Monday, March 25, 2019< ^ >
SHollenbeck has set the subject to: REGEXT WG at IETF-103
Room Configuration
Room Occupants

[08:31:09] healthyao2000 joins the room
[08:31:19] healthyao2000 leaves the room
[12:37:13] Meetecho joins the room
[12:37:18] VirtualQueue_KMNnpJqG joins the room
[12:48:14] Jody Kolker joins the room
[12:49:33] Tim April joins the room
[12:49:52] Roger Murray joins the room
[12:50:47] Yoshiro Yoneya joins the room
[12:51:52] Eduardo Alvarez joins the room
[12:52:59] Antoin Verschuren joins the room
[12:53:24] George Michaelson joins the room
[12:53:25] Antoin Verschuren leaves the room
[12:53:25] Antoin Verschuren joins the room
[12:54:02] <Antoin Verschuren> Welcome to IETF 104
[12:54:17] <George Michaelson> ok. if nobody else has taken the gig I can do this
[12:54:28] <Antoin Verschuren> This is the REGEXT WG
[12:54:45] <Antoin Verschuren> Ah George, Please!
[12:56:38] <George Michaelson> RFC8495/8521 made publication
[12:57:20] tapril joins the room
[12:58:37] <George Michaelson> Fee extension and Bundling are in IESG Evaluation
[12:59:04] <George Michaelson> Jim Gould. Verification code
[12:59:20] <George Michaelson> withdrawn (lack of interest)
[13:00:34] <George Michaelson> Grover, Centre for Internet and Society
[13:01:04] <George Michaelson> lot of discussion on list,
[13:01:15] <George Michaelson> Jim: yes, but no closure or traction.
[13:02:06] <George Michaelson> Five Docs newly adopted.
[13:02:17] <George Michaelson>
[13:02:21] <George Michaelson>
[13:02:25] <George Michaelson>
[13:02:30] <George Michaelson>
[13:02:34] <George Michaelson>
[13:02:43] <George Michaelson> Jim calls for discussion.
[13:04:08] <George Michaelson> Scott Hollenbeck does OpenID
[13:04:49] Simon Pietro Romano joins the room
[13:05:28] Joseph Chiu-Kit Yee joins the room
[13:06:20] <George Michaelson>
[13:08:44] tapril leaves the room
[13:08:45] <George Michaelson> tested in thick and thin registry
[13:10:32] <George Michaelson> backend Identity providers from viagenie and verisign
[13:11:36] Sean Baseri joins the room
[13:12:33] <George Michaelson> ongoing policy development work will cause delays in ICANN (roles, jurisdictions and other attributes)
[13:15:44] <George Michaelson> cannot assume browser client only
[13:17:25] <George Michaelson> Adam Roach. params on URLS, BCP190 checks to make sure its in line. avoid problems IESG evaluation time
[13:17:41] <George Michaelson> Andy Newton. device flow, possible to add that later on? not now?
[13:17:52] <George Michaelson> Scott Hollenbeck. Of course if we decide
[13:18:03] <George Michaelson> Andy Newton. complicates things we don't need right now.
[13:20:11] <George Michaelson> Jim Galvin. RDAP falls into API more than screen scraping used at present. ICANN space there will be e.g. reputation provider clients. need to think more about this.
[13:20:36] koji joins the room
[13:21:19] <George Michaelson> Andy Newton. non-browser clients. author of one. how auth would work. probably in a .cfg file somewhere -lots of things we talk about are automation, non-interactive
[13:21:42] <George Michaelson> Access tokens, API keys, in a. file, apply
[13:21:50] <George Michaelson> Jim Galvin. even for bulk?. Andy. Yes
[13:22:17] <George Michaelson> Rick Wilhelm. talk about ongoing policy Development, in ICANN, its the multi-stakeholder model. its not waiting for ICANN as an org, its the ICANN community
[13:22:54] <George Michaelson> Alex Mayrhoffer. Agree with Andy. will be OOB process. eg bulk for registrars. will get API, put in .cfg. will expire. Prefer things in headers not query string
[13:23:54] <George Michaelson> Jaromir Talir. RESTfulness of RDAP. try to much insist on REST expectaction. We need to reconsider a "session" model. easier to implement auth
[13:23:59] Tim  April joins the room
[13:26:03] <George Michaelson> Aiming to focus on work to be used, useful.
