IETF 86 Technical Plenary Minutes Orlando, FL, USA Monday, 11 March 2013 Minutes by Cindy Morgan, IETF Secretariat 1. Welcome Bernard Aboba welcomed the audience to the IETF 86 Technical Plenary. 2. Reporting 2.1. IRTF Chair Lars Eggert delivered the IRTF Chair report: http://www.ietf.org/proceedings/86/slides/slides-86-iab-techplenary-6.pptx 2.2. IAB Chair Bernard Aboba delivered the IAB Chair report: http://www.ietf.org/proceedings/86/slides/slides-86-iab-techplenary-3.ppt Bernard announced that Russ Housley has been selected by the board as IAB Chair for 2013-2014. 3. Technical Topic I: The End of Plain Old Telephone Service (POTS) Hannes Tschofenig introduced the first technical topic for the evening, exploring the technical, policy and regulatory issues surrounding the de-commissioning of Plain Old Telephone Service (POTS). Henning Schulzrinne, Chief Technical Officer, US Federal Communications Commission, delivered the following presentation: http://www.ietf.org/proceedings/86/slides/slides-86-iab-techplenary-4.pptx The IAB has also set up a wiki page with information on the End of POTS: http://trac.tools.ietf.org/group/iab/trac/wiki At the end of the presentation, the microphones were opened for questions from the audience. A summary of the Q-and-A session follows: JOHN LEVINE: I'm confused between IP technology and the Internet, because the requirements of a phone network and the requirements of the Internet are totally different. The Internet is voluntary interconnection, and it throws packets away with wild abandon. The phone system is mandatory interconnection, and you want low jitter and high reliability and stuff. I know you can build them both out of IP routers, but does it make sense to make them one network, or does it make sense to have them be two networks doing logically separate things? HENNING SCHULZRINNE: That's a good question, and I don't think there's going to be a single answer. Increasingly, I think you'll see three models that are emerging: (1) physically separate networks on different wavelengths, for the voice over IP, PSTN version of it--a packet cable type of thing; (2) completely integrated over-the-top things like Vonage in the U.S.; and increasingly (3) a middle ground where the separation is virtual--a large carrier does interconnection through VPNs across the Internet, using a session border controller so you don't just interconnect randomly. It's a VPN, not a separate physical network; that will become increasingly common because of the costs. Voice is such an overall small fraction of the traffic that it will become fairly difficult on an interconnection basis to just maintain a completely separate network between distant networks. You'll either end up with a separate transport network, or we'll end up with Internet transport. Regardless of whether it's logically separated, physically it will be the same--the same fiber, the same MPLS environment, the same network operations system. The whole reason for doing that is for cost saving. So I wouldn't say that they'll remain separate networks. BHUMIP KHASNABISH: You mentioned there is no cutover date, but what we're increasingly seeing in some enterprises is that there is no PSTN network anymore--you cannot send any faxes. So when you say there is no cutover date, is that a general statement, or about enterprises versus residential networks? HENNING SCHULZRINNE: It's not likely that a country will say as of X date we will turn off analog circuits and only IP is available. What will happen is more regional types of cutovers, like your phone company saying, "If you want TDM you're responsible for providing your own conversion of circuits over packets technology; we won't provide a TDM circuit to you anymore." Or, "We will not provide an analog loop to you anymore; you'll have to use our voice over IP product. You'll still get the same voice service, but it's no longer provided by an analog copper loop." That will happen individually--at least in the U.S.--as opposed to a national mandate. If you think of a transition from analog to digital, that's pretty much what we've gone through before. There are still analog switches, and there is still multi-frequency signaling out in the network. These transitions tend to be fairly gradual. Obviously in some areas, companies will simply say, "Sorry, no more old services for you." Whether they're allowed to do that is a policy issue. BHUMIP KHASNABISH: There are many residential customers whose home security services run on those POTS lines, so what should those consumers do? HENNING SCHULZRINNE: That is a question [the FCC is] struggling with. I don't think there's a good answer to that yet. PHIL HALLAM-BAKER: I'm less interested in how the actual messages get there and back than I am in the actual user experience. We're still trapped in the old PSTN user experience because the handset hasn't changed. And then you have the Internet handset, going through your computer, and you have usability problem. I mean, you pick up the phone, and it works. You start Skype, and then it tells you that you have to upgrade. So