <robertml> Andy talks:
[16:51:18] <robertml> XML Pipelining with Chunks (xpc)
[16:51:51] <robertml> Andy thinks BEEP is too complicated to such a simple job as IRIS.
[16:52:11] <robertml> .. and insufficient lib-support.
[16:53:03] <robertml> Two options... use something existing, or build something new.
[16:53:47] <robertml> HTTP lacks SASL and chunking.
[16:54:44] <robertml> (at least functionally chunking is not "in http")
[16:54:56] <robertml> So Andy decided to invent a new protocol
[16:55:21] <robertml> xpcl is so simple that it can be built in a day.
[16:55:23] <robertml> xpc
[16:55:52] <robertml> Has SASL.. or SSL/TLS via port wrapping
[16:57:31] <robertml> For normal XML-usage, you need the entire XML-text in memory, before it can be handled.
[16:58:08] <robertml> Pipelining cuts the xml into chunks
[16:58:25] <robertml> The server processes one chunk at a time, without having received the entire request.
[16:58:54] <robertml> xpc complements lwz
[17:00:01] <robertml> Brian, CERT/NSA: is xpc crisp-specific or can it be used for any XML-app?
[17:00:17] <robertml> Andy: it can probably be used by most applications.
[17:01:13] <robertml> me: what kind of requests would use xpc?
[17:01:57] <robertml> andy gives an example.
[17:02:26] <robertml> (if the server doesn't list an additional section)
[17:03:07] <robertml> Adopted by the WG
[17:03:38] <robertml> nagahashi-san (sp?) will talk about rreg
[17:03:46] <robertml> sic
[17:03:54] --- bhoeneis has joined
[17:04:19] <ggm> Nagahashi Kengo,
[17:04:42] <robertml> background Depends on whois-protocol. Lack of referral transaction.
[17:04:55] <robertml> ^^^ current query transaction of IRR
[17:05:21] <robertml> Typical operation of IRR: NRTM (near real time mirroring).
[17:05:32] <robertml> Why not real time?
[17:05:45] <robertml> Current whois based protocol does not support referral query transactions.
[17:07:14] <robertml> current status of draft: initial draft has been published... based on the areg doc
[17:07:26] <robertml> Would like to discuss the draft on the ML.
[17:08:23] <robertml> Will the WG adopt it?
[17:08:31] --- yone has joined
[17:08:43] <robertml> (something in japanese which probably means End Of Slides)
[17:09:05] <robertml> Question from Djengis (sp?)
[17:09:16] <robertml> Does CRISP do the real time mirroring?
[17:09:26] <mjo> Cengiz
[17:09:30] <robertml> ggm says it's out of scope in itself
[17:09:31] <robertml> thanks
[17:10:12] <robertml> Andy: points out that there is a serialization way, to support mirrorring
[17:10:51] <robertml> cengiz: please have the IRIS include actual mirroring (??)
[17:11:17] --- klensin-ietf has joined
[17:11:47] <robertml> CathyM: Thisbrings up another point.. what's the resolution method? We just got through how to do top-down. There's a natural structure for dreg and (now) areg and ereg.
[17:12:00] <robertml> What's the way to know where to look?
[17:12:28] <robertml> kengo: probably like areg
[17:13:38] <robertml> ggm: Joao Damas, ISC, couldn't come, but relayed the message that the draft only included information models from areg.
[17:13:52] <robertml> Would be better to state how to present routing in XML.
[17:14:31] <robertml> The rreg is quite different from areg, dreg. It is not given that it's part of crisp.
[17:14:39] <robertml> because the behaviour is different.
[17:15:53] <robertml> Andre, RIPE NCC, relayed via ggm, that irr-toolset is used by scripts which use rpsl.
[17:16:21] <robertml> (I don't get it.. someone?)
[17:16:34] --- yone has left: Disconnected
[17:16:35] --- yone has joined
[17:19:11] <robertml> hollenbeck: groups are welcome to take new work if it makes really sense, but dont forget existing tasks.
[17:19:23] <robertml> (hollenbeck, right?) :)
[17:19:34] <Hollenbeck> right
[17:19:35] <geoff> yes
[17:20:14] <robertml> ggm suggests additional authors
[17:20:29] <robertml> ggm asks if it could be presented again in Paris.
[17:21:14] <robertml> cathy asks if kengo talks to routing people in their natural habitats (mailing lists)
[17:21:18] <robertml> kengo says yes.
[17:21:32] <robertml> andy asks if rpsl is actually needed.
[17:23:14] <robertml> larr asks if maybe rpsl can be encapsulated in the xml.
[17:24:20] <robertml> andy: the reason for having xml is because there are lots of parsers out there.. to have rpsl in xml doesn't make sense.
[17:24:22] --- anewton has left: Logged out
[17:24:53] <robertml> ggm thinks it's on draft level minus 01, but it was just to get it on the table. There is a genuine interest from people.
[17:25:09] <robertml> author is asked to find co-authors and present again in Paris.
[17:25:24] <robertml> kengo done.
[17:25:33] --- mjo has left
[17:25:53] <amarine> thanks for scribing! :)
[17:26:08] --- Hollenbeck has left
[17:26:12] --- CathyM has left
[17:26:17] --- geoff has left
[17:26:22] --- robertml has left: Replaced by new connection
[17:26:32] --- yone has left
[17:26:38] --- robertml has joined
[17:26:46] <robertml> network is back
[17:27:01] <robertml> Larry was Larry Blunk from MERIT.
[17:27:06] <robertml> Meeting is over.
[17:27:12] <robertml> See you in Paris, july 31- aug 5
[17:27:23] --- robertml has left
[17:29:43] --- amarine has left
[17:33:09] --- klensin-ietf has left
[17:36:57] --- dcrocker has left: Disconnected
[17:48:31] --- bhoeneis has left: Disconnected
[17:50:25] --- levigner has left: Disconnected
[18:41:52] --- dcrocker has joined
[18:51:55] --- dcrocker has left
[18:53:10] --- levigner has joined
[19:12:10] --- ggm has left
[19:34:56] --- levigner has left: Replaced by new connection
[19:34:57] --- levigner has joined