[09:50:03] --- pesherb has joined
[09:50:10] --- pesherb has left
[10:42:12] --- LOGGING STARTED
[11:51:02] --- renee has joined
[11:55:08] --- scribe has joined
[11:55:35] <scribe> Plenary session, IETF, March 21, 2007, 5 p.M.
[12:00:29] <scribe> NEW SPEAKER: We're starting in about a minute.
[12:02:59] <scribe> NEW SPEAKER: Okay. I think we are better get started.
[12:03:02] <scribe> If you could sit down, if you're planning to sit down. If
[12:03:05] <scribe> you're planning to continue talking, please leave the room.
[12:03:12] <scribe> And if any of your friends are outside, you
[12:03:14] <scribe> might waive them down, because they're going to miss the best
[12:03:15] <scribe> part.
[12:03:23] <scribe> So, I'm Brian carpenter, and I'm chair of the IETF for a
[12:03:27] <scribe> little bit longer. This is IETF 68. If you haven't
[12:03:31] <scribe> noticed, you're in Prague, with very nice bridge, Old
[12:03:37] <scribe> Town and so on. I think we're all thrilled to find ourselves
[12:03:37] <scribe> in the City.
[12:03:45] <scribe> We have a pretty crowded a agenda and we actually have two
[12:03:48] <scribe> plenaries tonight. Plenary one and two.
[12:03:54] <scribe> Plenary one is the regular one where we cover IESG and I E O C
[12:03:57] <scribe> matters, I won't read it out. We'll come do it
[12:03:58] <scribe> one step at a time.
[12:04:03] <scribe> And plenary two is when the IESG disappears from the stage and
[12:04:09] <scribe> we hand over to the internet and routing area A Ds, to will then
[12:04:14] <scribe> run a plenary on the routing addressing problem which was
[12:04:18] <scribe> discussed at the previous IETF.
[12:04:25] <scribe> And we need to be out of here at 7:30, for a number of
[12:04:26] <scribe> reasons, including hundred ger.
[12:04:30] <scribe> Hung ger.
[12:04:39] <scribe> The first item as always is going to be an inter op report. There
[12:04:45] <scribe> is something special here, a contractor
[12:04:51] <scribe> running the knock, so after a bidding process, we have various
[12:04:53] <scribe> other people who have helped to make everything work. And
[12:04:58] <scribe> I'm going to switch to another presentation in a second, with the
[12:05:02] <scribe> audience for that. In the meantime, would Morgan like to get
[12:05:04] <scribe> up here to be ready to talk in a couple of minutes. NEW SPEAKER:
[12:05:10] <scribe> So, I have to see if I can make this bit work.
[12:05:22] <scribe> Yes. This is the one. So, just to say thanks to
[12:05:27] <scribe> NeuStar, our main host, to sooes in this case, whatever you prefer, which
[12:05:33] <scribe> is the registry here, C S net, which is a network I know
[12:05:43] <scribe> from my old days in the research networking world. And last mild, which
[12:05:49] <scribe> is probably a last, here dial tell Lee com, the nok cakt
[12:05:53] <scribe> err is Varian networks. And all those people who make it happen, the volunteer
[12:05:58] <scribe> team who are listed there, and the secretary rat services
[12:06:02] <scribe> team who are are listed at the bottom. So a little round of applause
[12:06:04] <scribe> for all these people.
[12:06:04] <scribe> ( Applause.)
[12:06:16] <scribe> So let me go back and find the presentation that we need. And
[12:06:24] <scribe> Morgan, you're on. NEW SPEAKER: Thank you. This is
[12:06:34] <scribe> about the IETF 68 network, and I'd like, as Brian said,
[12:06:35] --- bhoeneis has joined
[12:06:37] <scribe> somebody is providing the network staff, along with all of the
[12:06:40] <scribe> volunteers that have helped and the contributors who have been
[12:06:44] <scribe> here, assisting us to make sure that everything is running smoothly.
[12:06:47] <scribe> The initial set up went very easily. We got everything up and
[12:06:52] --- Stephen Farrell has joined
[12:06:53] <scribe> running Sunday evening, with partial connectivity earlier in
[12:06:53] <scribe> the day.
[12:06:59] <scribe> We did find and some of you have noticed an initial routing
[12:07:04] <scribe> issue, to certain sites that were fixed on Monday. We took
[12:07:08] <scribe> a little time there as the provider that we needed to speak with to transit
[12:07:11] <scribe> that was not available on Monday morning.
[12:07:16] <scribe> Since then, we have seen very little problems and expect that
[12:07:19] <scribe> everything else is going to continue on that way. We've
[12:07:25] <scribe> deployed 69 '80 two 11 try mode Cisco radios
[12:07:29] <scribe> throughout the building. Services available, in all the
[12:07:32] <scribe> meetings, basically the lower lobby, the lobby, the
[12:07:37] <scribe> mezzanine area, there are also A Ps on the upper levels that are
[12:07:42] <scribe> providing service in the executive lounge and to some of the guest
[12:07:45] <scribe> rooms, those of you who are lucky enough to get that.
[12:07:51] <scribe> Wireless users is 749 at current count. We're
[12:07:56] <scribe> Broadcasting four S S O V. Two on '80 two 11 A. And two
[12:08:00] <scribe> on the B G. One is open, and one is secured can something else.
[12:08:13] <scribe> Here's basic coverage of mezzanine level. The lower lobby,
[12:08:19] <scribe> lobby level, charting with associations, as you
[12:08:25] <scribe> can see fairly busy. The vair Lynn stuff that's assisting here
[12:08:31] <scribe> is Steven trade Dell. Andrew clap, David bros ser,
[12:08:33] <scribe> they've been helping, and then myself.
[12:08:41] <scribe> Cisco systems has, in addition to the A Ps and the switches that
[12:08:46] <scribe> were part of donation earlier in the year have also donated the
[12:08:50] <scribe> two border route terse that we are using to provide external
[12:08:55] <scribe> connectivity and again thank you to the people for providing
[12:08:56] <scribe> the external connectivity.
[12:09:00] <scribe> Speaking of the volunteers, again, list of names of
[12:09:04] <scribe> everybody else who has been assisting in the nok, helping the
[12:09:10] <scribe> network run smoothly. We've been pushing traffic but we
[12:09:16] <scribe> still have plenty of room to go. There's a gig bit link from
[12:09:20] <scribe> ses net, so we shouldn't have think problems. And I will
[12:09:27] <scribe> introduce Michael ces from cease Nick who will talk about the
[12:09:27] <scribe> upstream configuration.
[12:09:34] <scribe> NEW SPEAKER: Hello. I'd like to explain you the
[12:09:40] <scribe> situation with upstream. From the Hilton, there are two
[12:09:50] <scribe> fibers, unfortunately, there is last 50 meter shared fiber from
[12:09:59] <scribe> Hilton to next building, to the business center. Then there
[12:10:03] <scribe> are two dual op streams. The first one is the op fiber to
[12:10:08] <scribe> test necessary, which is two '85 two. It's one giggling and
[12:10:13] <scribe> then ses net has one two up streams, one is to jam, which
[12:10:22] <scribe> is type holder, which is the ten gig lines, you may see.
[12:10:27] <scribe> And other connection is to tell Lee network, and it's it's
[12:10:31] <scribe> used for inter connecting the work.
[12:10:39] <scribe> As Morgan told you, there was on Monday morning, some
[12:10:45] <scribe> troubles like isometric routing. It has been viewed some, on
[12:10:54] <scribe> down the jam to site with, sorry, the coming together with more specific
[12:10:55] <scribe> routes to this path.
[12:11:03] <scribe> This is also another coupling going to dial tell com. IP
[12:11:10] <scribe> network, and this network is also tell yeah, which seems to
[12:11:15] <scribe> be something like single point of framework, but there are some links which
[12:11:21] <scribe> are not listed here, inter connecting ses net to the world and also
[12:11:27] <scribe> the other tell com. Like we have links to Poland, Poland
[12:11:34] <scribe> Austria and slow vacuum ka, and dial also has to slow vac
[12:11:38] <scribe> yeah to tubes which are able to handle more than one gig of
[12:11:38] <scribe> traffic.
[12:11:44] <scribe> So, this is the upstream con fig, and probably some
[12:11:49] <scribe> interesting numbers for the, what is the network using for
[12:11:50] <scribe> used for.
[12:11:57] <scribe> The first one is showing what is the long tail. It means most
[12:12:04] <scribe> traffic from, going to this network is H T P, it's
[12:12:10] <scribe> supports '80. Number three is S S H. No. 4 is you know, H
[12:12:16] <scribe> T T P S. But what's nice, but nice for some
[12:12:23] <scribe> sniffers, not nice for people, No. 10, is op three which
[12:12:30] <scribe> is not something. So maybe if you are using this protocol,
[12:12:34] <scribe> some people have your password.
[12:12:44] <scribe> So, this is one interesting thing which is provided by ses net
[12:12:49] <scribe> stats. We are not looking to pay all, just for the
[12:12:53] <scribe> headers. And it is for last day.
[12:13:00] <scribe> So, this is, there is last slide for me, this is IP
[12:13:05] <scribe> traffic last day, but during the day, the colors are
[12:13:11] <scribe> saying, so you see that there is some ambient traffic over S S
[12:13:19] <scribe> H, so it means you are working all the time. There is some web
[12:13:25] <scribe> traffic which means you are are interested in what's done in
[12:13:29] <scribe> the public country, and some ACK bent traffic, you may see,
[12:13:36] <scribe> I mean, some deep traffic, you may see the, as you are
[12:13:47] <scribe> yesterday, around noon, someone tried to down load a
[12:13:49] <scribe> package from local network.
[12:13:55] <scribe> So, that's all for my side, and I would like to recommend to
[12:14:02] <scribe> you to visit some other stats which are provided by steen,
[12:14:10] <scribe> and the address for these stats are on the local network. So that's everything
[12:14:11] <scribe> for me. Thank you.