you have to wait 5 minutes for that. I have literally spent 15 minutes doing sysadmin to make a voice call. Skype is bad, Google Talk is bad, and that is the thing that the room here needs to fix. Until you can make doing a voice call over the Internet as easy as using the telephone system is today, it's not done. [Applause.] HENNING SCHULZRINNE: Yup. There's not much to add to that. LEE HOWARD: That's not my problem. [Laughter from the audience.] The fact that the packet-switched networks don't work the same as circuit- switched networks is part of why they're cheaper. Maybe the users of packet-switched networks didn't have the same requirements for the level of voice [quality]. They said, "I can get voice calls over Skype, and it's free! That's amazing, as compared to paying a telco to carry it over a packet-switched network." Maybe there is a reason for that difference in cost. Just because this network doesn't have the same application as the network you want doesn't mean I have to re-engineer my network. HENNING SCHULZRINNE: We can either be satisfied to simply say we'll have a TDM-circuit switched network forever because of the reason you mentioned, or they can trust us as a technical community to be able to engineer reliable networks that are at least as good as the existing TDM network. It's up to you which of the challenges you accept. KEVIN FALL: It seems that largely the innovation of the Internet has been software delivered over the Internet, but we deliver the software with end-user license agreements that go on for pages and pretty much disclaim anything. "Don't use this for anything safety-critical; don't use this for anything financial, etc." If the Internet is what we have to depend on, do you see a change in procurement or something related to software assurance that gives us a framework in which we might be able to get something more than, "We disclaim all uses for this particular piece of software"? HENNING SCHULZRINNE: I see the beginnings of one that's coming from a different angle, in cyber-security. It was acceptable until recently that if a certain widely-used programming language happened to botch your system, well, that's what you got for using a free service, even if it's maintained by a very large company. Generally speaking, if you look at any technology, as soon as it starts affecting real people in real ways, lawyers find a way to make sure it won't be a totally one- sided deal. Regulation will step in at some point. But to disclaim everything and say good luck is not a longterm acceptable solution. I don't see a procurement one, but I do see an awareness that will have a lot more emphasis on validation, acceptability testing, interoperability testing. DAN WEBER: Right now we have a situation on our Internet services where there is no guaranteed quality-of-service for just about anything. Will there be an FCC mandate to ensure that quality-of-service is guaranteed amongst ISPs, where we don't even know how they're handing our traffic right now? HENNING SCHULZRINNE: There's many ways to look at quality, and I can't speak for the commission, but the notion of mandates with fines seems unlikely. Everybody hopes that the first choice of regulators today is that competition takes care of these issues, which means competition plus disclosure. Make sure consumers know which providers have what quality of service. In a general sense, making sure consumers know which service providers provide which quality-of-service. One would hope that the cream of these networks rises to the top, and you will actually have networks that are reliable. That doesn't mean that there won't be requirements, especially when public safety is involved; making sure that the network has a reasonable architecture that can survive expected disasters, storms, etc. KEITH MOORE: You said there's no cutoff date, but I think there will be some sort of date at which these carriers will be no longer required to provide this last-resort universal service. Am I wrong about that? HENNING SCHULZRINNE: That actually gets more complicated than I could put in a slide. What makes this difficult in the U.S. is that there are two levels of telecom regulation: state and federal. Wireless is generally federal, but landlines in many states are still regulated. Many of these obligations are state-level regulations. There are about 20 states that have these last-resort obligations in one way or another. That is a process that is ongoing, and not everyone is happy with that. DAVE CROCKER: I was fascinated to hear the reference to caller-ID spoofing. It occurs to me that this is where VOIP can leapfrog the PSTN, because we could solve this almost tomorrow by using DKIM. I mean that only a little bit seriously. It's really an entry into that we have access to far better meta-data exchanges. I think this is all about packaging and what's provided to users, and we're not likely to get that from the current providers because they've got an installed base that they're supporting. There's massive business opportunities if people actually try to do better phone service, not