OpsAWG / Ops-Area Minutes taken by Tom Taylor OPS Area Begun 02:00. (Times are recording offsets.) NOTE WELL displayed. Audio recording at: http://www.ietf.org/audio/ietf90/ietf90-ontario-20140723-0900-am1.mp3 1. Administrivia - scribes / minutes, etc Slides at www.ietf.org/proceedings/90/slides/slides-90-opsawg-0.pdf ADs Benoit Claise and Joel Jaegglia chairing. Benoit presentation (slides as indicated above). Concerns as AD. He has been trying various solutions to increase energy level and productivity. - weekly interim meetings - design teams for specific Yang modules - Github Juergen Schoenwaelder provided more details on how the teams have been using Github to interact, and testified that their productivity has improved. 10:00 2. Potential new work: Transport Independent OAM in Multi-Layer network Entity (TIME) 10 min presentation by Qin Wu 10 min discussion draft-ww-opsawg-multi-layer-oam Slides at http://www.ietf.org/proceedings/90/slides/slides-90-opsawg-10.pdf 2. Discussion of TIME 3. Background 4. The Problem 5. "" 6. Strong technology dependencies 7. Layering OAM architecture in management plane 8. Weakness of cross-layer OAM 9. Lack of OAM above Layer 3 10. Issues of abstraction (cut off) 21:39 Skip to 12 Planned TIME activities for lack of time The key issue is OAM for overlay networks. Sounded like Peter Frumm, T-Systems, but no such registrant: He was surprised to see his name as contributor on second slide. On last slide, protocol development action item: not likely the proper point of attack. Instead, need much strengthened management plane to tie things together. Major concern shared by Qin is to avoid layer violations. Erik Nordmark question: scope. Seems like ping and trace are out of scope (speaking to a particular slide). (Qin not clear.) Erik also had an architectural point: to correlate between layers and also with topologies being developed in other WGs, one of the key issues is a common system of naming. 31:00 3. Potential new work: Network Function Virtualisation Configuration (NFVCon) 10 min presentation by Haibin Song 10 min discussion Slides at http://www.ietf.org/proceedings/90/slides/slides-90-opsawg-6.pptx - History - What are the main use cases - Protocol in scope - Data model in scope - Out of scope - Gap analysis - Security considerations - IETF related work 39:20 Comment: premature -- the ETSI Energy group is working on this but have not yet determined whether the work makes sense. Haibin responded that the work could be aligned. The commenter noted that although ETSI up to now has refrained from development of data models and protocols, there is some impulse to do work related to particular country codes. Georgios Karagiannis: set up a liaison to coordinate? Important to cooperate. General agreement to avoid duplication. 43:50 4. github as tool - 15 min presentation by Martin Thomson Slides at ?? For managing the production process for I-Ds. Have had quite a bit of success using Github in a number of WGs including HTTPbis. Editorials simply accepted -- push of a button. Tools have been developed to move Github documents into the IETF submission process. Scott Bradner: discussion in IETF Trust. Working with Github would be considered an IETF contribution for IPR disclosure purposes. Martin described work done to make sure people are aware that by particpating they incur IPR disclosure obligations. Robert Sparks: dependent on Github to maintain record relevant to IPR claims. Point made that this implies significant changes in working procedures -- discipline in passing Github comments to list. Does require a lot of training as well as discipline. Speaker favours the direction. Concern that we have a complicated document development chain already, and tracing through from marked-up source to final document will be even more challenging. Suggestion of tools to integrate with Github. Martin's summation: key value is improving collaboration. Looking at best engineering practice. Fred Baker (?) noted that the application would use a small fraction of Github's capabilities. Because of those capabilities, it is possible to do oneself considerable damage. 1:00 5. open mic No comments. 1:01:50 OpsAWG 1. Admin / changing seats, etc. chairs Welcome slides: see continuation of: http://www.ietf.org/proceedings/90/slides/slides-90-opsawg-0.pdf Warren Kumari presented Note Well and agenda. Latter accepted. 