Session 2012-11-05 1520-1720: Salon C Audio Archive http://www.ietf.org/audio/ietf85/ Roll Jabber Log http://www.ietf.org/jabber/logs/roll/2012-11-05.html Agenda IETF 85 http://tools.ietf.org/agenda/85/slides/slides-85-roll-0.txt scheduled waiting time object / deterministic routing in industrial networks http://tools.ietf.org/agenda/85/slides/slides-85-roll-4.pdf Diagrams explaining MP trickle-mcast http://tools.ietf.org/agenda/85/slides/slides-85-roll-5.pdf issues for trickle-mcast http://tools.ietf.org/agenda/85/slides/slides-85-roll-6.pdf RPL Routing Pathology In a Network with Mix Nodes http://tools.ietf.org/agenda/85/slides/slides-85-roll-7.pdf Multicast Protocol for LLNs http://tools.ietf.org/agenda/85/slides/slides-85-roll-8.pdf 1) Agenda/admin (Chairs - 5mn) - Adrian Farrel provided a Note Well on IPR, “disclose early!” 2) WG Status (Chairs - 10 mn) - Two documents in Last Call: draft-ietf-roll-trickle-mcast-02 draft-ietf-roll-security-threats - Two RPL documents in IESG review (AD Eval) draft-ietf-roll-p2p-measurement draft-ietf-roll-p2p-rpl 3) Security Framework and Applicability Statement template (Michael - 10m) - Michael polled group asking who had read the document, a few hands shown. - Comments from the group supported the document and people felt it was a good starting framework. - Michael requested issues should be raised on the list a few items recently raised included mesh links and metering. - Michael noted that some metering applications may require separate applicability discussion and documents, water versus gas, rural versus urban, power versus unpowered, etc. 4) RPL applicability in industrial networks (Pascal - 10mn) - Co-chairs asked who had read the document, quite a few hands shown. - The room was asked to show support for the document and specific subject area. Discussing and demonstrating how RPL can be used for industrial networks would be positive. - General comment from the room that a new objective function would with sending or receiving routes from the system manager. The information might include time and channel information. - Co-chairs indicated that WG list will be polled for adoption of this draft. 5) Industrial Deterministic Routing Extension for Low-Power and Lossy Networks (M. Wei - 10mn) [45] http://tools.ietf.org/agenda/85/slides/slides-85-roll-4.pdf - Pascal highlighted that a number of discussions are still required. These include time sync for RPL and abstraction of the slots and subsequent notification. Additionally what is required for layer-3 and how IPv6 sees the channels? Presenter agreed to discuss these requirements and architecture further. - Dominic requested clarification on the motivation for the metric (and origin) being proposed and what the objectives are (including reducing time to end-node). Co-chairs highlighted that creating dependencies (piggybacking MAC layer to L3) between layers is dangerous. Generally cross-layer dependencies have caused problems for IETF technologies. - Robert highlighted that the document makes too many assumptions that you can use L2 information at L3, what would happen if you tried to route from one LLN to another? Presenter responded that they did not consider routing across the backbone. 6) Multicast Protocol for Low power and Lossy Networks (MPL) draft-ietf-roll-trickle-mcast-02 (Jonathan - 10mn) [55] - General comment on operatiocnal complexity and the group is still unclear on what users will want to use multicast for. Furthermore, perhaps forwarding mechanisms can be proposed in the draft while further discussion takes place on deployment and use cases. - Some further comments made regarding how packets would be processed and inspection of outer and inner headers. Co-chairs asked group to respond and comment on list with use cases and selection of multicast mechanisms. - Co-chairs highlighted that although the document was in last call the consensus is that additional discussion and work is needed, also security aspects also need to be considered. - Co-chairs asked AD (Adrian) if this work should be progressed in this (ROLL) working group. Conclusion was that the work should continue in ROLL for now and updates on progress (via the mailing list) should be provided to the other relevant working groups including 6man working group. 7) Loop Free DODAG Local Repair (Jianlin Guo - 10mn) [65] draft-guo-roll-loop-free-dodag-repair - General comments mentioning that the described function looks nice but how necessary is it? Seeing as RPL has loop or repair detect, you can accept the risk and detect it or change the poison timers. Is anything else actually required? - Co-chairs polled the room asking of WG should pursue the work? No hands shown. 8) RPL Routing Pathology In a Network With a Mix of Nodes Operating in Storing and Non-Storing Modes (JeongGil Ko - 10 mn) [75] draft-ko-roll-mix-network-pathology-01 - Comment made asking how a root node knows to insert a routing header or not? Presenter responded that first thing to check is route to node, and next is the next node a storing node or not. If next hop stores routes then you will send to the next hop as that node will insert routing header, if next hop is non-storing node then current node is responsible for inserting a routing header. - A few comments made that supporting mix nodes with RPL may be a slippery slope. One issue is that storing nodes in the middle of the network to have full knowledge of the topology. Also is this functionality required now or is this future work. Additional this functionality may add significant overhead. A trac ticket was previously created that actually summarises this issue and the authors should look to review the ticket and address in the draft.