[12:14:12] <scribe> ( Applause.)
[12:14:18] <scribe> NEW SPEAKER: So. Thank you. More began and the crew, because
[12:14:21] <scribe> without you guys, we wouldn't have a meeting connection that worked.
[12:14:30] <scribe> Right now, if I'm not mistaken, this is the next one I
[12:14:38] <scribe> need, so, John. John Lynn bering is vice-president of secretary rat
[12:14:42] <scribe> services of NeuStar and he's speaking on behalf of our host. NEW SPEAKER:
[12:14:49] <scribe> Thank you, Brian. And I understand time is of the essence.
[12:14:53] <scribe> For these meetings, so we have relatively short
[12:14:57] <scribe> presentation. I had a hundred and 84 slides talking about how
[12:15:01] <scribe> wonderful NeuStar is, and all of its services, but
[12:15:03] <scribe> apparently, we've only paid for ten minutes.
[12:15:10] <scribe> Just as a clarification point NEW SPEAKER: That's 18
[12:15:13] <scribe> slides per minute. NEW SPEAKER: We have some great
[12:15:14] <scribe> mathematicians here.
[12:15:18] <scribe> NEW SPEAKER: I'll take all the help from the audience possible.
[12:15:22] <scribe> And I speaking of the audience shs I wanted to give a special
[12:15:26] <scribe> word of welcome to those on the bad N T channel. I want you all
[12:15:29] <scribe> to know that I am here for your entertain ment.
[12:15:37] <scribe> So very happy to be here in Prague with all of you this week and
[12:15:44] <scribe> hope you've enjoyed your week as well. As we tag on to the
[12:15:54] <scribe> next slide, I am here basically to give the elevator speech for
[12:15:58] <scribe> NeuStar, very quick overview, just a few slides. And starting with
[12:16:05] <scribe> this one, you know, NeuStar is serving network operators
[12:16:08] <scribe> and community providers by offering seamless inter
[12:16:13] <scribe> operability solutions. And probably the most famous example of
[12:16:17] <scribe> that is that in North America, NeuStar clears or routes about
[12:16:21] <scribe> two-thirds of a trillion phone calls a year. And it's able
[12:16:27] <scribe> to do this because it has a very stringent policy of neutrality
[12:16:30] <scribe> with all the operators. And that element of the neutral tea
[12:16:34] <scribe> flows through and continues toes inform
[12:16:37] <scribe> all the services that NeuStar works on with the
[12:16:43] <scribe> internet. And so, the IETF and NeuStar share this very
[12:16:52] <scribe> similar core value or core goal of inter operability. As we
[12:16:56] <scribe> look at, continue to take another quick look at NeuStar, here
[12:17:01] <scribe> is a list of fact toydz, of which I'll just quickly go through a
[12:17:04] <scribe> couple of them that I think are especially interesting. One
[12:17:10] <scribe> is that we're, rearly e really a global company and you can
[12:17:15] <scribe> see that we have growing options around the world. We're
[12:17:18] <scribe> growing in size, we're pushing a thousand employees with our
[12:17:21] <scribe> growth trajectory. We have 30 global points at
[12:17:25] <scribe> present. Not to go through all the points, but
[12:17:29] <scribe> one interesting static is that we resolve over 22 percent
[12:17:36] <scribe> of the world's domain names. And we are processing about
[12:17:40] <scribe> over, well, greater than 50 xwil yon resolution transactions
[12:17:41] <scribe> per month.
[12:17:48] <scribe> So, some fact toydz about NeuStar, what's the overall
[12:17:49] --- pesherb has joined
[12:17:52] <scribe> picture in the connection from NeuStar to the IETF, and I'm
[12:17:56] <scribe> asked that question a lot. And this kind of slide which I will
[12:17:59] <scribe> quickly go through basically says, you know, it's technology
[12:18:02] <scribe> becomes more advanced and the protocols become
[12:18:06] <scribe> more sophisticated and the service offerings continue to become
[12:18:10] <scribe> more robust and the end users continue to have higher and higher speshingt
[12:18:14] <scribe> taitions, NeuStar reliance on the IETF's standards is
[12:18:18] <scribe> absolutely essential in order for us to deliver on the
[12:18:19] <scribe> commitments that we make to our customers.
[12:18:26] <scribe> This is a slide that we put together just to track back within
[12:18:31] <scribe> the history of NeuStar, and the history of the IETF. You
[12:18:36] <scribe> can see that you know, some of these foundation Al RFCs
[12:18:44] <scribe> like ten 34 and 37 six one and RFC 2821 have
[12:18:48] <scribe> give even birth to some of these fantastic technologies like E
[12:18:52] <scribe> number and SIP, and these technologies have resulted in
[12:18:56] <scribe> essentially really, in NeuStar's estimation, the IETF having
[12:19:00] --- Stephen Farrell has left
[12:19:00] <scribe> created this one can wonderful global market.
[12:19:03] <scribe> So I think that's a tremendous acknowledgement of the
[12:19:07] <scribe> contributions with the IETF has made, and NeuStar is
[12:19:07] <scribe> grateful.
[12:19:12] <scribe> NeuStar is very grateful for all the work that all of you have done.
[12:19:20] <scribe> Here's a list of services that NeuStar offers. You know,
[12:19:24] <scribe> whether it be mobile I am or local port ability or tell com
[12:19:30] <scribe> routing, managed DNS or identity Clare raition. You can
[12:19:35] <scribe> see that nearly all the services at NeuStar offers and works with
[12:19:38] <scribe> stem from the IETF community.
[12:19:45] <scribe> Here's some names that may be familiar to some of you. These are some
[12:19:50] <scribe> NeuStar folks who are active in the community. Certainly,
[12:19:53] <scribe> hoping that everybody will continue to become more and more
[12:19:58] <scribe> active, and I believe that NeuStar will continue to make more
[12:20:02] <scribe> and more staff available to participate in the IETF
[12:20:04] <scribe> community, as we start to move forward.
[12:20:11] <scribe> Some of the current engineering topics, of special interest to
[12:20:16] <scribe> us are identity, real time carrying, secure DNS, but you
[12:20:20] <scribe> can see that as our future goes forward, we see that the IETF
[12:20:23] <scribe> community is essential, and that's one of the reasons why
[12:20:29] <scribe> NeuStar wants to to be as supportive as it can for the IETF community.
[12:20:37] <scribe> In my small role within NeuStar, I am very happy to say and very
[12:20:41] <scribe> proud to report, that I am allowed to wshing with the NeuStar
[12:20:45] <scribe> secretary rat services, which as some of you may know was a
[12:20:49] <scribe> continuation of for tech, when NeuStar bought for tech,
[12:20:53] <scribe> about fifteen months ago. And we have really three
[12:20:59] <scribe> functions, reporting to Ray pel tear, and Brian carpenter and
[12:21:04] <scribe> the I A O C. Our three tasks are basically the organization of the
[12:21:09] <scribe> meetings. We have been working diligently trying to get the
[12:21:14] <scribe> cookie scenario resolved. And we know there's still some traffic out
[12:21:17] <scribe> there. We're going to keep working on that. But, there
[12:21:21] <scribe> is a lot of work that goes into the meetings, and we want
[12:21:25] <scribe> them to be successful so that you don't have to worry about the
[12:21:28] <scribe> logistics, but that you can have an efficient meeting.
[12:21:32] <scribe> We also work with tools development and provide the IETF infra
[12:21:36] <scribe> structure behind the scenes and provide standard
[12:21:39] <scribe> support. And we're all very happy to be here at your
[12:21:43] <scribe> service, and happy to participate, and we thank you very much
[12:21:49] <scribe> for your time and my boss, Mark foster wanted me to pass on a
[12:21:53] <scribe> special word of thank you for allowing NeuStar to host this
[12:21:56] <scribe> meeting. We thank you very much and wish you the best for the
[12:21:57] <scribe> continuation of your meeting.
[12:21:58] <scribe> ( Applause.)
[12:22:05] <scribe> NEW SPEAKER: Don't get can away with without something to carry
[12:22:11] <scribe> home, John. So, this is the host plaque, a new tradition
[12:22:15] <scribe> we just instituted recently. This is presented to NeuStar as
[12:22:20] <scribe> the host of the 68 this IETF in Prague, and the check
[12:22:25] <scribe> Republic. So take this home and show it to Mark foster.
[12:22:25] <scribe> ( Applause.)
[12:22:30] <scribe> NEW SPEAKER: Just one last thank you, this really is a
[12:22:33] <scribe> beautiful plaque and we'll leave it at the registration desk if
[12:22:36] <scribe> anybody would like to take a look at it. Thank you very
[12:22:36] <scribe> much.
[12:22:40] <scribe> NEW SPEAKER: Okay. So, what's the time. Oh, we're
[12:22:42] <scribe> only three minutes late, which is not bad.
[12:22:58] <scribe> So back to a agenda. The next bit is the report which by
[12:23:00] <scribe> popular request, we've tried to shrink
[12:23:03] <scribe> down to ten minutes, since some of you seemed to get bored
[12:23:06] <scribe> when it took an hour. I'm going to give it on behalf of
[12:23:10] <scribe> myself and Ray. It's being put together well in advance. It was
[12:23:13] <scribe> finished at least three hours ago. I should say, by the
[12:23:18] <scribe> way, as far as possible, I've got all
[12:23:23] --- resnick has joined
[12:23:23] <scribe> these presentation meetings, on the Wednesday plenary, if you
[12:23:28] <scribe> wanted to get ahold of them. On the IETF web site.
