Agenda of the P2PSIP WG meeting at IETF 85 TUESDAY, November 6, 2012 1700-1830 (Afternoon Session III), Salon C Conflicts: iri, fmc, netmod, ccamp, tls, ippm Version: 0.2 (revised by Marc Petit-Huguenin) Jabber scribe: Vijay Gurbani Note taker #1: Jean Mahoney ---------------------------------------------------------------------- Administrativia ........................................ 5 min B. Rosen, CJ. Bernardos Note Well IPR Blue sheets WG Status Update ....................................... 10 min B. Rosen, CJ. Bernardos The chairs go through the WG status slides. Dean Willis will take ownership of draft-ietf-p2psip-base-23. Chairs are reviewing draft-ietf-p2psip-drr-03 & draft-ietf-p2psip-rpr-03 after the WGLC. Brian asks if there any reason not to submit to IESG. No objection from the room. Authors have to provide new version of draft-ietf-p2psip-diagnostics, addressing the comments received during the WGLC. If there are extension changes, the draft will go another WGLC A WGLC will be issued after this meeting on draft-ietf-p2psip-self-tuning-06. draft-ietf-p2psip-disco-00 & draft-ietf-p2psip-share-00 were recently posted as WG documents. ACTION: Need reviewers, please read and comment to the list. A WGLC will be issued after this meeting draft-ietf-p2psip-service-discovery-06 Brian ask if there is any reason not to do it. No objection from the room. RELOAD status .......................................... 10 min o draft-ietf-p2psip-base-22 * Presenter: Dean Willis Open issue 1: Reissuing certificates. ACTION: Take that to the list. Open issue 2: Qualcomm IPR. Cullen - We've disclosed and filed. Qualcomm disclosed IPR and added a feature, which isn't relevant. Can we separate that feature into an extension draft and remove the IPR? Hasn't been investigated. Brian - if it's low-cost effort to remove it, that's a great solution. I'm not a lawyer, but I don't think it's a big problem. Cullen - Cisco is shipping this tech. No one has investigated this - when the IPR note showed up with the feature. Dean - do we care? Brian - I don't think we care. If it's easy to isolate, but it's not worth a lot of work. ACTION: Dean to look into it. Open issue 3: NODE-MULTIPLE, need gen and text. Cullen - in NODE-MULTIPLE, to have permission to store things, it's an array. To check a signature, you have to go through i=0, generate see if it matches and keep going. With small numbers, it's ok. But it seems a lame design. I think we decided to live with it. Should say "test" not "text". Marc - Another option is to use service discovery. Cullen - I don't want to remove it because lots of pieces touch it. ACTION: The document should say this is a poor man's service discovery. New apps should use service discovery. Open issue 4: Removing node ID from JoinReq and LeaveReq. Cullen - it was discussed and decided to keep. It was convenient in the code. It doesn't hurt, not a bug. Dean - we'll leave it. Follow up - IESG discusses. 188 discuss items from the ADs, some simple, to the doc is unreadable and unable to be implemented. It needs to be restructured. Brian - ACTION: Concentrate on the other items and then see what the doc looks like. Dean - ok Update after WGLC of DRR and RPR documents ............. 10 min o draft-ietf-p2psip-drr-03 & draft-ietf-p2psip-rpr-03 * Presenter: Ning Zong Summarized the WGLC comments, which were addressed with confirmation. Carlos confirms that the chairs are reviewing the document and will be submitted to the IESG soon. P2PSIP Overlay Diagnostics ............................. 10 min o draft-ietf-p2psip-diagnostics-09 * Presenter: Haibin SONG Issue: Name of the Document. No objection to renaming. Issue: Authorization. Song - think option 2 is complicated, would like to go with 1st option. No disagreement. Issue: Registry. No disagreement. Dynamic Diagnostic Info. No disagreement. Next steps. A new version will be submitted within one month. A SIP Usage for RELOAD ................................. 15 min o draft-ietf-p2psip-sip-07 * Presenter: Thomas C. Schmidt Issue 1: AOR matching overlay name? Four choices: 1. Strict matching 2. Right-most partial matching 3. Leave it to the enrollment server 4. Make it configurable in the Overlay Configuration Document Has implications on look-up strategy Marc - I'm for strict matching. Cullen - I can't see why we wouldn't do strict. I don't have a real opinion. Thomas - [??] Henning - I assume you match on the UTF string. That needs to be documented. Can't assume all ASCII. Dean - The SIP AOR has to match the overlay name. Cullen - my fantasy world there is one overlay with all the SIP AORs. Brian - then strict wouldn't work, leave it to the enrollment server. Cullen - why is there a relation at all? Just a string to identify a user. I need to read the drafts. Dean - how do select an overlay? Brian - it doesn't look like one big email cloud. Dean - when I share documents it looks like one big namespace. Henning - we have a more organized world, and one AOR on your busness card. It needs to be strict matching. Partial matching doesn't mean anything. The implicit assumption that it's a mapping between AOR and overlay, but not always the case. Say - if you working in the environment, where the overlay matches AOR, if you are in a different universe, ignore this section. Marc - the spec says - if the AOR doesn't match -- Thomas - the lookup section says you have to have strict matching. If we don't use strict matching, then we have to change this section. Dean - My sendmail server has a list of local domains, in DNS I have MX records. Could do something similar to MX records. Thomas - like registering proxy servers. Dean - like registering mail servers. Thomas - we've talked about several options here. Brian - ACTION: let's take this to the list. What Henning said made sense to me, need to craft the text. Issue 2: Format of contact_prefs. Brian (as an individual) I prefer SIP style but I don't care. Marc - RFC 2533 is more machine friendly Brian (as chair) - just pick one. Issue 3: Anonymous GRUU. elaborate or point to Marc's doc? Brian - do we lose anything if we drop it? There's Marc's draft. Marc - My draft isn't finished. I don't think it's a bad idea to keep it. I wouldn't tie it to my draft. Thomas - I'll just refer to anonymous GRUU document. Issue 4: Security. No one had comments. Issue 5: Code Point. No objections. Open discussion. Using EKU for RELOAD X.509 Certificates ................ 10 min o draft-petithuguenin-p2psip-reload-eku-00 * Presenter: Marc Petit-Huguenin Any interest in doing this? Any opinion? Silence. Marc - I'll be presenting again in Orlando. Configuration of Access Control Policy in RELOAD ....... 5 min o draft-petithuguenin-p2psip-access-control-03 * Presenter: Marc Petit-Huguenin Adopt as a working group item? Thomas - I found this extremely useful. Brian (indvidual) - I like it too. It makes it simple to put the policies in place. Brian (chair) - there were positive comments on the list. I will call a Hum for making this a WG item. In favor? there were hums. Opposed? no hums. Brian - ACTION: chairs to discuss with ADs to create a new milestone. Anonymization for RELOAD ............................... 10 min o draft-petithuguenin-p2psip-reload-anonymous-01 * Presenter: Marc Petit-Huguenin Jan Seedorf - There are several research papers on onion routing. IPT papers. ACTION: I'll send pointer to the list. RELOAD Interop Testing Event ........................... 5 min July 27-28 Berlin Germany http://implementers.org/mailman/listinfo/reload Send Marc email for config and enrollment service Brian, to Marc - thank you for your work on the interop events. Next steps ............................................. 5 min B. Rosen, CJ. Bernardos Brian - Please read WGLC drafts and send comments to the list.