[13:26:35] Eduardo Alvarez leaves the room: Connection failed: connection closed
[13:26:44] <George Michaelson> Mario Loffredo, three drafts
[13:26:45] <George Michaelson>
[13:26:52] Eduardo Alvarez joins the room
[13:27:00] <George Michaelson> sorting and paging, partial response and reverse search (one pack)
[13:28:08] <George Michaelson> first draft adds new params and extensions to manage sorting/paging and so new values. changes conformance arrays. defines a cursor (pagination in a huge result set)
[13:30:02] <George Michaelson> asks Q: metadata elements? if yes, need to agree structure, more general metadata section (rate limits, server info)
[13:30:13] <George Michaelson> Asks Q: which pagination method? one, or both?
[13:30:56] <George Michaelson> Mario's personal opinion, do both in spec. let operators specify pagination method
[13:31:13] <George Michaelson> Jim Galvin; Q or discusion on this document?
[13:31:27] <George Michaelson> Jim Gould Verisign: possible for server to implement paging alone ,no sorting?
[13:31:35] <George Michaelson> Mario: yes. this is in the draft
[13:32:11] <George Michaelson> Jim Gould: method to identify what is supported? capabilities?
[13:33:10] <George Michaelson> Jim Gould: identifiers?
[13:33:12] <George Michaelson> Mario: yes.
[13:33:44] <George Michaelson> Andy Newton: pagination method, one or both? if we support both, servers MUST implement both, or decide which?
[13:34:11] <George Michaelson> Mario cannot report paging parameters in initial query. its in the response.
[13:34:30] <George Michaelson> some responses cursor based pagination not suitable so forced to use offset.
[13:34:52] <George Michaelson> but if user does submit sort parameter, can use own index, primary key to implement cursor based pagination
[13:36:08] <George Michaelson> Andy. interaction model is client emits query, cannot be satisfied, resubmit in form which can be satisfied. May want discussion to say capabilities. do not want interaction never reach result.
[13:37:04] <George Michaelson> Jim Gould: similar to registry mapping on EPP side. may make sense to make a server capability query
[13:37:17] <George Michaelson> Mario: let RDAP
[13:37:28] Eduardo Alvarez leaves the room: Stream closed by us: Replaced by new connection (conflict)
[13:37:29] Eduardo Alvarez joins the room
[13:37:30] <George Michaelson> operators decide
[13:38:42] <George Michaelson> RIchard Wilhem: differentiated capabilities per server, how to interrogate. Some under SLA might have response time requirements. mechanism to adjust capabilities to meet SLA (eg if busy) and decline sorting, paging
[13:39:46] <George Michaelson> Mario: now on second draft. Partial response
[13:40:48] <George Michaelson> parameter called 'fieldSet' to get specified fields
[13:43:33] <George Michaelson> decision coming about what should be in field sets.
[13:44:53] <George Michaelson> Jim Gould. falls into similar category as the prior case.
[13:45:11] <George Michaelson> put these into the draft may not be feasible. actual servers return will be based on policy.
[13:45:32] <George Michaelson> define field set client use, but (like capabilities need meta query to get what is available)
[13:46:27] <George Michaelson> Mario: we define fieldSets in our opinion, interoperability between client and server.
[13:46:45] <George Michaelson> Jim Gould: id is easy. but other things, decisions beyond leave out of drafy
[13:47:24] <George Michaelson> Andy Newton. disagree with Jim(s) if understood what is in doc, server can override but client can code to what is in spec and deal with variances
[13:47:53] <George Michaelson> Mario Third document reverse-search
[13:50:05] <George Michaelson> also Qs. should support, but on any entity, or only on specified entities? is notation suitable
[13:50:25] <George Michaelson> e.g. binary match logic encoding? email match X and address match Y (boolean AND)
[13:50:45] <George Michaelson> non-technical issue: privacy consideration section adequate? need more amendment?