2. document updates document authors Scott Bradner and Warren Kumari chaired this segment. Status updates on IDs requested by ADs. Chairs had not expected slides, but some were provided. draft-ietf-opsawg-capwap-alt-tunnel - Major revision to be posted shortly - Next version should be ready for WGLC draft-ietf-opsawg-capwap-extension - Gang Liu: IEEE 802.11 review just completed, comments to resolve, - Next vversion should be ready for WGLC draft-ietf-opsawg-capwap-hybridmac - New version posted recently based on comments draft-ietf-opsawg-coman-probstate-reqs draft-ietf-opsawg-coman-use-cases - In WGLC. One set of comments received. Looking for more. draft-ietf-opsawg-large-flow-load-balancing - Working on comments received from OPS-DIR, SEC-DIR, GEN-ART. - Main update on security issue, other issues. - Expect update in next few weeks. draft-ietf-opsawg-vmm-mib Slides: http://www.ietf.org/proceedings/90/slides/slides-90-opsawg-9.pdf - Read-write issue -- resolved. - Michael saying cannot guarantee notification settings in low-level code - New bullets an improvement draft-taylor-opsawg-mibs-to-ieee80231 - Ready for WGLC. - Submit as draft-ietf and WG will do it. draft-tempia-opsawg-p3m - No report draft-winter-opsawg-eap-metadata - Slides at: http://www.ietf.org/proceedings/90/slides/slides-90-opsawg-11.pdf draft-ww-opsawg-multi-layer-oam - Expect -01 draft well before IETF91 draft-xue-opsawg-capwap-alt-tunnel-information - Being submitted as WG document - Next version should be ready for WGLC 1:20:20 3. Operators and IETF 20 min presentation by Jan Zorz Slides at http://www.ietf.org/proceedings/90/slides/slides-90-opsawg-8.ppt Jan Zorz introduced himself, operator for many years before joining ISOC. On-line survey results. More to come. Scott Bradner comment: no mechanism for feedback to the IETF on how well a protocol was designed. Used to have working groups for the purpose, but the IESG of the time dissolved it because of its open-endedness. Perhaps need to reconstitute such a feedback mechanism. Lee Howard: budget thing -- lot of people above have to approve, so managers don't really have the power Jan suggests. Would be interested in working with WG Chairs to provide a digest of current work and instructions on participation, to be presented at a 'NOG meeting. Confirmation that this would be very helpful. Benoit Claise: what new must we do? Jan: intend to glean operator suggestions and present them in the I-D. Michael Behringer question re sample selection. Self-selected after lightning talk requesting participation at the 'NOG (NANOG, RIPE) meetings. Suggestion that other organizations less likely to be familiar with IETF and therefore better targets for outreach. Lots of relating of experiences in attempted participation and outreach. - operator not sure her comments were taken into account - plans for outreach at LACNOG Santiago end October, Sao Paulo December Brian Carpenter: tension between ops people and IETF because IETF is working to change their way of working. Fact of life. Also, outreach has to be responsibility of people in their own countries -- outreach from a central group does not scale. Don't have the same people who worked on outreach in the past. Opportunity is still there, will still get good feedback. Wes George: cultural issue, love-hate relationship. Sometimes a matter of "IETF, what you're doing hurts, please stop!", but sometimes: "I have a problem and you're not helping me with it." The oft-times IETF response: "You're doing it wrong!" is not helpful. Religious -- have to stop doing that. Warren -- would it be interesting to survey IETF participants to see if they consider themselves to be operators? Mike: problem with time horizon. IETF too slow for implementers. Useful stuff has to have much shorter time scales. Cost and risk of investment over a number of years. Kent Watson: don't want to lower the bar. Mailing list more reactive, but regular interim meetings proactive. Wes Hardaker: thanks for doing the survey. Has presented work in progress, gathered feedback at different 'NOGs. Anyone coming to an IETF meeting for the first time will have a very hard time understanding what is going on - one explanation for non-attendance. DNSOps trying to make it easier for newcomers. Sometimes a matter of relative priority -- many claims on their time and budget. Have to make relevance to them clear so they can evaluate priority correctly. NANOG plus IETF would be 6 meetings per year -- difficult for many people. Affects family. Important to have agendas published early. 2:04:00 4. PCAPNG: The default file format of Wireshark Michel Tuxen Slides at http://www.ietf.org/proceedings/90/slides/slides-90-opsawg-7.pdf Wants to bring work on PCAPNG file format to IETF PCAPNG is actually a new file format, not just a second version of PCAP. Brian Carpenter: PCAP vs. PCAPNG -- not interoperable. Could a PCAPNG file be exported from WIRESHARK and interpreted by a PCAP system? Conditional. How is the user community with this? Orthogonal to file format. Metadata includes filters used. Ops? Network Management? Scott Bradner: both covered by OPSAWG. Example of prototyping where this would be really useful. Some interest in the room to work on the document and review it. Warren Kumari: willing to work on it. Not a frequent type of job for the IETF. Advice: find every tool you can and then bring the work in to standardize. Wikipedia has an extensive list of tools. IETF are not the experts in this area. Microsoft: standardized format the key outcome. Warren: will talk it over with the ADs. Michael: want networking expertise to make sure all desirable information is captured by the format. 2:20:45 5. Open mic Benoit summary: what we did discuss. BUT should have been discussing open issues in our WG documents here. Blew many travel budgets without accomplishing primary aim. Scott partially accepted the admonishment. Meeting closed at 2:23:00