emulated phone service. HENNING SCHULZRINNE: Thank you; that's exactly what I think we should be striving for, and I am hoping that as we take up the challenge that we can get, for example, secure identities relatively quickly. But that requires, in some cases--if you work for a carrier--that you talk to your management to make that happen. We're still pushing a rock uphill on that one. OLAF KOLKMAN: In this Q-and-A there is a lot of resonance with the regulatory aspects and the discussions, debates, and negotiations that went on during WCIT. In your answers, you gave a very western- hemisphere regulatory perspective. I think we should recognize that there are other regulatory perspectives we should consider. These came into the WCIT debates. I think Bernard already mentioned that there will be an IAB session summarizing the WCIT results, and I think if you are interested in the debates and want to know what happened there it would be valuable for people to go to that session. HENNING SCHULZRINNE: One of the difficulties that a policy discussion has on a national level is that many of the nice divisions that we used to have, where telecom was safely on one side and the Internet was separated on the other, are now unfortunately gone. Many of the distinctions are now legal rather than technical. PETER LOTHBERG: I'm getting a little confused, because I'm having a hard time telling the difference when you're saying telephone this and telephone that. In the Internet world, we have a network that moves packets. On top of that, we have an application, and the application in this case is called voice. If you look at the problem that way, the service provider of voice is the end user. You have 300 million phone companies in the U.S. that you have to regulate; everybody becomes a phone company because they are providing a voice service. Or, I'm paying someone to provide and monitor equipment at my house to provide the voice service so I can pick up the telephone. In the old days, the phone network was a purposeful network that delivered the phone service. Now, you can of course build a purposeful network using Internet technology to deliver voice service, but that's not what we're going to do, because it's not cost-effective. So, I think you have to figure out how to get the wording right--who is providing the service, who is doing voice, etc., so that we don't get this mishmash where we don't know who is doing what. HENNING SCHULZRINNE: I think the world would be a whole lot easier if we had a clean layer separation between the different service layers. The real world doesn't quite have that happening, for a variety of technical and non-technical reasons. The model of separation of layers in terms of companies that provide that is not the way that many countries now operate. This is true for video; think of the cable companies that offer anything from content to the physical infrastructure. It is not true for the telephone service, largely. I'm not arguing in any way that that's ideal or should be the only model, but it is a model that will probably be with us for a while; we can't just wish it away even if we want to encourage a variety of service models. BERNARD ABOBA: I would very much like to thank Dr. Schulzrinne for once again giving us an excellent plenary talk. 4. Technical Topic II: IEEE 802 Proposed OUI Registry Restructuring Bernard Aboba introduced the second technical topic for the evening, the OUI Registry Restructuring proposed by the IEEE 802. Glenn Parsons, Chair, IEEE Registration Authority Committee (RAC), delivered the following presentation: http://www.ietf.org/proceedings/86/slides/slides-86-iab-techplenary-5.pdf More information on the topic is available in the internet-draft draft-ieee-rac-oui-restructuring [https://datatracker.ietf.org/doc/draft-ieee-rac-oui-restructuring/]. The draft will be discussed further in the OPSAREA meeting at IETF 86. At the end of the presentation, the microphones were opened for questions from the audience. A summary of the Q-and-A session follows: KERRY LYNN: The data links we're used to have 48-bit MACs; I think you're trying to explain to us that we have at least 100 years left of life in the 48-bit space. Some 802 technologies that are coming out hard, like the 802.15, have 64-bit address space. Is there any sort of plan to bidirectionally bridge between these two address spaces? GLENN PARSONS: Before we get into that, one thing I didn't mention is that we're actually contemplating a 128-bit MAC address. But that aside, bridging in between a 64-bit MAC address and a 48-bit MAC address is not in the scope of the Registration Authority Committee (RAC). That would actually be in the scope of IEEE 802.1, but so far there has been no proposal to do such a thing. However, if someone came forward with proposal, we would certainly entertain the possibility. BERNARD ABOBA: If people have any more questions about this topic, you can bring them to the OPSAREA meeting. 