OPS-AREA Open meeting minutes (based on notes taken by Bert Wijnen and Joel Jaeggli) Monday, July 26, 2010 Afternoon Session II Area Directors and Meeting Chairs: Dan Romascanu and Ron Bonica (remote) agenda: http://www.ietf.org/proceedings/78/agenda/opsarea.txt meeting-materials: https://datatracker.ietf.org/meeting/78/materials.html#wg-opsarea 1. Note well, agenda bashing, note takers, jabber scribe, blue sheets - 5 min 2. OAM Workshop proposal and plans - Dan Romascanu for Ron Bonica - 15 min slide 4 at: http://www.ietf.org/proceedings/78/slides/opsarea-3.ppt planned for October. OAM for the forwarding plane. Dan reads the announcement that is expected to be done Wednesday during the plenary. It is an IAB-IESG joint workshop October 12-14. George Mason University, Fairfax VA 3. Revision of RFC 1052 - Dan Romascanu - 5 min Dan: Quite a few hands went up on: who knows this document Dan: the document is from late 80s, so over 20 years old. Was the basis for NM architecture for the last 20 years in IETF Some things they got quite right. Separation between data modeling language and protocols. Idea was to have one data model which would work over multiple protocols. Prognosis that CMIP would become THE PROTOCOL in the NM space did not happen. The temporary solution (SNMP) lasted a couple of decades. So right now we;re in 21st century. NM landscape has change to large extent. We no longer believe in "one protocol fits all", so the concept now is "a multiprotocol toolkit" So IESG discussed the situations and also approached the IAB to revise RFC1052. But IAB is welcoming the initiative, and will put 1052 to historic and would encourage, possibly sponsor a revision of the doc, but there is no expertise in the IAB for this at this point. So Dan is asking the OPS AREA community if we can work on a revision of 1052. This type of work needs good consideration and serious resources. Joel Jaegly: it seems that that doc was quite easy to produce a synthesis at that time. But even in the OPS area it seems an insurmountable effort at this point in time,. Maybe a workshop would produce some valuable effort. This seems to be a larger task than what we are capable of taking on. Dan: the world and the Internet is now much more complex than 20 yrs ago, so the magnitude and degree of the work is now much more complex. Maybe start with an interaction with the operators. WE are now approaching a broader space. JJ: Yeah. Bert: At 1052 time we had a big problem (inter-SDO) and inside IETF (HEMS, SGMP..) so we needed this doc The IAB workshop in 2002 we had a similar problem (operators telling us they were not using SNMP). DO we have a similar problem? Dan, No, not so much operational problem like 8 years ago but we better have an architecture-like document upon which we can base our decisions when chartering WGs Wes: I think we need such an architecture document (used his lunch discussion on some possible future work to use NETCONF but not YANG) to guide us as to when to use/suggest using which protocols. Dan: I thought about what would go in the document. I have not yet figured that out, but I think we would not need to list protocols Wes: but we should at least document the wisdom we have. JJ: it strikes me as congruent with experience in other areas. Using NETCONF with bolting other language onto it seems bad. We could encode that in an architectural principle. Andy Bierman: Using something other than YANG, then they talk about XSD or RELAXNG, otherwise it would never fly. If someone really wanted to use XSD and just depend on description clauses, I think we should reject that. Dan: this I consider as a long term item and it is a trigger that hopefully leads to something. 4. TM Forum work on data models - Nigel Davis - 30 min Nigel Davis. Presents a slide set http://www.ietf.org/proceedings/78/slides/opsarea-1-1.ppt aim: - To identify key areas for collaboration between IETF and TeleManagement Forum (TMF) and to poll for interest and potential participants - To identify areas for potential work in IETF Nigel goes over the - materials/tools/methods they have/use in TMF (slides 3-5) - modelmapping and models (slides 6-8) - Open source tooling (slide 9) - Concept area overlap (slide 11) - Why engage? (slide 12) - conclusions (slide 13): - Working together, a list if questions like: - where - what - how - Possible formats to work together (how can you engage?) Q&A: Yuri (from UVA): WE need to follow lifecycle (we found this useful) what is your opinion? Nigel: TMF