Routing Area Open Meeting (rtgarea) IETF 100 (Singapore) =============================================================================== Area Directors: Alia Atlas (akatlas@juniper.net) Deborah Brungard (db3546@att.com) Alvaro Retana (aretana@cisco.com) Area Secretary: Jonathan Hardwick (jonathan.hardwick@metaswitch.com) Wiki: https://trac.tools.ietf.org/area/rtg/trac/wiki/WikiStart Scribe: Jonathan Hardwick (jonathan.hardwick@metaswitch.com) Location: Canning, Raffles City Convention Center, Singapore Time: November 16, 2017, 1550-1750 (3:50pm-5:50pm) ------------------------------------------------------------------------------- 1. Administrivia ---------------- Chris Hopps: How many attendees are in fact WG chairs? Attendance seems low. Is this just another chairs meeting? (show of hands showed that about 50% are chairs, 50% not. A few people are at their first IETF - welcome!) Chris Hopps: Should we think about combining the RTG chairs meeting with this one? Alvaro Retana: There may be some topics we can bring in from the RTG chairs meeting, but it would not have worked this time due to the scheduling constraints of the SEC ADs. Alia Atlas: We encourage chairs to look for document shepherds on their mailing lists. We encourage more people to take an interest in reviewing documents! Alia Atlas: I will be standing down as AD at the next IETF. I expect to close TRILL and to put I2RS into hiatus before stepping down. I also expect to merge the OSPF and ISIS working groups and to recharter SFC. Deborah Brungard: Please get your NomCom comments in for the new Routing AD. The feedback period runs to the end of next week Alvaro Retana: You can give feedback to NomCom about candidates, or also on what you think is needed in the routing area, so they can talk to the candidates about that. Alvaro Retana: In MANET, we will have a discussion about the future of the working group. The spring working group is to be rechartered. ------------------------------------------------------------------------------- 2. Dynamic Link Exchange Protocol (DLEP) ------------------------------------- Rick Taylor Greg Mirsky: You should look at the work on links with variable discrete bandwidth in CCAMP and TEAS with RSVP-TE and IGP extensions. Amy Ye: I work on microwave links. But in our device we have layer 3 functions. Your requirements sound quite different. Rick Taylor: We are aiming to reach out to people like you. There is crossover. We are trying to get microwave people to add DLEP to their links. DLEP is an event driven way of driving routing, not prescriptive, so may differ from CCAMP's approach. Justin Dean: I am the MANET chair. In your presentation, you mentioned that we had developed DLEP in isolation. I don't agree we've done it in isolation. Three radio vendors have adopted this and put it into their radios. Now we want routing people to look at it. Rick Taylor: I meant that there has not been much exposure in the IETF. Justin Dean: DLEP can help routing protocols decide when to take a link up or bring it down. When we first started in MANET, we did not have DLEP so we had to rely on network drivers. Without DLEP it is hard to route around a problem link in a mesh radio network. Greg Mirsky: Y.1731 has a 'bandwidth notification extension' (the extension is BN.1731) which may perform a similar function. We noticed this when we were adding the discrete bandwidth available enhancement to OSPF and RSVP-TE. It is worth being aware of. Rick Taylor: SDN is prescriptive but DLEP is reactive. Making them work together is a challenge. BFD and DLEP should work together really well. Once DLEP has told you a links is there, BFD can keep it up. We also want to talk to the other routing protocols to find out if there are any best practices for using DLEP with those protocols. We also have this question to the ADs: is MANET really the right place for DLEP? IJsbrand Wijnands: Multicast does not work well over wireless. Do you have anything to improve it? Rick Taylor: DLEP does not do anything specifically for optimizing multicast. It allows a radio system that does do multicast to tell the router about it - but it is up to the radio system how the multicast works. Alvaro Retana: MANET has something in the charter for wireless propagation of multicast. PIM has also discussed it. Lou Berger: DLEP supports radios with an Ethernet service model or an IP service model. There are interesting differences to the use cases. Rick Taylor: We have work ongoing to make IP service model more usable. Lou Berger: DLEP is a reporting protocol, so in a variable rate system it will give notifications about the changes in bandwidth. This is a different model to the CCAMP model that Amy mentioned. Rick Taylor: Agreed. DLEP is primarily a channel to allow reporting events up to the router. It is not used for management, so you can't use it to set link attributes, bandwidth etc. Deborah Brungard: So instead of using SDN to configure parameters, you learn it from the link layer. This is different to CCAMP's model. Unlike LMP, where you are managing the link properties. Rick Taylor: In SDN the controller may assume it can set the link bandwidth e.g. I need 173 Mb. With DLEP you get told that the bandwidth is only 120Mb, so you don't just read the config file and assume that it's 173 Mb when it's not. Jeff Tantsura: You could use BGP-LS northbound to inform the controller when the effective bandwidth changes, but it's not clear whether this would be fast enough. The timing could be critical. Lou Berger: DLEP is like LMP for a different market sector. It would have made sense to combine them. DLEP is still growing. Alvaro Retana: If you have thoughts on where we should work on DLEP, let us know. Come to MANET and tell us. ------------------------------------------------------------------------------- 3. Working Group and BoF Reports -------------------------------- - Routing Directorate Report - WG Chair Reports See the wiki for the contents of the reports: https://trac.ietf.org/trac/rtg/wiki/IETF100summary Only comments that are not readable in the wiki are recorded in the minutes. BIER: (Tony Przygienda) The BIER architecture was published. So we went out for beers until 3am. Which is why Ice has no voice. Now we need to recharter. (Alia Atlas) I aim to recharter BIER before stepping down. OSPF: (Alvaro Retana) The segment routing architecture is now scheduled for a telechat, which will unblock the OSPF segment routing draft. (Alia) Wouldn't it be great if the ISIS segment routing draft could be ready at the same time? ------------------------------------------------------------------------------- 4. Open Discussion / Any other business ---------------------------------------- Alia Atlas: Please be aware of the following. (i) Path Aware Research Group (it would be nice to have larger routing presence there). (ii) ANRW in Montreal for IETF 102 is open for papers. Alia Atlas: Question 1: How many are active in any of the open source projects related to networking? (quite a few) Is anybody unaware of these projects? (nobody) Qusetion 2: How many have ideas of how we can engage better with it? (A few) Question 3: Do you feel we can be more aware of what is going on elsewhere? (Universally, yes) Jeff Tantsura: In RTGWG, we have several interesting topics, including virtualization, slicing and multi-access. If you have other topics then please bring them. Deborah Brungard: Please make more use of the hackathon. Recently, PCE won an award there. More of this, please. Meeting closed at 1745. ===============================================================================