IETF-81 Roll Minutes Session 2011-07-29 1300-1515, room 208AB - Audio stream - roll chatroom Co-chair: JP Vasseur (present), David Culler (on jabber) Agenda/Slides: http://www.ietf.org/proceedings/81/agenda/roll.txt Jabber logs: http://www.ietf.org/jabber/logs/roll/2011-07-29.html Minutes: Laurent Toutain and Meral Shirazipour – Thanks to Dan King for merging minutes. 1. WG Status (Chairs - 10 mn) [15] * Co-chair mentioned the applicability statements already discussed on the list; the first applicability draft has been made WG draft, so far there are no competing documents. JP opened the Mic for comment. - Thomas Heide Clausen thinks it would have been better to wait before WG adoption - Pascal Thubert responded that it was a good thing as it triggered great discussions on the list * Updates on various documents - Roger K. Alexander briefly mentioned the roll security draft - Pascal: OF-zero IESG review process has already started on the list (draft-ietf-roll-of0) - Hysteresis draft passed WG last call (draft-ietf-roll-minrank-hysteresis-of-04) – JP mentioning that publication request will follow soon. - p2p drafts in progress (to be discussed during that WG meeting) - 6man: behind in review, outstanding comments (from INT AD) to be resolved shortly. - JP went over milestones; reminded that top priority right now is applicability statements; discussed re-chartering (basically, rechartering discussed is deferred); and provided refreshed on applicability statement definitions. 2. Applicability Statement for the Routing Protocol for Low Power and Lossy Networks (RPL) in AMI Networks draft-popa-roll-applicability-ami-00 , (Jorjeta Jetcheva - 10mn) [25] - Authors are grateful for the comments received on the list - Trickle parameterization is still ongoing - List comments on PAP2, SG-Net and Open-SG/SG-Net, authors comment: these do not address the same requirements the draft is looking at. General discussion: [JP allocated 20 minutes from buffer considering the number of comments and fuitfull suggestions at the Mic] - Thomas: real work scenarios are not taken into account (bidirectional traffic etc.). - Co-chair asked to have a list of requirements on the list for authors to take into account (general to RPL). - Jari Arkko thinks the job of the document is to say what the characteristics of the application is (not if it works or not). It should cover how it scales, if it works well in each direction, etc. - Stewart Bryant asks Thomas to please provide what topology or scenarios he thinks the protocol will not work for. - Thomas pointed to a scientific paper that has numbers that shows this may not work. Co-chair and others do not agree. To be discussed on the list. - Ralph Droms thinks Jari is right (we are not interested in if it works or not) but we are interested in: what is the better way to use it and how it is best to ' avoid' using it- i.e. asking for guidelines. - Cédric Lavenu: Agrees with Jari to have guidelines of when and how it works better or works optimally - Angelos Castellani: Give a way to implement routing table in so big topology. How to aggregate routing entries - Thomas Herbst would like to see simulations-deployment results as well. - Ulrich has implemented RPL and will participate in helping by sharing experiments on how RPL should be used. - Area Director (Adrian) wants an iterative discussion on the list on what people want to see in this document. He gave the example of BGP and how no one has proven yet that it works in all deployment scenarios - same will go with RPL. Adrian asks the WG to share on the list any scenario-topology that is important to them. 3. RPL P2P (Mukul Goyal - 30mn) [55] 3.1 "Reactive Discovery of Point-to-Point Routes in Low Power and Lossy Networks", draft-ietf-roll-p2p-rpl - Authors emphasized that they optimized for reachability in both direction and they discussed how to run Trickle for P2P-RPL - Pascal: clarified issue on interface with an address reachable in both directions - Thomas asks for multicast results to be shared by anyone who has results (?) 3.2 "A Mechanism to Measure the Quality of a Point-to-point Route in a Low Power and Lossy Network", draft-ietf-roll-p2p-measurement - Co-chair asks if the proposal is stable, author thinks it is. 3.3- "A Compression Format for RPL Control Messages", draft-goyal-roll-rpl-compression - Pascal: thinks it is very useful and supports it - David Culler (Jabber), thinks it is an effective solution - Ulrich: clarified that base implementation will need to be modified to support this. Why it is not done in RPL specification itself. - Pascal: the one critical point is energy and computation energy << transmission energy - Thomas would like to use this document to be published as an upgrade to RPL document and not an option. - Jabber, Re: Thomas' comment. I would argue against "two formats" - default and other. Better to have a defined compaction scheme where the "most general case" is the carried fully in-line - Pascal and author: answer to Thomas: in many cases you don’t need to do the compression - Jonathan (?) had a question on where authors get the context information - Ulrich Helberg also thinks it is a good idea to make this default - Co-chair’s summary: p2p drafts some are quite mature, last one on compression needs opinion on the list. 4. Adapted Multimedia Internet KEYing (AMIKEY): An extension of Multimedia Internet KEYing (MIKEY) Methods for Generic LLN Environments", draft-alexander-roll-mikey-lln-key-mgmt, (Roger K. Alexander - 15mn) [70] - Author went over the reasons for choosing MIKEY. - Micheal Richardson clarified if this is about updating the key or negotiating a new key. - Co-chair asks WG to please comment on this draft – it is a key issue. 5. The Direction Field in Routing Metric/Constraint Objects Used in RPL", draft-goyal-roll-metrics-direction , (Mukul - 5mn) [75] - Author explained that this is the merge of two previous drafts 6. Identifying Defunct DAGs in RPL", draft-goyal-roll-defunct-dags, (Mukul - 5mn) [80] - Author very briefly introduced the new draft 7. Update from Pascal Thubert on Applicability statements: RPL for Industrial, Automation (10mn) [90] - Pascal explained that RPL applies very well to pre-configured scenarios like industrial-automation cases. - The draft is to be published before the next IETF (Tom Phinney is the editor)