[13:53:02] <George Michaelson> Stephane Bortzmeyer. I have big problem with privacy considerations section. potential for harrassment, currently "follow the law" needfs more work
[13:53:44] <George Michaelson> Mario: know there are roles/users, registrants, can search their own domains, not sponsored by others
[13:54:39] <George Michaelson> country rules apply.
[13:54:59] rstory joins the room
[13:55:53] healthyao2000 joins the room
[13:57:17] <George Michaelson> grover: agree with stephane. actional recommend privacy details need to be in tech space (eg role params) specific entities perform function, then recomment role not optional
[13:57:42] <George Michaelson> contentious topic in ICANN. like to hear how much of that concern relevant here
[13:57:42] Eduardo Alvarez leaves the room: Stream reset by peer
[13:57:57] <George Michaelson> Alex Mayrhoffer. Privacy section needs more work
[13:57:57] Eduardo Alvarez joins the room
[13:58:18] <George Michaelson> like see single sentence in document "ok to implement subset of what is specified" otherwise give impression compliant server, client have to implement all
[13:58:39] Sean Baseri leaves the room
[13:58:42] Sean Baseri joins the room
[13:58:48] <George Michaelson> Andy: MUST NOT unless authenticated user.
[13:59:42] <George Michaelson> Jim Gould: Login Security
[13:59:43] <George Michaelson>
[14:00:18] <George Michaelson> RFC5730 only does 16 char passwords. needs fixing
[14:00:31] <George Michaelson> EPP doesnt support login warnings, events
[14:00:41] <George Michaelson> need client-server information
[14:02:22] Ondřej Caletka joins the room
[14:04:39] <George Michaelson> Robert Story, enumerate stuff in client
[14:04:49] <George Michaelson> Jim Log and security pol extension (EPP) draft covers this
[14:05:10] <George Michaelson> Martin Casonova. Q about agent field. free text
[14:06:20] <George Michaelson> Stephane Bortzemeyer. attack against registrar last year. specific advice move outside security consideration, point to more general document.
[14:06:32] <George Michaelson> Jim: welcome suggestions
[14:07:11] <George Michaelson> Jim Galvin: Milestone review
[14:08:44] <George Michaelson> conflict between RDAP and EPP for time to complete. many RDAP documents
[14:09:35] <George Michaelson> suggestion from AD (who is going to change, Barry Lieba coming in)
[14:10:06] <George Michaelson> deal with both equally. overlap, set of people interested in one, and other.
[14:12:19] <George Michaelson> GGM: RDAP is increasing importance. if both streams, we need two meetings.
[14:12:42] <George Michaelson> RIchard WIlhelm: things in RDAP which are forward looking. challenges just in initial implementation.
[14:12:57] <George Michaelson> some of RDAP is more tactical, other more longterm
[14:14:05] <George Michaelson> Jim Galvin: milestone management is process. nothing prevents WG going through docs as quickly as you can. give opportunity to get milestones other docs on it. success of docs is WG members wanting to progress
[14:15:02] <George Michaelson> Alex Meyroffer. Two workstreams have diferent size of the audience. EPP 1000 registries, 20,000 registrars worldwide, 10 registries. RDAP audience is hundreds of millions. lot more interest in that. spend time on protocol with widest with most impoact. EPP hidden protocol.
[14:15:09] <George Michaelson> RDAP public protocol. hard.
[14:16:44] <George Michaelson> Peter Koch: if EPP is the airline mechanism, then RDAP is the reporting mechanism to government (!) but don't over-assume RDAP is the ICANN outcome, they haven't actually closed on it. Privacy, Policy aspects haven't been adequately closed. big work. Challenge. cannot rush.
[14:17:10] <George Michaelson> WG needs broad focus. not just the travel agents talking to the airline
[14:17:21] <George Michaelson> Andy Newton: if have two meetings, maybe we need two WG
[14:17:46] <George Michaelson> if I did EPP work, I would not want it to be starved. if we think RDAP is an avalanche it may make sense to carve out, to get EPP work done.