[12:23:30] <scribe> Who we are this time, as I always have to point out the
[12:23:30] --- DonFedyk has joined
[12:23:36] <scribe> numbers for the people here, a snapshot at the close of registration
[12:23:40] <scribe> yesterday, there are always people who register quite late in
[12:23:43] <scribe> the week. So the total you'll see in the end will be a little
[12:23:48] <scribe> bit higher. But we have 11 29 people on site as of yesterday
[12:23:52] <scribe> evening. 1200 six4 in Dallas, but that was the meeting
[12:23:55] <scribe> total, so these numbers are very comparable. Dallas
[12:23:58] <scribe> because it was klaktly a year ago. 45 countries
[12:24:02] <scribe> represented. 36 countries represented in Dallas. We
[12:24:06] <scribe> typically do get more in the country of origin when we're
[12:24:11] <scribe> outside the U.S. And the pi chart shows you
[12:24:12] <scribe> the distribution of participation.
[12:24:17] <scribe> The big chunk is the U.S, but it's about one-third of the
[12:24:20] <scribe> people here. Whereas in the U.S, it's typically half of
[12:24:26] <scribe> the people. And you can see the others there, if the
[12:24:28] --- jlcjohn has joined
[12:24:31] <scribe> colors don't show up very well, the sequence is U.S, Japan
[12:24:37] <scribe> shs Korea, Germany, fans France, Finland, China and all the
[12:24:41] <scribe> others. And so, there's a spreadsheet with all the
[12:24:45] <scribe> numbers, which I guess we could make public if people are
[12:24:45] <scribe> interested.
[12:24:53] <scribe> The IETF standards activity since San Diego, three new working
[12:24:57] <scribe> groups have been charted. Six have been closed according to
[12:25:00] <scribe> the secretary's statistics. As I say every time,
[12:25:03] <scribe> whenever I count the working groups, there are always a hundred 20 of
[12:25:09] <scribe> them. There's some strange a arithmetic going on here. I
[12:25:13] <scribe> think somebody wrote a draft about that already. And he can work on
[12:25:13] <scribe> this problem.
[12:25:19] <scribe> 44 one new I Ds. That's O O drafts of which as usual,
[12:25:25] <scribe> half came in the final four weeks. A thousand 20 I Ts got
[12:25:29] <scribe> updated, six one percent were in the last four weeks. We
[12:25:36] <scribe> are working as, as has been said quite often, on a tool to make
[12:25:39] <scribe> submission less dependent on manual intervention, so
[12:25:44] <scribe> hopefully by the time we show these statistics next time
[12:25:47] <scribe> around, these will be statistics including usage of the tool.
[12:25:52] <scribe> A hundred 19 documents wement to IETF last call. The IESG
[12:25:55] <scribe> approved a hundred 30 documents. It was a push to get documents
[12:26:00] <scribe> approved before the change over of people in the IESG. And
[12:26:03] <scribe> six7 of those documents were standards drafts or B C P.
[12:26:07] <scribe> And I do always say, these are measures of quantity and not
[12:26:15] <scribe> necessarily quality. RFC editor actions, since the last
[12:26:19] <scribe> IETF, actually, this is by calendar month, because that's
[12:26:23] <scribe> how the statistics work. And those four months, they
[12:26:29] <scribe> published '95 RFCs, you can see the breakdown. The
[12:26:32] <scribe> cumulative changed from a hundred 42
[12:26:35] <scribe> to the a hundred 4. That's the first time for several IETF that
[12:26:39] <scribe> we reported the queue getting bigger instead of spawl err.
[12:26:43] <scribe> It tend to suggest we've actually now got to something like a steady
[12:26:46] <scribe> state, where the RFC editor is keeping up with the rate at which we're
[12:26:48] <scribe> sending them stuff.
[12:26:53] <scribe> And we'll see how that goes going forward. But to me, that's, you
[12:26:58] <scribe> know, that's a perfectly okay situation. The big deal is
[12:27:05] <scribe> in 2006, they published 4 59 RFCs, and I haven't got the
[12:27:09] <scribe> graph, but if you see the graph, you will realize this is
[12:27:13] <scribe> spectacular, compared to historical levels. So I would
[12:27:17] <scribe> give the RFC editor a round of applause. ( Applause.)
[12:27:25] <scribe> The new contract is imminent. They're currently still
[12:27:28] <scribe> operating under a letter of intent. There are some fine
[12:27:32] <scribe> details of the IP R clause, we got further news on that today.
[12:27:35] <scribe> That seems to be practically resolved. They've done quite a
[12:27:38] <scribe> lot of other work, including reorganizationing
[12:27:42] <scribe> their web site, which you may well have noticed. There's a
[12:27:49] <scribe> rather full err report on the I A O C's web site, and there
[12:27:50] <scribe> will be a URL for that at the end.
[12:27:57] <scribe> IANA actions during the same period, they processed 1100 60
[12:28:00] <scribe> IETF related requests, which is significantly more than we
[12:28:06] <scribe> reported last time. The reason for that is actually the very
[12:28:09] <scribe> small bullet at the bottom. They've done a lot of work to improve
[12:28:11] <scribe> their tooltion and metrics. And they're now probably
[12:28:14] <scribe> counting things that they were doing before but not counting.
[12:28:19] <scribe> Examples of what sort of thing they've been doing, the winner is ten
[12:28:23] <scribe> requests, '79 six pen requests in four
[12:28:28] <scribe> months, and '81 port requests, which is a little bit
[12:28:32] <scribe> worrying, all the queues are below 59 items, or
[12:28:36] <scribe> at least they were a few days ago. NEW SPEAKER: It's
[12:28:40] <scribe> media type. NEW SPEAKER: Yes. MIME is media type
[12:28:45] <scribe> requests. Correct.
[12:28:52] <scribe> They reviewed 300 plus I Ds during last call IESG review.
[12:28:57] <scribe> That's very important that eye nan na reviews the IANA considerations
[12:29:01] <scribe> in advance of the I O C getting written this stone. And the
[12:29:04] <scribe> most important thing in a sense is that the we finally signed the
[12:29:08] <scribe> S L.A, service level agreement, between IANA and the
[12:29:12] <scribe> IETF receipt tiff by the I USC. And we're
[12:29:15] <scribe> rounding out the implementation of that, which is why the
[12:29:18] <scribe> statistics are getting more reliable, I think. That's very
[12:29:22] <scribe> important step forward, as the first time since the year
[12:29:24] <scribe> 2000, that we've actually mod fid the terms of our agreement with
[12:29:25] <scribe> IANA.
[12:29:33] <scribe> Budget numbers, this is a really interesting stuff. People
[12:29:35] <scribe> asked to know how the meetings are doing, so the preliminary
[12:29:39] <scribe> numbers for San Diego was income was in terms of K, the direct
[12:29:43] <scribe> meeting expense was 40038 K. That's the money we spent
[12:29:48] <scribe> on the meeting and the surplus goes into the I S budget. And
[12:29:52] --- tlyu has joined
[12:29:53] <scribe> I have to be careful how I phrase the next bit. And I want
[12:29:57] <scribe> you to hold judgment on that phrase until you see the next
[12:29:57] <scribe> slide.
[12:30:02] --- alexis has joined
[12:30:02] <scribe> We have to make a surplus on the meetings, because
[12:30:06] <scribe> otherwise, we can't construct our annual budget to cover
[12:30:07] --- weiler has joined
[12:30:07] --- admcd has joined
[12:30:13] <scribe> everything. Including the secretary rat services. So,
[12:30:18] <scribe> here's an attempt at the budget summary. Left hand column is
[12:30:23] <scribe> the unaudited actual numbers for 2006. The right hand column
[12:30:25] <scribe> is the plan for dwo thousand 7.
[12:30:30] <scribe> Paragraph 2007. And so what you'll see is our total
[12:30:34] <scribe> meeting income is around two point four million.
[12:30:35] --- JeffH has joined
[12:30:43] <scribe> Our income from I S O C was 1.2 5 million last year. Income
[12:30:46] --- Leslie has joined
[12:30:47] <scribe> is the wrong word. Contribution is the correct work. If you're
[12:30:51] <scribe> used to thinking in terms of cash flow, it looks like income.
[12:30:56] <scribe> And the contribution this year will be one point six 5
[12:30:59] <scribe> million, according to the budget. sthas direct contribution from I
[12:31:05] <scribe> S O C specific sponsor ships, obtained from I S O C added
[12:31:09] <scribe> together. Am I saying it right, Lynn? Good. So you
[12:31:14] <scribe> see the totals there. And the expense totals, since we
[12:31:19] <scribe> have a single contract with NeuStar secretary yacht services,
[12:31:23] <scribe> there's one act that covers both the secretary services and the
[12:31:26] <scribe> meeting expensees, but they're split about 50/50 if you go
[12:31:27] <scribe> into the details.
[12:31:35] <scribe> And that's going to go up somewhat next year. The RFC editor
[12:31:37] <scribe> total, that includes the copy editing, right? NEW SPEAKER:
[12:31:42] <scribe> Yes. NEW SPEAKER: The IETF gets some direct support for
[12:31:46] <scribe> example, support of IESG retreat meetings and things like that
[12:31:51] <scribe> and the IETF trust costs a little money, and then there's the
[12:31:55] <scribe> general costs of running I O S. Which are going up quite a bit
[12:31:58] <scribe> next year, but if you want to know about that, talk to Ray, because I don't know
[12:32:02] <scribe> why. That's the projected budget for 2007.
[12:32:08] <scribe> The plan for 2006 was a little bit more, so we came in under
[12:32:11] <scribe> budget. Which means we spent a little bit less than we
[12:32:14] <scribe> expected to, which means that the contribution from I S O C
[12:32:17] <scribe> was a little bit less than had been planned.
[12:32:22] <scribe> 2007 to make these numbers add up, regrettably shs we had to increase the
[12:32:30] <scribe> meeting fee to $600. And the numbers for 2007 are on track so
[12:32:34] <scribe> far. Which means the monthly numbers are what they should be
[12:32:39] <scribe> to a close prox approximation. It isn't
[12:32:43] <scribe> meaningful to do actuals until after this meetings when we get the first
[12:32:46] <scribe> cut at the numbers for this meeting, because our budget goes in
[12:32:50] <scribe> three big chunks through the year because of the three
[12:32:52] <scribe> meetings. Anything to add, Ray? NEW SPEAKER: Okay.