5. RFC Editor Report Heather Flanagan delivered the RFC Editor report: http://www.ietf.org/proceedings/86/slides/slides-86-iab-techplenary-0.pptx 6. Introduce Incoming IAB Members The IAB took the stage and introduced themselves: Outgoing members - David Kessens - Danny McPherson - Jon Peterson Incoming members - Eliot Lear - Xing Li - Andrew Sullivan Continuing members - Bernard Aboba (outgoing IAB Chair) - Jari Arkko (incoming IETF Chair) - Mary Barnes (Executive Director, ex officio) - Marc Blanchet - Ross Callon - Alissa Cooper - Spencer Dawkins - Joel Halpern - Russ Housley (incoming IAB Chair) - Dave Thaler - Hannes Tschofenig Lars Eggert (as IRTF Chair) and Heather Flanagan (as RSE) joined the IAB on stage for the open microphone session. An edited transcript of that open microphone session follows: 7. Open MIC ROB AUSTEIN: This is for the RFC Series Editor: please make note that there are people in the world who speak a dialect of English other than that supported by the Chicago Manual of Style. While I understand that uniformity is a good thing, it would be nice if we didn't have to have the arguments about the use of "that" and "which." When someone writes grammatically correct English in some dialect or another, perhaps we should consider leaving it alone. HEATHER FLANAGAN: Thanks for the input. BRIAN HEDSTROM: I'm a new attendee, so maybe I'm opening a can of worms; I don't know. The work I'm dealing with here is developing very rich diagrams, information models, and in ASCII text it is really not possible to do that. When can we move to PDF, for example? [Laughter from the audience.] HEATHER FLANAGAN: There has been no decision yet about what the implications of the requirements will necessarily be. If you haven't read the requirements draft [https://datatracker.ietf.org/doc/ draft-iab-rfcformatreq/], I encourage you to do so and give me your feedback. It will impact what happens later. KERRY LYNN: If you talked about the tools I missed it. Is there a plan to continue supporting the existing tool set, specifically the validator--keeping that moving apace with style changes and so forth. It appears to be available sometimes, and other times not. Then the other thing is that recently it seems like the output of XML2RFC has changed. I wonder if there could be more advance notice and some sort of a feedback system where people can report issues that they are seeing with the tool set. RUSS HOUSLEY: We had a long debate about XML2RFC when we reached the point where 50% of the approved internet-drafts sent to RFC Editor were accompanied with XML files. We saw that the community had embraced the tool, and so we decided that it had become a critical tool, which means that we will spend money to support it. We contracted have it re-written because the original form was not maintainable, and as part the transition to the new version, we've seen a lot of announcements, including a warning that the new version is more strict about adherence to the schema. The back versions remain available. On Saturday we shifted the xml.resource.org website to use the new version, but there is a link on the web page to the the old one in case you have trouble with the new one. We really have been sending a lot of messages saying, "We're shifting these tools; please try them out and let us know if you have an problems." We have very few open bugs at this point, and that's why we made the shift. JOHN KLENSIN: To follow up on Rob Austein's question a little bit: there are number of people in the community whose version of English is different from the one Rob was asking about, and whose technical writing of English is considered in some cases a dialect of Martian. As the Martian participation in the IETF increases, in terms of technical writing of English, what do you expect to do about that? HEATHER FLANAGAN: It's a big concern that I don't necessarily know how to address. People who come to the IETF don't want remedial training in writing, and we're not really set up to do that. And yet, sometimes it's needed. Today we do a light-touch edit to make the document as clear and consistent as possible given that we think that it has come out of a consensus call in a Working Group. Doing a heavier edit for a non-native English speaker -- it's a tough area, and I'm always open to hear about ideas. One person suggested looking at the demographics by companies of where authors are, and maybe reach out to the companies and suggest that they develop some assistance on their end. I don't know if that's a good idea, but any idea is certainly welcome. MARC BLANCHET: And related to the [ongoing mailing list] discussion on diversity, we also must consider that we have non-native English speakers writing documents. JOHN KLENSIN: I agree and think those are parts of the same issue, but having said that, I want to be very certain that in this particular situation, I'm talking about the ability to write good technical English, and what we've observed in the IETF (and certainly the broader community) is that has nothing to do with whether one is a native speaker or not. We've got members of this community who are not native English speakers who write better technical English than most of us. STEWART