[14:18:16] <George Michaelson> Jim Galvin. right way to manage. as chair, not inclined to carve in two. strong relationship, same people in both groups (to some extent)
[14:18:45] <George Michaelson> Jim Galvin: propose 5 tasks per milestone, two from each, one floater
[14:19:02] <George Michaelson> This is just project management
[14:20:43] <George Michaelson> (discusses backlog of six things, suggesting we have to pick two)
[14:21:59] <George Michaelson> Jim Gould. do not wish to split. but also do not want delay
[14:21:59] healthyao2000 leaves the room
[14:22:12] <George Michaelson> scott hollenbeck: see the possibility of splitting the WG. not sure I like
[14:24:08] <George Michaelson> could we expand beyond five? could we do three per stream, not two? some can delay
[14:24:32] <George Michaelson> Jim Galvin. some things will sit, and take up a milestone slot.
[14:24:56] <George Michaelson> no harm having WG docs, no committed milestone, not ready to deadline, keep on 5 slots, can assign to something.
[14:25:02] <George Michaelson> pick items we can commit to as milestones
[14:25:25] <George Michaelson> five milestones, more is lot to manage.
[14:26:12] Sean Baseri leaves the room
[14:26:12] Sean Baseri joins the room
[14:26:55] fneves joins the room
[14:28:24] <George Michaelson> Andy some agree Scott, worried RDAP work not coming into room because visibly no traction. much hidden demand
[14:28:44] <George Michaelson> GGM: privacy is going to need a document of its own, its a big, big issue. fundamental collision LEA and individual privacy.
[14:28:49] <George Michaelson> Jim maybe too many open questions
[14:29:16] <George Michaelson> Barry Lieba: (AD) shame to re-split. won't object if WG decides, but believe capable being managed as-is
[14:29:39] <George Michaelson> can negotiate how many milestones appropriate: wanted limit, but five is not magic. can decide what is right answer. docket model
[14:30:21] <George Michaelson> if document on privacy that important to get done, shoiuld put milestone on it. aggressively work on it, so doesnt block other documents. any WG only has limited capacity
[14:30:28] <George Michaelson> if split, get TWO WG with less capacity, wont help
[14:31:04] <George Michaelson> Jim Galvin: will take to list.
[14:31:19] <George Michaelson> can say seven, manage, its what the poll said. speak to list
[14:31:49] fneves has set the subject to: REGEXT WG at IETF-104
[14:31:52] Yisong Liu joins the room
[14:32:37] <George Michaelson> Tom Harrison: RDAP Mirroring Protocol
[14:32:39] <George Michaelson>
[14:34:00] <George Michaelson> APNIC needs this for NIR WHOIS mirrors coming into APNIC, switching to RDAP mirror, we get i8n outcome
[14:34:13] <George Michaelson> mechanism is snapshot pointer, delta set
[14:34:44] Yisong Liu leaves the room
[14:35:12] <George Michaelson> serial number sync mechanism.
[14:35:29] <George Michaelson> default type refs to avoid duplicate info unchanging
[14:35:38] <George Michaelson> can refer to objects by reference avoid sending
[14:35:44] <George Michaelson> refreshh timer can be specified
[14:36:08] <George Michaelson> believe interest in domain name, not seeking adoption. will address privacy.
[14:37:36] <Jody Kolker> I can't imagine that our legal department would ever approve a client to have our complete whois database.
[14:37:53] rstory leaves the room: Disconnected: closed
[14:38:28] <George Michaelson> stephane bortzmeyer. privacy issue. much more sensitive data. interest in idea? legal would refuse to do bulk
[14:38:41] <George Michaelson> stephane flag redacted missing
[14:39:13] <George Michaelson> richard escrow covered by different mechanism. don't see gTLD use case. understand your use case but not generally applicable
[14:39:35] <George Michaelson> marc blanchett: registrar ID to RDAP server URL
[14:39:47] <George Michaelson> (no slides)
[14:40:17] <George Michaelson> RDAP to thin registry, needs link to registrar RDAP server. how to find the server URL?
[14:40:49] <George Michaelson> since have IANA registry for bootstrap, idea is to have IANA registry for reg ID to RDAP
[14:41:25] <George Michaelson> ICANN using IANA as registry, not governed by IETF. gives to any accredited registry. use number as RDAP JSON key pointer.
