6lo: IPv6 over networks of resource-constrained nodes WG Minutes Meeting: IETF 90 Thursday July 24, 2014 Time: 0900-1130 Morning Session I Location: Tudor 7/8 Chairs: Samita Chakrabarti Ulrich Herberg Technical Advisor: Ralph Droms Minutes: Kerry Lynn Thomas Clausen Jabber: xmpp:6lo@jabber.ietf.org Scribe: Ole Troan Audio archive: https://www.ietf.org/audio/ietf90/ietf90-tudor78-20140724-0900-am1.mp3 URLs: https://www.ietf.org/mailman/listinfo/6lo http://datatracker.ietf.org/wg/6lo/ ========================================================= 09:00 Meeting Opening, agenda bashing, and note well [Ulrich Herberg] A slight agenda change from the posted agenda was discussed. 09:03 Draft Status since IETF89 [Samita Chakrabarti] draft-ietf-6lo-btle: Currently with IESG, but is going to come back to the WG for review as substantial edits are needed due to BT-SIG recommendations. Markus Isomaki also provided a short status on the microphone -- BT-SIG has updated their core specification to include IP support and will likely publish that spec.BT SIG plans to have an interop event in Fall and after that the BT-SIG document update will be confirmed and then it will be a good time to update the btle IETF draft. draft-ietf-6lo-lowpanz: 05 version of this draft has been submitted for IESG submission (AD review) during the IETF week after passing the 2 WGLC and an IPR call. draft-bormann-6lo-6lowpan roadmap: Expired. Carsten Bormann (CB) agreed to update roadmap. draft-ietf-6lo-ghc and draft-ietf-6lo-lowpan-mib completed WGLC and draft-ietf-6lo-6lobac and draft-ietf-6lo-dect-ule have been adopted as WG document. There are number of new documents submitted as individual drafts in 6lo WG. 2 documents on ipv6-over-foo(PLC and NFC), 3 enhancement documents, 3 security related proposals are among the notable ones. Further status on other documents from the existing milestones: 5 documents adopted, 2 WGLC completed. Also draft-thubert-6lo-forwarding-fragments is on 6lo milestones, so the WG needs to have discussion on this document and determine whether the WG needs to work on fragment forwarding issues in 6lowpan. draft-thubert-6lo-forwarding-fragments previously received comments from the WG - the draft should address those comments in the upcoming versions. Samita Chakrabarti: Is there a need to generate a generic 6lo guideline I-D? The milestones will be updated soon. Also the chairs requested the WG to actively provide more feedback on the existing documents in the 6lo WG page. A LLN plugfest was held on July 20 (Sunday) at IETF jointly with 6lo, 6TiSCH, and ROLL. Kerry Lynn: Would like to ask that future plugfests will be announced earlier for travel planning. 0917 draft-ietf-6lo-lowpan-mib: (Juergen Schoenwaelder) - MIB provides valuable insight into what stack is doing - LOWPAN MIB fits between IPv6 and L2 MIBs in architecture - Possible DoS attack for severely constrained nodes can be detected - Showed slides of what happens during the plugfest, discussing the contiki implementation - Stresses that the document defines counters that go into different levels of network management implementations Pascal Thubert: One counter would have been of much interest: counter of packets being lost during the period when the one buffer is filled by a fragment waiting to be reassembled. I would like to figure out how often this situation occurs. Do you know? Juergen Schoenwaelder: Special counter for packet drop while stack is not processing packets normally; need to decide if a special counter is needed, or the simple discard counter is enough. We need to discuss that. For the second question, it obviously depends on the link, packet size, etc., so I do not really know. Pascal Thubert: 1, 2 or 3 buffers, still have the problem. If you have the information of drops, however, then you can actually do something, tune things. Michael Richardson: What do you think the IETF administrative overhead of adding a counter to a document id? Better to "think hard" and issue a spec -- but I do prefer to push out a module, and then rev it, sorta "we'll rev this every two years, throw your counters in". Doing that, or publish a draft at rev-47? But, is worried that either can appear as a DoS attack on the IETF. Ulrich Herberg (Chair): Ideally, publish now and know we are perfect. Have been through that process of developing and reviewing MIB modules: obsolete RFC with new MIB module. Not do that every other month. Brian Haberman (AD): The only problem I have with Michael's question is, if we go