[ Thanks to Barbara Stark for taking minutes - Chairs ] opsawg Minutes; Wednesday, April 6, 2016, 1400-1600 Chairs: Warren Kumari Scott Bradner Tianran Zhou Minutes: Barbara Stark Jabber: Joel Jaeggli Scott was presented with shirt and farewell speech by ADs, standing ovation, and a rousing song in Dutch by Bert Wijnen. Agenda posted at https://datatracker.ietf.org/meeting/95/agenda/opsawg/ Meeting materials posted at: https://datatracker.ietf.org/meeting/95/materials.html/ Warren went through Administrivia slides 1-5 (https://www.ietf.org/proceedings/95/slides/slides-95-opsawg-4.pdf). Eliot Lear presented slides on Manufacturer Usage Descriptions (MUD) https://www.ietf.org/proceedings/95/slides/slides-95-opsawg-2.pdf Eliot presented all slides without interruption. Dan Romascanu: It is important to clarify relationship with vendors. Eliot: Agree. For example, need to determine who signs certificates. Andy Bierman: Likes draft. IoT security needs to be figured out. Benoit Claise: Need to figure out home for this work. Additional support was expressed. Dan Romascanu presented on Automatic attachment of end stations and network devices Dataplane Probing https://www.ietf.org/proceedings/95/slides/slides-95-opsawg-0.pdf Dan presented all slides without substantive interruption. Eliot: Interesting work. Want to work with Dan on this and look forward to doing that. Shahram Davari: Why here and not IETF? What new protocol is needed? Dan: Can be applicable to other types of tunnels. Not asking for new protocol. Describing solution to an operational problem. Shahram: There seems to be a mix of requirements, discussion, other in doc. Dan: Yes. It needs to be cleaned up. Plan to split documents in separate I-Ds. Diego Lopez expressed interest and commented. Petr Lapukhov presented on Dataplane probing https://www.ietf.org/proceedings/95/slides/slides-95-opsawg-1.pdf Petr presented all slides without interruption. Linda Dunbar: Do you use Data Plane or a separate management channel? Petr: It's internal to the device. Warren Kumari: Does it sound like silicon vendors might be willing to do this? Petr: Talking to some and they might be willing. Shahram: I like this. -------- end of opsawg ; start of ops area -------------------------- Ops Area (see slides 8-9 of the agenda) Ops Area ADs: Benoit Claise Joel Jaeggli Mehmet Ersue presented on "NFV Trends and OPS work?", slides 1-6 https://www.ietf.org/proceedings/95/slides/slides-95-opsawg-5.pdf Benoit presented slides 7-11, and then asked "What should Ops Area do?" Linda: There are many parts. How do you ensure policy consistency between multiple instances? How do you monitor if an instance is under attack? Resource management. Will provide next time some strawman thinking to help put together a more solid proposal. Benoit: Need to identify gaps. Dan: The output should be an NFV taxonomy document (as suggested on slide 11). We have avoided defining "service" for a long time. We need to define "service". Benoit: We definitely need to determine IETF focus. There is work on SLAs. Hui Deng: We need to unify parts to deploy end-to-end service. Diego Lopez: 3 aspect we must solve. Plaftorm / mobility; ; Mehmet: 2 points -- the first concerned assurance ; the second is that the taxonomy work is important and many orgs are craving such a product Ram Krishnan: Tons of information is available in open source around Openstack. Maybe there is something to bring in. Chongfeng Xie: It may also be useful to provide operational experiences. Benoit: Who wants to participate in creating the taxonomy? Who wants to participate in the gap analysis? Several hands were raised for each question. Anyone who is interested should contact Benoit. Tianran Zhou presented on Service Component Composition https://www.ietf.org/proceedings/95/slides/slides-95-opsawg-3.pdf Open Mic: Lucy Lynch: Thank you to Scott. You will be missed.