[14:42:02] <George Michaelson> Qs sent to ICANN RDAP pilot "does use case really exist"
[14:43:42] <George Michaelson> some feedback received on the list
[14:44:55] <George Michaelson> Andy not all Qs on mailing list
[14:45:42] <George Michaelson> don't like algorithm
[14:46:01] <George Michaelson> Jim Gould. problem came up before ICANN put URL in. I believe the problem is solved
[14:46:55] <George Michaelson> Andy: what URL is being updated? we need to know.
[14:47:14] <George Michaelson> Andy separate bootstrsap files have to be dealt with.
[14:47:36] <George Michaelson> Marc: temporary solution until draft. if draft not appropriate.. ANdy here is way to do it
[14:47:52] <George Michaelson> Gavin Brown. registry is not to be consumed by clients, by servers.
[14:48:19] <George Michaelson> are in process of bootstrapping. registrars:registries == millions of emails. simplify
[14:48:36] rstory joins the room
[14:48:57] <George Michaelson> IANA maintains but ICANN populates. RDAP client doesnt need, will never consume. is link in the response from gTLD rdap server
[14:49:26] <George Michaelson> Jody Kolker, Q for Marc, registrars, ccTLD not iana accredited. hundreds of them. reseller
[14:49:36] rstory leaves the room
[14:50:18] <Jody Kolker> yes
[14:50:32] <Jody Kolker> Please update jCARD.
[14:50:38] <George Michaelson> Andy Newton JCARD issues discussion. no slides. will send notes to ML. discuss on list. how many people think we need something done about JCARD
[14:51:28] <George Michaelson> Alex Meyrhofer. Standard for registrry lock to prevent (re)delegation of a name. having side meeting wednesday 2pm. 15ppl seating. Paris Room
[14:51:37] <Jody Kolker> I would like to remote attend.
[14:51:44] <George Michaelson> Wednesday 2pm
[14:52:02] <George Michaelson> No remote attendance in signup room but will do summary
[14:52:07] <George Michaelson> (Jim said)
[14:52:14] <George Michaelson> maybe somebody can zoom you in
[14:52:27] <Joseph Chiu-Kit Yee> maybe jabber scribe?
[14:52:32] <George Michaelson> ROW #8 9th May Bangkok, Thailand CFP in progress.
[14:52:46] <George Michaelson> yes, maybe somebody ELSE can volunteer to scribe :-)
[14:54:15] Meetecho leaves the room
[14:54:48] Eduardo Alvarez leaves the room
[14:54:58] Ondřej Caletka leaves the room
[14:54:58] Sean Baseri leaves the room
[14:54:58] Roger Murray leaves the room
[14:54:58] Joseph Chiu-Kit Yee leaves the room
[14:54:58] Antoin Verschuren leaves the room
[14:54:58] Tim April leaves the room
[14:54:58] Simon Pietro Romano leaves the room
[14:54:58] Jody Kolker leaves the room
[14:55:33] VirtualQueue_KMNnpJqG leaves the room
[14:55:38] George Michaelson leaves the room
[15:04:08] Yoshiro Yoneya leaves the room
[15:04:11] Yoshiro Yoneya joins the room
[15:04:37] Yoshiro Yoneya leaves the room
[15:08:06] Tim  April leaves the room
[15:11:19] koji leaves the room
[15:14:50] George Michaelson joins the room
[15:14:51] George Michaelson leaves the room
[15:15:27] fneves leaves the room
[15:15:32] fneves joins the room
[15:15:59] koji joins the room
[15:19:51] George Michaelson joins the room
[15:21:13] fneves joins the room
[15:33:24] fneves leaves the room
[15:34:40] koji leaves the room
[15:36:33] koji joins the room
[15:39:04] koji leaves the room
[15:52:59] fneves leaves the room
[16:06:08] George Michaelson leaves the room
[16:13:21] George Michaelson joins the room
[16:33:48] George Michaelson leaves the room
Powered by ejabberd - robust, scalable and extensible XMPP server Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!