[12:33:00] <scribe> Actions taken by the I O C and IETF trust since San Diego, and of
[12:33:04] <scribe> course, Lucy Lynch is sitting there is the I O C chair and
[12:33:09] <scribe> this bit is on her behalf. We finalized
[12:33:13] <scribe> the new RFC editor contract except for the IP R details which
[12:33:17] <scribe> are just in the final details. We signed yesterday with
[12:33:20] <scribe> IANA. Which is the other side of those two previous points.
[12:33:26] <scribe> We issued an RF I on secretary services, because under the
[12:33:29] <scribe> understanding we have in the I O C and what the community expects
[12:33:31] <scribe> us to do, that was something we needed to do.
[12:33:33] --- pk has joined
[12:33:38] <scribe> We completed the RFC copyright transfers from I S O C to IETF
[12:33:43] --- randy has joined
[12:33:45] <scribe> trust, which means all the I S would be copyright and the I
[12:33:48] <scribe> sock's rights are now being assigned to the IETF trust as
[12:33:49] <scribe> planned.
[12:33:54] <scribe> We reached agreement to attach an open source agreement to the
[12:33:57] <scribe> secretary rat tools. We regard that as quite important that
[12:34:01] <scribe> those tools should be open source. And we received source
[12:34:04] <scribe> files recently enough that, I don't think we've finished
[12:34:07] <scribe> analyzing them yet. But they're there.
[12:34:14] <scribe> IETF is a registered trademark. And which we think is
[12:34:18] <scribe> important to you know, protect against any kind of misuse,
[12:34:21] <scribe> it's not that the trust is going to go around and suing people for
[12:34:26] <scribe> using it on their slides but with we wanted to be sure we had a
[12:34:27] --- tonyhansen has joined
[12:34:29] <scribe> reasonable measure of protection on the trademark. So you will
[12:34:32] <scribe> see a new version of the logo appearing, which has the symbol.
[12:34:35] <scribe> I patched it into this slide.
[12:34:43] <scribe> And we defined and implemented a document retention policy,
[12:34:47] <scribe> which sounds boring until you realize that marshal
[12:34:51] <scribe> spenlt I don't know how long shredding documents
[12:34:55] <scribe> that have been kept for many years. And like any good
[12:35:00] <scribe> corporate entity, we developed policy so we know what needs to be
[12:35:03] <scribe> kept. And the full reports on all these activities are on
[12:35:06] <scribe> the URL, and at the bottom of that slide.
[12:35:12] <scribe> That's all. We're going to have Q and A in a while where you
[12:35:17] <scribe> can raise any questions about any of this. That comes after
[12:35:17] <scribe> the next bit.
[12:35:28] <scribe> Future meetings, IETF 69 Chicago, hosted e by Motorola. IETF 7 on
[12:35:32] <scribe> Vancouver, the host is still to be announced. IETF 7
[12:35:37] <scribe> one, Philadelphia, the host is com cast. Pleasees me
[12:35:41] <scribe> particularly, because com cast is in the process of pulting
[12:35:46] <scribe> in one of the largest deployments in the world. So that's just my
[12:35:50] <scribe> thing. So I really hope that I will see many of you in Chicago,
[12:35:52] <scribe> Vancouver and Philadelphia.
[12:35:58] <scribe> And that is actually the end of this report. And as I say, we
[12:36:01] <scribe> can take questions when we get to the Q and A session a little bit
[12:36:02] <scribe> later.
[12:36:12] <scribe> So we've got to 1730. Logically, 1735
[12:36:17] <scribe> physically. Andrew, chair of non com is going to give his
[12:36:22] <scribe> report and then strange things will start to happen up here. NEW SPEAKER:
[12:36:51] <scribe> Okay. This is the nom com report. Give you guys an
[12:36:56] <scribe> update on what we've been doing for the last six months or so. Okay.
[12:37:00] <scribe> So, the, we're going to go ahead and start with the
[12:37:03] <scribe> congratulations and acknowledgement. The members of non com
[12:37:07] <scribe> all worked very hard. There was lots of work to be done,
[12:37:11] <scribe> material to review, meetings to go to. Discussions,
[12:37:13] <scribe> debate and everybody trying to do what's best for the IETF.
[12:37:18] <scribe> Folks on this listed slide are here, I'm going to go ahead and make you
[12:37:22] <scribe> stand up and give you a around of applause that's well
[12:37:26] <scribe> deserved. Is anyone here? Here's Fred. Come on.
[12:37:26] <scribe> ( Applause.)
[12:37:35] <scribe> There well R we go. NEW SPEAKER: We're all in favor of
[12:37:39] <scribe> transparency, but your slide is in visible. A font issue,
[12:37:42] <scribe> if you could highlight it with your mouse or something, that
[12:37:45] <scribe> might help people know whether they're on the slide and whether
[12:37:48] <scribe> they ought to stand up. They should NEW SPEAKER: They
[12:37:51] <scribe> should know whether or not they're on the slide or not.
[12:37:55] <scribe> Let's see. We'll ask Brian if he can help me out. NEW SPEAKER:
[12:37:59] <scribe> It's sort of a gray, you chose a gray font and I don't know how
[12:38:02] <scribe> to change the color in P D F. NEW SPEAKER: Highlight
[12:38:07] <scribe> it. NEW SPEAKER: That doesn't work in P D F?
[12:38:09] <scribe> Yes, it does. NEW SPEAKER: NEW SPEAKER: Not
[12:38:11] <scribe> in presentation mode. NEW SPEAKER: Take it out of
[12:38:20] <scribe> presentation mode. NEW SPEAKER: Hit select.
[12:38:22] <scribe> (Several people talking.)
[12:38:25] <scribe> NEW SPEAKER: Is that better everybody? NEW SPEAKER:
[12:38:28] <scribe> No. NEW SPEAKER: All right. Yes, it's quha I get for
[12:38:32] <scribe> using the little apple fancy default slide stuff. Right. NEW SPEAKER:
[12:38:34] <scribe> That's exactly right. NEW SPEAKER: All right.
[12:38:38] <scribe> However, if the, Brian, your mouse is not working, okay.
[12:38:39] <scribe> Here we go. All right.
[12:38:41] <scribe> He uses the track point.
[12:38:52] <scribe> IESG nom knees,
[12:38:56] <scribe> NEW SPEAKER: Which one is the -- there we go.
[12:39:07] <scribe> NEW SPEAKER: I'll use control A. That will work.
[12:39:12] <scribe> Okay. These are, if you haven't read your mailing list in
[12:39:16] <scribe> quite a while, these are the new IESG members for this year.
[12:39:20] <scribe> You should all know this by now. These are the I A B
[12:39:25] <scribe> members, and the I A O C members. NEW SPEAKER: And
[12:39:28] <scribe> you still have Barry's name spelt wrong. NEW SPEAKER:
[12:39:32] <scribe> My apologies, Barry. NEW SPEAKER: They always do. NEW SPEAKER:
[12:39:38] <scribe> Okay. Here's some quick numbers. And really, the important
[12:39:42] <scribe> take away from this, they can't read them?
[12:39:43] <scribe> NEW SPEAKER: No. NEW SPEAKER: Even with the
[12:39:47] <scribe> highlighting. I'll summarize quickly for you. We don't have
[12:39:51] <scribe> enough people volunteering for non com. NEW SPEAKER: Next
[12:39:54] <scribe> slide. NEW SPEAKER: Well, I mean, more
[12:39:58] <scribe> specifically, the number of people we've had volunteer for non com has
[12:40:02] <scribe> dropped precipitously. Now, some of the numbers going back
[12:40:05] <scribe> and historical numbers are harder to get, but assuming that we have
[12:40:08] <scribe> about the name number of eligible volunteer pools and maybe a
[12:40:12] <scribe> little bit of faulty assumption, because going back, we had
[12:40:15] <scribe> more people coming to IETF, but the sheer number of people who
[12:40:23] <scribe> volunteered, in 2 thousand 4 we had a hundred people. Now we have
[12:40:26] <scribe> six2. Prior years we had at least a hundred 40. And I
[12:40:29] <scribe> know that because people with those index numbers were selected.
[12:40:34] <scribe> So when it comes time to volunteer next year, please throw your
[12:40:37] <scribe> name in the hat. It's hard, in fact, last year, not
[12:40:40] <scribe> this year, we actually had two people repeat from the
[12:40:43] <scribe> previous year. So it's a very small set, and they should
[12:40:45] <scribe> be commendd for their dedication, but we need
[12:40:47] <scribe> to get some more people on that list.
[12:40:56] <scribe> Okay. There's a second call for nominations for security.
[12:41:01] <scribe> Because non comedy sided to move rus to the chair position and that
[12:41:04] <scribe> didn't leave a lot of candidates in security, and we issued a
[12:41:08] <scribe> new open call for that. We got initial 8 souls, and from that we
[12:41:11] <scribe> were able to select a A D candidate. For security.
[12:41:16] <scribe> The announcements were a little bit late as you may have
[12:41:22] <scribe> noticed. IESG and I S O C came up the 27 which of February,
[12:41:25] <scribe> I OB was the March 2nd. The target date is one month prior to
[12:41:30] <scribe> IETF, so February 18. fovr the tunneling suggestions,
[12:41:33] <scribe> please see last year's report. The same things apply for
[12:41:37] <scribe> this year there. And also, we noticed 37 '77 does not take
[12:41:40] <scribe> into account the extra notification week or so that you have to
[12:41:42] <scribe> do at the end of the process.
[12:41:52] <scribe> This is really the key slide which is why key is at the top.
[12:41:56] <scribe> The first point I thir I think I covered earlier, we don't have
[12:42:01] <scribe> enough volunteers for non come. Fleece please
[12:42:04] <scribe> throw your names in the hat. For this particular case, it's
[12:42:08] <scribe> basically we are you have a strong well respected incumbent,
[12:42:11] <scribe> nobody wants to quote unquote run against him.