to rev 47, where in the OID tree do we anchor this? Need to publish to get an identifier. Pascal Thubert: As we push the different IPv6-over-foo documents out, should we not have a MIB considerations section? Carsten Bormann: Basic idea about a MIB: you expect people to actually implement it. The other thing that comes up sounds more like an IANA registry of whatever counters came up this week. Counters that few implementations share do not need to go through the troubles of having a specification through the IETF. The idea of this doc should be having the counters that we all agree on will be useful and push the vendors to implement. Kerry Lynn: Follow-on from Carsten: there is some cross-pollination between Contiki and the IETF, would encourage to get your work into the Contiki mainline. Will be glad to help. Juergen Schoenwaelder: Usually, try to identify the common subset of counters, core set, more means more memory requirements, etc. Not yet convinced about the stack-drops-buffer-full-waiting-for-fragment counter that Pascal suggested, but that can be explained in the document and discussed. Peter van der Stok: Would like to encourage that it is published in its current state. Can use to force vendors to publish Samita Chakrabarti (Chair hat off): Also thinks that it is important to have this published, as a first example, and to move ahead with implementations. Ulrich Herberg (Chair): WGLC was over, no request for new counters then. I cannot keep this open-ended forever, do you need another week? Michael Richardson: What is the overhead of "putting that in later"? May be we need more revs as we come up with new counters. Ulrich Herberg: Infer that we do not need another WGLC for this. The one remaining question is the intended status. 0940 Intended status of draft-ietf-6lo-lowpan-mib (Ulrich Herberg) Suggests keeping the intended status as PS, and then add a section explaining that it is a data model, can be used with SNMP, but also by something else. Walking over a set of smaller issues, stack picture from slides into introduction, some educational material, etc. Thomas Clausen: Experimental drafts must now typically pass or fail some bounded experiment and transition to another status. Carsten Bormann: This could potentially be a cross-layer effort with COMAN OPSAWG WGLC announcement: Juergen Schoenwaelder/Chairs draft-ietf-opsawg-coman-probstate-reqs draft-ietf-opsawg-coman-use-cases/ Developed in the un-official COMAN activity. Wants to make 6lo aware of these, and ensure that there is review 0950 GHC (Carsten Bormann) Carsten Bormann: Suboptimal performance is "good enough" Kerry Lynn: Any open source implementations? Nalini Elkins: "Codematch" will match willing students with interesting projects Brian Haberman (AD): Soon datatracker will contain links to GitHub projects Ulrich Herberg: Likes to start a WGLC in 2 weeks. 10:00 draft-ietf-6lo-dect-ule: Jens Toftgaard Petersen Pascal Thubert: What does "NEMO" in DECT std refer to? Jens Toftgaard Petersen: "No emission" mode Pascal Thubert: You do not have to convince me that we should do IPv6 over this media. What is left to be done? Jens Toftgaard Petersen: There is nothing left. Samita Chakrabarti: Can you send comments on this draft on the list? After that we will go for last call soon. Abdussalam Baryun: Are there any test results? Jens Toftgaard Petersen: Already installed in the Gigaset elements, so quite a lot market experience, but I do not have any technical reports. 1020 LoBAC (Kerry Lynn) - MS/TP data link layer changes have been approved; only waiting now for a Frame Type assignment - ASHRAE has made the data link spec available; contact Kerry Lynn for a copy Motivates the use-case of industrial and building control. BACnet, is working on binding to IPv6. Presented overview of the technical approach in the doc. Carsten Borman: In the 6lowpan space we found that "having an uncompressed form" was a mistake, since people implement - but do not test - it, and therefore it contains bugs when turned on. Kerry Lynn: In order to support compression we need additional support from 6ND such as the distribution of 6CO options. Carsten Bormann: Not necessarily; if you have it then it is better, but you do not need it. Kerry Lynn: Are you suggesting that we eliminate non-compressed entirely? Carsten Bormann: In agreement. Kerry Lynn: Would like that we take this discussion on the list. Pascal Thubert: In that case, should we not do this in general, and not for this doc? Carsten Bormann: It