Registration Protocols Extensions (REGEXT) IETF 100, Singapore, Meeting minutes Co-chairs: Jim Galvin, Antoin Verschuren Mailinglist: regext@ietf.org ================ Existing Documents ------------------ Launch Phase is in publication requested - all good so far. Fee Extension (Roger Carney) --------------------------- Couple of open questions from implementation - next version (09) is ready to be published (adds just one item to the schema). Two questions: 1) "cd:avail" flag if only a partial set is returned - currently returning 0 (unless full set is given). Jim Gould argues for the opposite (should be 1 even in case of partial responses) - will bring it to the list. 2) classification on the command vs. object level - should be at the object level (proposal to move it up to the object level) - Roger agrees, if someone disagrees Roger will post the two last remaining items to the list, then roll -09 and then ask for WGLC Validate ------- -02 was posted in August. Only a handful of comments, and currently looking for implementors. Jim Gould: Option to use it for existing contacts, rather than supplying "new" contacts? Roger: Mainly intended for domain creates Registry Mapping (Roger Carney) --------------- http://www.verisign.com/assets/epp-sdk/verisign_epp-extension_registry_v01.html Will talk more about it later in the Working Session Milestone Review ---------------- Back in Chicago, got pushback on new documents from AD - should clean up document queue, 3-5 documents are going to go to IESG before end of year or latest Q1 2018. Discussion goal is to identify whether documents are ready from the WG perspective. 3-5 people required during WGLC to indicate concensus - can be hard to reach "non-authors" because of group size. - Change Poll: Chairs think it is ready, Jacques said they intended to implement Jim: Implementation from Neustar and Verisign. Think it is ready for WGLC - Jim will make a request on the list. Nobody objects against that plan. - Allocation Token: same. - dnsoperator-to-rrr: Needs more review, is not there yet for WGLC. Jim Gould: Did a recent review, and has a concern about lack of addressing of the trust relationship between dns operator and other parties. Jim Galvin (as individual): appreciates the 2 implementation efforts (CIRA and CZ.NIC), however it's difficult for gTLDs to support this for policy reasons. (Is defined out of scope in document). Roger: As Registrar, changing data outside of the registry/registrar loop is problematic for registrars. Communication "along the chain" needs to be figured out. Currently no intention to implement this. Jim (as Chair) notes the Milestone for this document is coming up soon - heads up to document authors. - bundling-registration: Defines just one way / use-case of bundling, but does not cover all options. Suggestion from the Chairs is to make it informational, and move it forward that way. Jiankang Yao: We have significant implementation / operation experience in that under .cn. Scott: If we go to informational, then add an Implementation Status section with the experiences? Ning: If there are other bundling mechanisms / use cases, please feed back to us and we can add them Edmund Chang: Wouldn't that be worthwhile as a standard if it would be subject to IDN bundling only? Jim: No way to enable/disable individual registrations in a bundle, and DNSSEC would need to be dealt with - org/org-ext: Antoin: Documents need a thorough review, but also implementation analysis / experience report. Suggests the authors push for review or more implementation. Ning: reseller draft is implemented, org drafts are not implemented yet, research ongoing on whether implementation feasible. Jim Galvin: Implementations in other contexts? Ning: Registrars and .cn Roger: Never found a reason to implement this Jody: Have no reason to implement this - considers that information that is of no concern to the registry. ICANN option to include it is considered an overkill. Jim Galvin: Falls into same category as the dnsoperator-to-rrr and bundling - whether that's a standards track document or informational candidate. Scott: Since when is implementation required for standards track? Jim Galvin: Wide applicability is, not necessarily implementation, right. Andrew Newton: Don't like the concept "i'm not going to do that" -> "we won't standardize it". Andrew Sullivan: If this WG is not the work to produce one canonical document for a problem, then please point us to where that work is performed. Jim: That's what the IANA extension is for - doesn't mean everything has to be standards track Ning: not sure we should always spend the energy on that adoption yes/no discussion or standards track / informational RDAP jcr (Andrew Newton) --------- Problem: RDAP defines JSON just in prose - no formalism -> difficult testing. Uses JSON content Rules (JCR draft-newton-json-content-rules) (Mario Loffredo, .it) - Following another approach than JSON Scheme... Thinks that JCR is better because fits better to the way RFCs are written. Thinks JCR might be less useful for other REST services, lacks ways to describe relationships between query and responses. Plus, only few implementations for JCR vs JSON Schema. Scott: Where would the base JCR document go? This WG is not the right place for the base JCR doc - but where is it? Jim Galvin: Dispatch? (??): Similar to CBOR. (??): Really useful, easy to use, i like it. Adam Roach (as AD): JCR base to be taken on in this WG? Andrew: No, didn't say that. Adam: Good- we just need to make sure underlying technology is pushed forward before we have apps on top of it here RDAP sorting-paging & partial response & reverse search (Mario Loffredo) -------------------------------- sorting-paging & partial-response: takes practices from REST services to RDAP. New parameters "count", "sortby", "limit/offset", and new properties "paging_count", "paging_links". Alternative to offset "cursor" - logical pointer to the next page. Discussing advantages/disadvantages of offset / cursor based variants. Field sets - two options: List of fields vs. named sets of fields. Flexibility vs. comfort. Reverse search: Registries already provide users with reverse searches (eg. domains based on nameservers). Q: Should reverse search be based on other entities as well? Should it be extended to the other types of searches? Andy: Search - before we do anything like this, we need to get the OAuth stuff done. Plus, this is very foundational work, and should be standards track work. Scott: Please get your plate clean - RDAP is moving into operational status, and we'll encounter more similar things which we need to work on. Discussion about registrars requiring lists - comment: We don't want lists in EPP. Interim Meeting feedback -------------------- First two interim meetings were very productive, last one was zero participation (4 or 5 people on the call). AOB --- Andrew: a) BCP about operating servers - sometimes hard to contact the operators b) Draft about http vs. https, sometimes people cannot operate https Stephane: Extend 5731 RFC to allow registration via DNAMEs Kim: The only use case was the one presented - not work on this? Jim Galvin: Push that to the mailing list? Jim: Discussion about broadening the charter to adopt other "registration-related" documents, once the document "plate" is clean. Before london, several documents should be off our WG, and we can adopt more documents. Alex: Search is underspecified, would appreciate working together with others to get that fixed Scott: Was similarly confused when writing the section to capture the opinion of the working group (WEIRDS)