[12:42:16] <scribe> They'll get re appointed. If they want to do it, what this
[12:42:19] <scribe> does is results in a problem for non com.
[12:42:22] <scribe> Should you want to move that person somewhere else.
[12:42:29] <scribe> Also, including Al. It's all the qualities, that's
[12:42:33] <scribe> another possibility, sure. The other, another issue is
[12:42:36] <scribe> that the interaction of 37 '77 is between the confirming body
[12:42:40] <scribe> and the non com is not entirely clear. We, there's some
[12:42:44] <scribe> information that actually back up slides on this, that you can read
[12:42:48] <scribe> on line if you want more detail of what's going on somewhat
[12:42:50] <scribe> there. But it was an issue of whether or not the non com has
[12:42:53] <scribe> to present a dpul slate or partial slate before the confirming
[12:42:57] <scribe> body will take action on that. We need to get that sort of stuff
[12:43:01] <scribe> sorted out because that contributed a little bit to some of the delay in
[12:43:02] <scribe> getting the slates out.
[12:43:10] <scribe> The other issue of course is perpetual with non can com, the
[12:43:13] <scribe> contention between openness and confidentiality. We want to
[12:43:17] <scribe> be as open as possible and not do the smoke ee back room
[12:43:19] <scribe> decision process. But at the same time we want to protect the
[12:43:23] <scribe> confidentiality of everybody involved to the extent that it
[12:43:26] <scribe> makes sense to do so. One radical suggestions would be to amend
[12:43:30] <scribe> 37 '77 to make lists of willing can daiingts or perhaps
[12:43:33] <scribe> nominees depending on how we want to construct this public.
[12:43:37] <scribe> That's an issue for the community to decide and something the non com
[12:43:41] <scribe> wanted to throw out there for your consideration. And perhaps
[12:43:41] <scribe> debate.
[12:43:46] <scribe> I believe this is my last. Yes, the rest of these are back up
[12:43:49] <scribe> slides which you can read on line. And feel free to e-mail if
[12:43:50] <scribe> you've got any questions on them.
[12:43:54] <scribe> NEW SPEAKER: I believe I'll pass the baton to Lynn; is that
[12:43:57] <scribe> correct? NEW SPEAKER: Yes. NEW SPEAKER: Just
[12:44:01] <scribe> one comment while Lynn and Daniel and whoever are approaching, which
[12:44:06] <scribe> is that Andrew gave a much longer version of that presentation to
[12:44:11] <scribe> the I S G over breakfast this morning and we had fifteen people
[12:44:14] <JeffH> it's "3777"
[12:44:15] <scribe> discussing the issues he's raced and didn't reach any
[12:44:20] <scribe> conclusion in half an hour. So I certainly encourage people to talk
[12:44:22] <scribe> about it, don't imagine that we're going to settle it in the
[12:44:26] <scribe> time available today. There is clearly a number of issues
[12:44:31] <scribe> around non com process that two non comes in succession are
[12:44:36] <scribe> raised and there's a message in there for agenda or something.
[12:44:41] <scribe> Daniel, I think you're, Daniel is the chairman of the
[12:44:43] <scribe> board of the internet society and you saw the budget numbers,
[12:44:46] <scribe> so you know why he's a really important guy. NEW SPEAKER:
[12:44:48] <scribe> I only paid for two minutes.
[12:44:56] <scribe> Actually, I'd like first to reiterate what Andrew just said.
[12:45:00] <scribe> I mean you get there and I don't like the word either, you
[12:45:04] <scribe> get the governance that you sdee serve. So if you're not
[12:45:07] <scribe> willing to serve on the non com, you get quha you deserve.
[12:45:09] <scribe> ( Applause. ) NEW SPEAKER: Thank you.
[12:45:12] <scribe> NEW SPEAKER: After having said that, I have the pleasure of
[12:45:15] <scribe> actually recognizing the outgoing people and some of them
[12:45:19] <scribe> welcoming back, but all of them who are actually managed to get
[12:45:24] <scribe> rid of a job, which is always good, and I'll take these in the opposite
[12:45:29] <scribe> order in which they've been given to me. From the I A B, we have
[12:45:33] <scribe> leaving Dave Meyer, is he here? Dave. Come up.
[12:45:34] <scribe> ( Applause.)
[12:45:42] <scribe> NEW SPEAKER: All of them will get such a nice plaque as
[12:45:45] <scribe> well, but we're nicer to them than we are to NeuStar, because
[12:45:49] <scribe> we're going to ship them to them. So this is only a template
[12:45:53] <scribe> so they don't have to luing them back and show them at the
[12:45:54] <scribe> registration desk.
[12:46:01] <scribe> We have from the IP burner somebody. I haven't seen him. He was
[12:46:02] --- miaofuyou has joined
[12:46:06] <scribe> around. He's hiding. Okay. Anyway he gets a round of applause.
[12:46:07] <scribe> ( Applause.)
[12:46:14] <scribe> NEW SPEAKER: Also the outgoing folks from I A B, IESG, we
[12:46:18] <scribe> have David somebody and the ops and management area. I saw
[12:46:20] <scribe> him, he can't hide. Where is he.
[12:46:20] <resnick> Bernard Aboba
[12:46:21] <scribe> ( Applause.)
[12:46:27] <scribe> As has already been mentioned we have had rus as the security
[12:46:30] <scribe> areas director and he will come in a second.
[12:46:39] <scribe> ted Hardie as the applications area director.
[12:46:40] <scribe> ( Applause.)
[12:46:49] <scribe> And Will somebody at the routing area director.
[12:46:50] <scribe> ( Applause.)
[12:46:50] <resnick> David Kessens
[12:46:54] <resnick> Bill Fenner
[12:46:55] <scribe> And last but not least and I'm going to miss him quite a lot, we
[12:46:58] <scribe> have Brian carpenter who managed to get rid of his job. And
[12:47:05] <scribe> I will certainly miss his very short success sinkts and to the
[12:47:09] <scribe> succinct and to the point presentations. Thank you, Brian.
[12:47:09] <scribe> ( Applause.)
[12:47:18] <scribe> So I guess NEW SPEAKER: I guess I get to do to next bit which
[12:47:21] <scribe> is let the new people show their faces. I have to remember
[12:47:25] <scribe> who they are. I A B who will be on stage tomorrow, the new
[12:47:30] <scribe> people are Barry leeb err and I advise you to figure out how
[12:47:34] <scribe> his name is spelled, otherwise he'll get annoyed at some
[12:47:38] <scribe> point and I'm momentary blanking, Danny. Danny
[12:47:40] <scribe> McPherson. On the I A B.
[12:47:43] <scribe> There he is.
[12:47:43] <scribe> ( Applause.)
[12:47:44] <resnick> Leiba
[12:47:47] <scribe> NEW SPEAKER: The usual trick, the light is in your eyes
[12:47:48] <scribe> when you're up here.
[12:47:56] <scribe> On the I A O C, of course, there is actually no change in
[12:48:00] <scribe> the non com representative, and there have been other changes
[12:48:04] <scribe> that Lucy might mention shortly.
[12:48:09] <scribe> Is that right, what I just said? Yes. That's right.
[12:48:16] <scribe> And the IESG, we've got four new area directors coming in.
[12:48:20] <scribe> And I will give these in no particular order except what my brain
[12:48:24] <scribe> recovers. Tim vok is coming into security. He'll be up
[12:48:29] <scribe> here soon. So you'll see these faces. Dave is coming into
[12:48:31] <resnick> Tim Polk
[12:48:34] <scribe> routing. Chris, I don't know where he is, but he's coming
[12:48:34] <resnick> Dave Ward
[12:48:38] <resnick> Chris Newman
[12:48:39] <scribe> in to applications. And what's the other area I've forgotten. NEW SPEAKER:
[12:48:41] <scribe> Applications. NEW SPEAKER: Yes, operations and
[12:48:48] <scribe> management. See, we always forget operations. That's Ron, who
[12:48:52] <resnick> Ron Bonica
[12:48:52] <scribe> is at the bok. So those at the back.
[12:48:54] <scribe> ( Applause. ) NEW SPEAKER: Most important, the new
[12:48:59] <scribe> guy in my chair is of course, once again, rus housely.
[12:49:02] <scribe> Come up here, rus.
[12:49:07] <scribe> So we know this thing called the passing of the dots.
[12:49:27] <scribe> NEW SPEAKER: So I am now Dotless, which means I am just
[12:49:35] <scribe> a citizen. I have just become tech support for rus.
[12:49:36] <scribe> ( Applause.)
[12:49:45] <scribe> He's just told me which of the two slide sets I'm allowed to
[12:50:09] <scribe> show you. I'm indulging myself, the first slide is mine. rus
[12:50:12] <scribe> didn't know I stuck this on the front. I want to thank a few
[12:50:15] <scribe> people and what I want to say about this list of names, I can
[12:50:20] <scribe> give you a reason why every single one of these people helped
[12:50:23] <scribe> to get my job done in the last two years. And I've certainly
[12:50:27] <scribe> forgotten people. I think the message here is the IETF works because
[12:50:30] <scribe> we all help each other and long may that continue. Okay.
[12:50:32] <scribe> Now, rus. NEW SPEAKER: thank
[12:50:33] <scribe> you, Brian.
[12:50:39] <scribe> Basically, the accepting the baton from Brian,
[12:50:42] <scribe> but it's coming after a long series of other people who have
[12:50:47] <scribe> come before us, so I'm standing the shoulder of those six
[12:50:52] <scribe> people I'm actually the 7th chair, 7 layers, whatever.
[12:50:57] <scribe> But several of them have given various different words of wisdom
[12:51:00] <scribe> over the years, and so I thought I would share a few of them.
[12:51:04] <scribe> Fred Baker says, I know I'm the chair because everybody sits
[12:51:07] <scribe> on me. And I found this picture of a chair and thought of
[12:51:07] <scribe> his quote.