is easy to remove uncompressed from new specifications, hard to tell people to take it out of existing 6lowpan 15.4 code. Pascal Thubert: If we want to deprecate it, then let us start the process. Samita Chakrabarti (Chair): We will take this to the list. Expecting some updates to this document? Kerry Lynn: Frame assignment, and a couple of correction. But, it is pretty mature and straight-forward, and I would like one more cycle then go for WGLC. 10:30 draft-popa-6lo-6loplc-ipv6-over-ieee19012-networks: Daniel Popa Subir Das: Can we get a copy of the IEEE spec at no charge? Brian Haberman (AD): Not 802; may involve a liaison request. Abdussalam Baryun: Why std. track and not experimental? Daniel Popa: Already developed, market need. 10:39 draft-hong-6lo-ipv6-over-nfc: Y. Hong Presentation of the NFC architecture, in which this is to be integrated. Comparison NFC-BT-BTle, and of the applicability of NFC. Then, technical presentation of the IPv6 mapping to NFC Showing photos of their test platform, and development work. Hopes to have running code at the next meeting. Suresh Krishnan: This draft needs address mapping details before adoption. 1050 draft-thubert-6lo-rfc6775-update-reqs (Pascal Thubert) Thomas Clausen, Ulrich Herberg: Currently this proposal is very RPL-centric. Pascal Thubert: It can be made more protocol independent. Erik Nordmark: The need to express most recent registration is independent of routing. Carsten Bormann: How much can be done with options as compared to "breaking" the existing protocol? Backward compatibility is not just the new scheme working with the old, but vice versa. PT: Router expresses its capability. Erik Kline: Is there anything that prevents proxy from being implemented in the 6LBR? Pascal Thubert: No, the functions are logical and may be combined in the same device. The 6LBR has to be the root Erk Nordmark: Current draft combines requirements and solution. Advise factoring this into separate drafts. Samita Chakrabarti: Think that this draft needs modifications for generalizing. I think that I heard that there may be some other requirements. If you can work with other group members, then that would be good. Suresh Krishnan: Agree with Samita -- but, there's something that I do not know how to make more generic, and there's a lot of thought needed to do that. You can form an address, bind it with some CGA at some point, and then use it later once we have demonstrated proof of ownership. This is interesting, we should discuss this. 11:08 draft-sarikaya-6lo-cga-nd: Behcet Sarikaya Presenting the idea, slides Erik Normark: What is the OII and UII? An EUI64, if instead of that I use the 64 and make that a SeND hash, I can now prove that I have those 64 bits. But, unless these are also my Interface ID, I haven't proven anything. That's not very flexible, SeND is a lot more flexible. Pascal Thubert: We have this central point that can validate first come first served, this registrar that will remember the UII that was used the first time this address was used, and enforce it. Samita Chakrabarti: Take it to the list, we are out of time. 1120 Diet-ESP (Daniel Migault) draft-mglt-ipsecme-diet-esp draft-mglt-ipsecme-diet-esp-requirements draft-mglt-ipsecme-diet-esp-iv-generation draft-mglt-ipsecme-diet-esp-payload-compression Thinks that the requirement document and the diet-esp document are pretty mature and would like to ask for adoption Brian Weis: This approach is perhaps too ambitious. Part of your requirement is "not reduce the security of IPSEC" And then there's the alignment that I am not sure is completely compliant. Also, for the IV compression, you want to be sure... Kerry Lynn: Possible to support group mode? 11:30 6tisch Plugfest: Ines Robles Really quick presentation of the slide of the plugfest on Sunday IETF where 6TiSCH, ROLL, 6lo implementations feature testing took place. Several implementations from different vendors, research organizations and Universities joined the plugfest demonstrating 6lowpan, 6Tisch, and Roll updates. Slides are available at: http://www.ietf.org/proceedings/90/slides/slides-90-6lo-8.pdf Next IETF plugfest is planned on IETF92. Wrap-up by Samita Chakrabarti - will take action on the many things presented and issue appropriate calls on the mailing list. Please take a look at last part of chairs slides for the need of a "6lo" usecase document describing different ipv6-over-foo usage based on the L2 technical requirements[http://www.ietf.org/proceedings/90/slides/slides-90-6lo-12.pdf]