BRYANT: I support expenditure of resources on better compilers to get from XML to the text we get as output, but I would strongly like to encourage more expenditure of resources on editors. When I write, I like to put 100% of my focus on getting clear English, clear technical descriptions, and getting it right. It really kind of annoys me, the thought of spending time trying to compile it through actually writing raw XML. In terms of WYSIWYG, the only tool that I know of is Bill Fenner's fairly old tool that is no longer free. Please, can we have some thoughts on tools that allow us to focus on the text, and not the XML syntax. RUSS HOUSLEY: I disagree; I don't think this is the community that should be building those tools. I don't want to have the VI / EMACS debate for XML editing tools. STEWART BRYANT: I didn't necessarily say that we should write it ourselves, but since we're the only users of this tool, maybe we should spend some money on someone who is an expert doing it. JARI ARKKO: Stewart, we do have the code sprint on Saturdays before the IETF, so you're welcome there. STEWART BRYANT: I'm not the greatest coder. KEITH DRAGE: In terms of getting your documents better written in English, I think that's the Working Group Chairs' problem. They've got to appoint document authors or co-editors to help with that part. Let them concentrate on the technical details and get someone else to help them with the English. HANNES TSCHOFENIG: Do you think that will help with diversity, if we push aside people who provide good content but aren't the best ones to provide the text? It's important in the community to be listed as a document author. KEITH DRAGE: But the document has to be understandable. If the person you're getting to edit it can't do that for them, then you need to get someone else. It should be happening as you're developing the document; you can't do it at the end of the process. MARY BARNES: I don't think Keith is suggesting that you take off the original author with the ideas and stuff; he's talking about a collaborative effort. I've done that with our good friends from the University of Naples, and it works out well when you do that. JOHN SCUDDER: Do you have a veto stamp where you can say, "I can't make this readable; please send it back to the Working Group"? HEATHER FLANAGAN: Yes. TONY HANSEN: When I'm writing a draft, I want to concentrate on text. However, there are formatting issues, boilerplate, cruft, and I can't share my text until it looks like all that stuff, which is one reason I work on the tools to try to make it easier. I'd love to say we don't need to have all that stuff, but we left that point 30 years ago. I wish we could share documents like, "subject, content," but I don't think we can. I had a conversation today with Henrik about something called structured text that could easily be converted into whatever fairly straightforwardly, and yes, it does take that extra step to take what you want to talk about and get it in that format where it has all the other cruft so it can be shared, but it looks very interesting. There might be tools along that line at some point. JOHN KLENSIN: Tony, the difficulty with what you said is the job the generic markup is supposed to do for you, and if it's not there is a bigger problem. With the more general problem of working group chairs saying, "You can't be an editor on your own document; you have to have somebody paired with you," it works well sometimes and badly other times. If we're worried about diversity, telling Martians they need someone else editing their documents isn't a good plan. SAM WEILER: This is not a question for Heather. So, NomCom. Acknowledging that the community doesn't know how we got into the situation of not having a new Transport AD, I want to share an opinion that I hope will help us along the way. I believe that among the tasks of the NomCom, should they choose to accept the role, is reviewing the requirements that they are given (by the IESG in this case), perhaps even second-guessing them, and deciding where to go from there. BERNARD ABOBA: Thank you, Sam. NomCom will have a session at the Wednesday Plenary, and there is also the TSV Area session where you can discuss this to your heart's content. DAVE CROCKER: Regarding NomCom deciding the requirements, I would like to lodge a request to everybody to think about what I'm going to ask for at the next plenary. There is some feeling that the discussion on the mailing list about this surfaced two interpretations of NomCom's authority with respect to job requirements. I was astonished by the certitude of the interpretation that didn't match my own, but the discussions seem to have converged on something like consensus, so can we have a hum on this [at the Administrative Plenary]? RUSS HOUSLEY: If you really think we've reached consensus on a decision, can you please write up what you think it is and put it on the list? It is easier to take a hum on a written version. JARI ARKKO: And we do understand what the model is going forward. BERNARD ABOBA: Thank you everyone; enjoy your dinner. [ENDS]