[12:51:13] <scribe> Hair rolled basically said there was this view from the
[12:51:18] <scribe> community side, and a very different view, a true err view from
[12:51:19] <scribe> the chair's side.
[12:51:24] <scribe> And Brian said, many thanks to Harold for spending four years
[12:51:25] <scribe> making the shoes bigger.
[12:51:33] <scribe> So, all of those are things that affect the job that I am
[12:51:33] <scribe> inheriting.
[12:51:41] <scribe> As security area director, I had a goal of continuous
[12:51:45] <scribe> incremental improvement in the security of the internet. I hope
[12:51:48] <scribe> I've achieved that. And basically, I'm carrying that
[12:51:52] <scribe> philosophy forward, but it's continuous, incremental
[12:51:57] <scribe> improvement of all aspects of the internet. And continuous
[12:52:01] <scribe> incremental improvement IETF standards development process. So
[12:52:03] <scribe> that's what I hope to achieve.
[12:52:09] <scribe> Many people have asked me how someone that does not work for a
[12:52:12] <scribe> very big company can take on this position, and the answer is
[12:52:18] <scribe> sponsorship. I'm receiving monthlyly check from ver
[12:52:22] <scribe> sign, which I greatly appreciate, and the N S A is providing a
[12:52:23] <scribe> travel budget for me.
[12:52:27] <scribe> So, the bottom line is that -- ( Applause.)
[12:52:37] <scribe> I'll tell N S A that. They all applaudd here.
[12:52:41] <scribe> Oh, they probably already know. ( Laughter.)
[12:52:49] <scribe> So, one thing I want to say is obviously, I'm on a steep
[12:52:53] <scribe> learning curve, learning how big those shoes are that I'm stepping into.
[12:52:58] <scribe> But we all have a different view of the IETF, and it's
[12:53:01] <scribe> important that I receive input from all of those different
[12:53:07] <scribe> views, and I welcome your input. Help me figure out what and how
[12:53:11] <scribe> I'm to achieve those the goals of continuous incremental
[12:53:13] <scribe> improvement. Thank you.
[12:53:13] <scribe> ( Applause.)
[12:53:24] <scribe> NEW SPEAKER: Brian says I don't get to sit down, I run the
[12:53:29] <scribe> meeting now. Okay.
[12:53:36] <scribe> So, is Lucy supposed to come up now? Is that right? NEW SPEAKER:
[12:53:43] <scribe> Yes. Okay. NEW SPEAKER: Great. So, and after
[12:53:52] <scribe> Lucy, we're doing the, so, we're doing I O C open Mike be
[12:53:57] <scribe> then I O S G open Mike. NEW SPEAKER: I'm Lucy Lynch
[12:54:01] <scribe> and for the last time I will say I'm the I O C chair,
[12:54:05] <scribe> formerly from the University of Oregon, now working for the
[12:54:08] <scribe> internet society. First thing I will do is tell you who is new
[12:54:09] <scribe> and who is leaving.
[12:54:15] <scribe> I'm leaving, Brian is leaving, Leslie is leaving, and
[12:54:24] <scribe> Steve cok err already left. Steve cok ker has been replaced
[12:54:29] <scribe> by Ed Baker, and my seat is now being filled by Bob hin den.
[12:54:33] <scribe> Rust rus is filling Brian's seat and oh
[12:54:36] <scribe> laugh will be filling Leslie's seat. This is a huge amount
[12:54:39] <scribe> of transition in a team that's been very tight for the last three
[12:54:45] <scribe> years, basically, from the I A S T the through. And I,
[12:54:47] <scribe> and this will be the first time ever that you'll three chairs change at once,
[12:54:50] <scribe> because this is the first time there are three chairs to change.
[12:54:56] <scribe> I would call folks up, because the Q and A should be addressed
[12:55:00] <scribe> to everybody. I have one item tead up to start the
[12:55:03] <scribe> conversation here. You will have seen my e-mail saying that the
[12:55:07] <scribe> IETF trust, which is the other half of the I O C's
[12:55:12] <scribe> responsibilities has come up with a simple assignment document
[12:55:18] <scribe> that allows authors of early RFCs to assign those
[12:55:20] <scribe> rights they think they may have in the
[12:55:21] <scribe> documents to the trust.
[12:55:26] <scribe> As Brian said earlier, the I sock IP R has all been
[12:55:30] <scribe> transferred. Authors now have the opportunity to say, when
[12:55:34] <scribe> I wrote this document, I believed this was part of the
[12:55:37] <scribe> community repository and it officially signs it over.
[12:55:40] <scribe> I think we have a couple of volunteers. If they would be
[12:55:42] <scribe> willing to come to the mic or come on stage?
[12:55:57] <scribe> I believe technically, they've already signed. But
[12:56:06] <scribe> collectively, I give you the author of RFC No. 1. And the
[12:56:08] <scribe> other half.
[12:56:13] <scribe> ( Applause. ) NEW SPEAKER: You gentlemen have any
[12:56:16] <scribe> remarks, feel free. NEW SPEAKER: You go first. NEW SPEAKER:
[12:56:22] <scribe> So this is an IETF process that's being started and I'm up here in stead
[12:56:24] <scribe> of at the mic down there.
[12:56:24] <scribe> ( Applause.)
[12:56:33] <scribe> Little bit like serving a non com so you don't get nominated. NEW SPEAKER:
[12:56:36] <scribe> But the serious part of this is, when I was asked to do
[12:56:39] <scribe> this, besides immediately going, let's see, I'm being asked
[12:56:44] <scribe> to give ae away rights that I thought I didn't have, and give
[12:56:47] <scribe> them to a group I thought already had it. Sure, why not.
[12:56:52] <scribe> And that's sort of the point. And Lucy has made it a couple of
[12:56:58] <scribe> times on the list and now up here, is, most of us maybe all
[12:57:00] <scribe> of us have been participating with a certain view of being
[12:57:05] <scribe> contributing. And this just documents it. And so my
[12:57:09] <scribe> attitude, since I'm not typing away on the line, is a good one.
[12:57:14] <scribe> I think this is the thing that more of us should be doing. NEW SPEAKER:
[12:57:23] <scribe> It's with a small amount of awe that I stand here. The RFC
[12:57:30] <scribe> started 38 years ago in April of 69. As a temporary informal
[12:57:35] <scribe> set of notes just to get them out so that we could start the
[12:57:38] <scribe> conversation. I fully expected that they would disappear as
[12:57:41] <scribe> we built the network, and we haven't been able to kill them
[12:57:41] <scribe> off yet.
[12:57:46] <scribe> It's been an absolute pleasure to work with the I A O C. It's
[12:57:51] <scribe> been a vigorous, sometimes strenuous course to
[12:57:55] <scribe> bring this operation into being, and to put all the pieces in
[12:58:00] <scribe> place. Lucy has been a phenomenal leader. I've been
[12:58:04] <scribe> absolutely delighted to do a small part in being part of
[12:58:09] <scribe> that, and signing over the RFCs, is exactly in
[12:58:12] <scribe> keeping with the intentions that we had right all the way along
[12:58:17] <scribe> from creating the RFCs to, through three plus decades
[12:58:18] --- dthaler has joined
[12:58:24] <scribe> later creating the I O C and I S operations and the associated machinery. So thank you very much, and
[12:58:29] <scribe> I hope that others feel similarly inclined and are building
[12:58:31] <scribe> willing to contribute. ( Applause.)
[12:58:39] <scribe> NEW SPEAKER: I'll take both of these. The official
[12:58:41] <scribe> record. These are originals. NEW SPEAKER: Yes.
[12:58:46] <scribe> You had your copy. NEW SPEAKER: And Steve, this is
[12:58:49] <scribe> for actually, what we're doing now is we're recognizing those
[12:58:56] <scribe> folks who are leaving I A O C and ltion served on the transition team and
[12:59:00] <scribe> Steve is a representative of both of those groups and we have
[12:59:02] <scribe> this sample and we'll ship it to you. NEW SPEAKER:
[12:59:06] <scribe> Excellent. NEW SPEAKER: Because right now it says IP
[12:59:09] <scribe> workshop. And Brian would like this, but -- NEW SPEAKER:
[12:59:12] <scribe> Let's see. NEW SPEAKER: You're going to insure this? NEW SPEAKER:
[12:59:17] <scribe> Oh, absolutely. So, in gratitude and recognition for your
[12:59:20] <scribe> service on the I O C, and also as a member of the transition
[12:59:25] <scribe> team, joining others of that group are Lucy Lynch, the
[12:59:30] <scribe> chair, Leslie, Brian carpenter, the transition team,
[12:59:38] <scribe> Leslie, Brian carpenter, Lucy Lynch and Steve cok err, and
[12:59:42] <scribe> Harold. Where you are. I guarantee you it will be in the same
[12:59:45] <scribe> number of pieces and you'll see this in the next month or so. NEW SPEAKER:
[12:59:48] <scribe> Thank you very much. NEW SPEAKER: Appreciate it.
[12:59:49] <scribe> ( Applause.)
[12:59:55] <scribe> NEW SPEAKER: So it's all fun and games until somebody puts
[12:59:59] <scribe> nuts in the cookies, we are now on on to the Q and A part of the
[13:00:06] <scribe> meeting. Would the I O A C members come close nouv to come close
[13:00:09] <scribe> enough. NEW SPEAKER: Could I start by asking you a
[13:00:10] <scribe> question. NEW SPEAKER: Sure, Fred. NEW SPEAKER:
[13:00:15] <scribe> I'm Fred xwaik customer, random individual. ( Baker.
[13:00:16] <scribe> Participant in the IETF.)
[13:00:20] <scribe> I can think of about 40 RFCs that could probably be
[13:00:27] <scribe> turned over to the trust very quickly. Are each of us who
[13:00:31] <scribe> have done this intending to do this or is there
[13:00:33] <scribe> some other process. NEW SPEAKER: This is not something
[13:00:37] <scribe> that is intended. This is something that is voluntary.