is looking at the whole lifecycle management. Some network modeling activities in ITU-T SG15 I agree we need to understand the cycle as a whole. Yuri: NGOS is a development of eTOM. Yuri: there is interesting work in ITU-T: NGM convergence model And there also a few WGs at OpenGrid forum: NML/NDL .. etc Dan: you mentioned JCA: Joint Coordination Activity. IETF has an unofficial participation (AD sits in meetings as individual and loosely liaising). They produced guidance docs, mappings etc etc IETF did suggest to include things like SNMP, MIBs, SMI, NETCONF etc Dan: If you start to develop any YANG modules, than I expect we will be willing to help and review/assist/advise. Also if there are specific requirements or pieces of work, they could be suggested as work to be done in IETF. To start work, we have BOF, BARBOF etc.. and WGs can be formed. 5. Demo on modeling Physical Resource Models with YANG Complex Types - Bernd Linowski - 10 min Bernd presents a slide set http://www.ietf.org/proceedings/78/slides/opsarea-0.ppt Q&A Dan: What do you plan to do with this work? Bernd: probably want to publish as an experimental document Mehmet: we implemented this twice. It is available as open source so we asked AD if he would support it as experimental document. And we may want a YANG expert to check if there are conflicts from a YANG point of view. Dan: David Kessens, can you put this question into the NETMOD WG. David Kessens: yes will do 6. Management of 6lowpan networks - Juergen Schoenwaelder - 15 min Juergen presents a slide set http://www.ietf.org/proceedings/78/slides/opsarea-2.pdf Q&A Wes: the results is very valuable, but I'd suggest to make it more generic. There are lots of small things out there that may need this. So pls absolutely continue; Dan: what does it mean "absolutely continue?" Wes: Yes, make it an informational. A lessons learned type of document. Dan: Juergen, is there enough common space between this and the smart grid (running SNMP in constrained environment); Bob Cole: In manet WG we are struggling with some of this (we'de be interested in the proxy stuff you're doing) also in the DTN RG. Benoit: When are we going to reach the limits in terms of OIDs we can have and how many we can poll? Juergen: On 6lowPan? Benoit: Yes Juergen: Our experience is that some devices work better than other so there seems no generic answer. Andy: This is interesting. There are aspects that may relate, like other protocols/mechanisms that work of from some MIB objects (like ifIndex). If other protocols are going to be dependent on some SNMP/MIB data, would such devices then be required to do SNMP? Your numbers are impressive (like 16 bytes for an object). Juergen: what makes sense on a device very much depends on that device. My idea was: if you want to run SNMP in a resource constrained environment, what does that mean and what can you do I will probably take Wes's advice and make it more generic. 7. WG status review - 3 min each WG Dan: proposes to skip this to allow the other topics to be done 8. Juergen Quitteck on: Energy Management - we've discussed for a year - we started with a small MIB - we found there is much more behind it - need more than just MIB modules we may also need a yang model for management. - we want to propose to create a separate WG for this Juergen then presents the slides http://www.ietf.org/proceedings/78/slides/opsarea-4.pdf Q&A Juergen Q: Should this be in OPSAWG or in separate WG Andy: I think it should be its own WG Anwar: What expertise we need to do this in IETF JQ: this is a new issue. People working in this are working on this in their companies John Parello There are other companies that really work on this and are participating in this work. Would be valuable to bring in people in for external source e.g.schneiber electric and other power distribution and devices vendors. Dan: if another forum wants to do it we can also assist them. If such a forum does not exists or if they think that they rather see it done in IETF (where this type of NM expertise exists), then we can do it. JQ: People working on these drafts (well known IETFers) are working on devices in this space in their companies. Anwar: I am supportive of this work Dan: straw poll: Who does really not care about the issue? No hands, So there is strong support for this work Dan: - Who thinks we need a separate WG: half of the room - In OPSAWG: no hands - Who thinks we should not do this in IETF: no hands 8. Open mic - any time left No time for this