[13:00:41] <scribe> And what we've done here is to sort of cherry pick the easy
[13:00:42] <scribe> cases.
[13:00:46] <scribe> I think I said in the message that I sent that these are willing
[13:00:50] <scribe> volunteers, that were lightly encumberd, they
[13:00:54] <scribe> didn't owe the IP R to a company they worked for, there
[13:00:58] <scribe> weren't multiple authors on most of these documents and they are clear that
[13:01:02] <scribe> what they are signing away is their own personal sets of rights
[13:01:05] <scribe> that they think they have. This isn't a blanket thing. I
[13:01:09] <scribe> know this will go through some iterations to figure out the
[13:01:13] <scribe> more complex cases. Folks who want to volunteer are welcome and they
[13:01:14] --- alexis has left
[13:01:17] <scribe> can talk to Ray. NEW SPEAKER: Do you want us to dash NEW SPEAKER:
[13:01:21] <scribe> I O C members, I'm serious, get your buts up here.
[13:01:25] --- JeffH has left
[13:01:35] <scribe> NEW SPEAKER: Do you want to start? NEW SPEAKER:
[13:01:39] <scribe> Sure. Elliot bear. I think you might have just answered a
[13:01:44] <scribe> question, but you know, there are a bunch of us who have the old
[13:01:44] --- JeffH has joined
[13:01:48] <scribe> RFCs, and I think it starts around 2218, and
[13:01:51] <scribe> before; is that right? NEW SPEAKER: Brian would
[13:01:55] <scribe> actually probably have the number in his head, since he usual lie
[13:01:57] <scribe> does it. It was mentioned in the e-mail. NEW SPEAKER:
[13:02:02] <scribe> 2026. Set those conditions but I don't think you can assume
[13:02:05] <scribe> that 2027 is the first one covered. Because they don't
[13:02:08] <scribe> appear in the exact chronological order in the numbering
[13:02:11] <scribe> sequence, but around about then. NEW SPEAKER: So there are
[13:02:16] <scribe> probably a bunch of us here today and will probably be at the next
[13:02:19] <scribe> IETF, and it's, if there's a process that we can most
[13:02:23] <scribe> easily, you know, do while we're here, you know, I'm happy
[13:02:26] <scribe> to, I've got a couple of them. I'm happy -- NEW SPEAKER:
[13:02:31] <scribe> Great. I think probably what we'll do is just as we've done
[13:02:35] <scribe> I E O C office hours, we'll set up something that looks like
[13:02:39] <scribe> trust office hours where e people with execute the document if
[13:02:42] <scribe> they want. We'll set it up tomorrow just before the plenary,
[13:02:46] <scribe> if people want to drop by, again, this is voluntary. NEW SPEAKER:
[13:02:51] <scribe> Question for you, Lucy. Would it be rationale to have an
[13:02:56] <scribe> individual send Ray an e-mail, which is P GP signed that says
[13:03:01] <scribe> something like to the extent that I own any IP R, it's fair. NEW SPEAKER:
[13:03:05] <scribe> We would need to talk to George about the validity of that
[13:03:09] <scribe> signature on an assignment document. So I don't know the answer to that,
[13:03:12] <scribe> but that's a good question. And I'm sure Ray will put it on
[13:03:15] <scribe> his list for things to ask George. In the meantime we can
[13:03:20] <scribe> probably provide on line a P D F of the document that they just
[13:03:23] <scribe> signed. NEW SPEAKER: EKR. NEW SPEAKER: I don't
[13:03:27] <scribe> think my question is going to be a surprise. So there is
[13:03:33] <scribe> inconsistency about snacks? San Diego, you you, there wasn't
[13:03:36] <scribe> anything on the first break on if first day and I was all warmd
[13:03:39] <scribe> up. And cleverly somebody put something out the second day
[13:03:42] <scribe> so I didn't say anything. But now we're at the point that there's
[13:03:45] <scribe> nothing at the first break. So there is this quite long,
[13:03:48] <scribe> sort of 4 to 5 hour break. I'm not saying that we have to
[13:03:51] <scribe> have that. But it seems sort of an executive decision made up
[13:03:56] <scribe> there and a lot of community input. So I really ask that
[13:04:01] <scribe> somebody get a pretty clear picture to have cookies at both
[13:04:04] <scribe> brakes and rather than making it executively. NEW SPEAKER:
[13:04:07] <scribe> EKR are you talking about the afternoon brakes. Not the
[13:04:08] <scribe> morning brakes. NEW SPEAKER: Morning break, that was
[13:04:10] <scribe> a mistake. My understanding. NEW SPEAKER: Yes. In
[13:04:15] <scribe> fact, we have gone to a process where it is just beverages
[13:04:21] <scribe> had the first half of the break, and the reason is that,
[13:04:25] <scribe> the period between breakfast and lunch is about the same amount
[13:04:29] <scribe> of time as it is between the end of lunch and the second break.
[13:04:34] <scribe> And so, we don't feel like we have an obligation or, to
[13:04:36] <scribe> provide food and snacks at that particular point.
[13:04:40] <scribe> And if you feel like you need some food or something, grab
[13:04:42] <scribe> something left over from lunch and carry it around with you. NEW SPEAKER:
[13:04:47] <scribe> That's what I'm asking you not to do is not make an executive
[13:04:50] <scribe> decision, but ask the community. NEW SPEAKER: That's
[13:04:55] <scribe> what we are, our posture is, and if the community has some thoughts
[13:05:00] <scribe> about that, I'm going to receive those and I will add it as part of
[13:05:02] <scribe> the survey after this meeting. NEW SPEAKER: I find
[13:05:06] <scribe> this very frustrating action here, because it seems like,
[13:05:10] <scribe> no transparency at the initial cost of the communities, for
[13:05:12] <scribe> judgment, and what you're presenting me with the this, and if
[13:05:16] <scribe> you want to like, you know, if you want that changed,
[13:05:20] <scribe> I'll revoke. It would be better if you ask the community. NEW SPEAKER:
[13:05:22] <scribe> We will ask the community and survey after the meeting. NEW SPEAKER:
[13:05:25] <scribe> Can that includes transparency of what it will cost. NEW SPEAKER:
[13:05:28] <scribe> Thank you. NEW SPEAKER: Actually, wait, wait.
[13:05:32] <scribe> Trick question. Ray did send a message about this earlier,
[13:05:36] <scribe> and it the cost varies from meeting to meeting. So there will
[13:05:39] <scribe> be an estimate, but it isn't a solid cost, please be aware of
[13:05:41] <scribe> that. NEW SPEAKER: I appreciate it's a rough estimate. NEW SPEAKER:
[13:05:45] <scribe> That was a twik question so I don't want to be tied to an
[13:05:48] <scribe> answer that is a moving target. NEW SPEAKER: And just so
[13:05:52] <scribe> we're clear, it's not going to be like unless you don't change (Several people
[13:05:53] <scribe> talking.)
[13:05:55] <scribe> NEW SPEAKER: Understood. NEW SPEAKER: Thank you. NEW SPEAKER:
[13:06:01] <scribe> So, I'm going to actually do, an insensitive thing here,
[13:06:04] <scribe> because our time is running short. I'm going to cut the line
[13:06:09] <scribe> after Mr. kesing and I'm going to go to John next, and I'm
[13:06:12] <scribe> sorry. But we're running -- Daniel is also
[13:06:17] <scribe> over here. So, John, Daniel, David, we need to move on.
[13:06:21] <scribe> So John. NEW SPEAKER: Two things. The first is that
[13:06:27] <scribe> I'd like the I O A C to generalize part of EKR's comment.
[13:06:31] <scribe> The expectation under which the I E S A was created that policy
[13:06:35] <scribe> decisions were going to get made by the community and the I O A
[13:06:39] <scribe> C was going to interpret them. And I believe there's a case to be made
[13:06:43] <scribe> that the balance is little bit off in terms of disclosure of
[13:06:47] <scribe> things which the community can comment. The other dash NEW SPEAKER:
[13:06:50] <scribe> John, just to interrupt you. I'll let other people take
[13:06:55] <scribe> comments here, but there is a point at which refining down to the
[13:06:59] <scribe> level of number of cookies at a break is micro management, not
[13:07:03] <scribe> policy. Go on. NEW SPEAKER: Agree completely. And it wasn't
[13:07:08] <scribe> trying to raise that issue. Just a very general observation.
[13:07:11] <scribe> brav paragraph the second question is one which I would like
[13:07:15] <scribe> to ask today but hear an answer to in Chicago because I think
[13:07:16] <scribe> it's a longer term discussion.
[13:07:19] <scribe> As we look at the portion of the registration fees, which
[13:07:20] --- randy has left
[13:07:25] <scribe> is connected to non meeting activities, how high does the I O
[13:07:30] <scribe> A C expect that money, that amount to go before we begin a
[13:07:34] <scribe> review of other ways of raising revenue for running
[13:07:38] <scribe> the IETF, which spread the burden from the attendees to the
[13:07:40] <scribe> non the a ten Dee's?
[13:07:44] <scribe> NEW SPEAKER: Good question. And this may be the point in which
[13:07:49] <scribe> I announce that as I'm stepping down as chair, we have a new
[13:07:52] <scribe> chair, and I'll pump that one to Curtis and squ him to put it
[13:07:56] <scribe> on the agenda for Chicago.
[13:08:03] <scribe> NEW SPEAKER: Thank you, John. I think Sam and then Dan kl yell. NEW SPEAKER:
[13:08:07] <scribe> Sure. So, I think this thing with the RFCs,
[13:08:10] <scribe> deciding rights is wonderful. I don't understand why you're
[13:08:11] --- apetrescu has joined
[13:08:15] <scribe> stopping at RFC 2026. Because honestly, 2026 doesn't
[13:08:18] <scribe> give you all the rights. This agreement gives you a lot more rights
[13:08:23] <scribe> than you do with RFC z 2026. It in fact gives you the
[13:08:25] <scribe> rights that you need to change the license had the future.
[13:08:31] <scribe> Why the heck don't you ask all others to sign something like this
[13:08:36] <scribe> until the IP R actually gets its act together and actually
[13:08:39] <scribe> coughs it up. NEW SPEAKER: Anybody jump up here. NEW SPEAKER:
[13:08:43] <scribe> Brian, as a retiring member of this committee, I actually
[13:08:46] <scribe> agree with them. I think there is no reason not to, but it
[13:08:49] <scribe> seemed more urgent to focus on the ones where the rights are
[13:08:53] <scribe> completely unclear, but I actually agree. I think that's a
[13:08:59] <scribe> very valid comment. NEW SPEAKER: Next. NEW SPEAKER:
[13:09:02] <scribe> Daniel. NEW SPEAKER: I'd like to go last with your
[13:09:05] <scribe> permission. NEW SPEAKER: Okay. NEW SPEAKER: I
[13:09:10] <scribe> think that the issue to do with negotiations with hotels and number of
[13:09:16] <scribe> cookies, et cetera, there's a large problem therein that
[13:09:20] <scribe> what you might think you're buying is the cookie, but
[13:09:25] <scribe> what you're actually buying is the meeting room. And there
[13:09:31] <scribe> are a large number of linkd pieces here that, when
[13:09:34] <scribe> you trying to work out whether you're -- you just don't get
[13:09:38] <scribe> the charge from the hotel for what the cookie costs. You get
[13:09:43] <scribe> a charge for a free room and then all the incident tals, it's
[13:09:47] <scribe> like the stupid meeting planners which are at the back, which we're
[13:09:52] <scribe> getting billed at probably ten dollars apiece that probably cost
[13:09:56] <scribe> 30 cents to make. I really don't think we want to get into the
[13:10:00] <scribe> point of what the contract with the hotel is.
[13:10:07] <scribe> Maybe the statement is, should there be cookies at all? Are
[13:10:10] <scribe> cookies good and leave it to that. I'm on a diet.
[13:10:16] <scribe> I've actually lost 25 pounds since
[13:10:19] <scribe> if last IETF. ( Applause. ) NEW SPEAKER: So the
[13:10:28] <scribe> cookie is a big one. NEW SPEAKER: Dave, random
[13:10:31] <scribe> member of the community. I would actually like to ask you to be
[13:10:36] <scribe> a bit more straight with the mic control during this kind of question
[13:10:41] <scribe> sessions, quite frankly, I'm not interested in micro management. I'm not
[13:10:45] <scribe> interested in discussions about cookies during a plenary session.
[13:10:50] <scribe> This time is way too expensive to spend time on this, and
[13:10:54] <scribe> next time I wish we really appreciate we don't hear anything about
[13:10:54] <scribe> cookies.
[13:10:59] <scribe> Thank you.
[13:11:02] <scribe> NEW SPEAKER: Daniel, I believe you're last. NEW SPEAKER:
[13:11:07] <scribe> Okay. As we have heard, there's actually a big turn over
[13:11:13] <scribe> in this I O C committee and we heard actually quite a tight group
[13:11:16] <scribe> running a good job over three years and actually probably
[13:11:19] <scribe> beforehand and what everybody in this room should
[13:11:23] <scribe> realize is that they just created something from nothing,
[13:11:27] <scribe> something that was basically run as a commercial operation and
[13:11:31] <scribe> brought it under community control and they did it in a very
[13:11:33] <scribe> professional manner, and I was on the, mostly on the side that
[13:11:38] <scribe> was asked for the money. But I had, at some view of
[13:11:41] <scribe> this, and I think we should really, I mean, we're
[13:11:44] <scribe> thanking a lot of people, but we should really thank those
[13:11:47] <scribe> people who have done a very well done professional job.
[13:11:48] <scribe> ( Applause.)
[13:11:56] <scribe> NEW SPEAKER: I'll turn the mic back over to rus, oh,
[13:12:02] <scribe> Leslie. NEW SPEAKER: I just want to say that having
[13:12:05] <scribe> been involved in that process for the several years that it
[13:12:08] --- csp has joined
[13:12:09] <scribe> took, I'm sitting back in my last I O C meeting the other
[13:12:12] <scribe> day, where we were discussing something about the disposition of the
[13:12:16] <scribe> source code that had come from secretary rat, and it was a real
[13:12:21] <scribe> pleasure to sit back and realize that we're having this
[13:12:21] --- mrex has joined
[13:12:25] <scribe> conversation that had been inconceivable just a few years
[13:12:31] <scribe> ago, and it's great that we've arrived here. So, con grad laitions to
[13:12:32] <scribe> us all. ( Applause.)
[13:12:38] <scribe> NEW SPEAKER: thajs to everybody and over to rus. NEW SPEAKER:
[13:12:44] <scribe> Okay. At this point, the I O C leaves the stage. The I
[13:12:49] <scribe> S G comes on up. I'd like the old IESG members to pass their
[13:12:55] <scribe> dots to the new IESG members at this time. And come on up.
[13:12:57] --- csp has left
[13:13:08] --- csp has joined
[13:13:13] <scribe> If you have questions for the I S G, find the mic phone.
[13:13:23] <apetrescu> it's IESG
[13:13:46] <scribe> NEW SPEAKER: Okay. Just real quickly, because some of these
[13:13:49] <scribe> faces might be new to some of the community. I'd like to
[13:13:52] <scribe> stoort down there request John, just quickly introduce
[13:13:54] <scribe> yourselves, starting with John. NEW SPEAKER: John
[13:13:59] <scribe> Peterson from the R a I area. NEW SPEAKER: Cullen Jennings
[13:14:02] <scribe> also from rye. NEW SPEAKER: Chris new man, aps. NEW SPEAKER:
[13:14:04] <scribe> Tim Polk, security. NEW SPEAKER:
[13:14:11] <scribe> Ross, routing. NEW SPEAKER: Sam, security. NEW SPEAKER:
[13:14:15] <scribe> ops and management. NEW SPEAKER:
[13:14:19] <scribe> NEW SPEAKER: Mark townsly, internet area. NEW SPEAKER:
[13:14:23] <scribe> David routing. NEW SPEAKER: Management. NEW SPEAKER:
[13:14:30] <scribe> My name is invest, transport. NEW SPEAKER: Okay. And now for
[13:14:34] <scribe> the open Mike, expense sore Spencer, you're
[13:14:39] <scribe> first. NEW SPEAKER: Spencer, dawk kins, I wanted to
[13:14:45] <scribe> that thank, for the non com report that he did, and wanted
[13:14:51] <scribe> to mention that the strong incumbent syndrome and things like
[13:14:56] <scribe> that that were described in the non com draft that was done last year, I
[13:15:00] <scribe> wanted to encourage you guys to be looking at that.
[13:15:06] <scribe> It has, it also has impacts to time line, that
[13:15:09] <scribe> need to be considered as you guys consider other things. NEW SPEAKER: Thank you.
[13:15:15] <scribe> We've already put that on the agenda for the I S G retreat. NEW SPEAKER:
[13:15:16] <scribe> Cool.
[13:15:22] <scribe> NEW SPEAKER: Philip Mathews. I'm not sure if this is
[13:15:25] <scribe> quite the right way to raise this, but I'll do it briefly and you
[13:15:28] <scribe> can decide how to handle it. I wanted to raise the question
[13:15:30] <scribe> about the amount of time we have to read the documents for this
[13:15:34] <scribe> meeting. I know that I dpelt quite overwhelmed trying to read
[13:15:39] <scribe> all the documents for this meeting. In fact Cullen sent an e-mail
[13:15:42] --- AWGY has joined
[13:15:42] <scribe> saying, I believe we had half a minute per page, if we read 24
[13:15:46] <scribe> hours a sdai for two weeks roughly, to read just the
[13:15:50] --- Stephen Farrell has joined
[13:15:50] <scribe> documents on the agenda for the R A I area, not talking about
[13:15:55] <scribe> all the other documents or the other areas. And I certainly
[13:15:56] <scribe> couldn't keep up with that rate.
[13:16:00] <scribe> And I talked with a few other people that I know that sort of
[13:16:05] <scribe> also agree that we sort of go into right only mode. Where we
[13:16:08] <scribe> write these dock ments but nobody has a chance to read most of them
[13:16:11] <scribe> when we come to the meetings. I don't know quite how to
[13:16:14] <scribe> handle it. My suggestions maybe is to increase the amount of
[13:16:17] <scribe> time. I don't know how you want to handle this. Do you
[13:16:20] <scribe> want to talk about it now. But I'll just raise this issue. NEW SPEAKER:
[13:16:23] <scribe> Obviously, there's a balance between shutting the
[13:16:27] <scribe> queues down so that the documents become stable, and people
[13:16:31] <scribe> actually doing work that they never, that gets
[13:16:34] <scribe> procrastinated, and so, do you want the most recent work or do
[13:16:38] <scribe> you want what -- what we found when we increased that time is
[13:16:41] <scribe> people post it on their own web site and say, sorry, I
[13:16:44] <scribe> finished this after the queue is closed and that doesn't really
[13:16:47] <scribe> help, because now you have even have more places to look for the
[13:16:50] <scribe> latest documents. Does anybody else have anything you want to say. NEW SPEAKER:
[13:16:53] <scribe> You can always join the I S G and then you get to read
[13:16:54] <scribe> everything.
[13:17:00] <scribe> NEW SPEAKER: I suggest peer pressure to write shorter
[13:17:02] <scribe> documents that are more concise. NEW SPEAKER: Wasn't
[13:17:07] <scribe> it mark Twain who said it takes longer to write shorter
[13:17:07] <scribe> documents. NEW SPEAKER: Yes. NEW SPEAKER:
[13:17:10] <scribe> Anything else?
[13:17:13] <scribe> NEW SPEAKER: I feel your pain. NEW SPEAKER: Okay.
[13:17:21] <scribe> Hi, Pete. NEW SPEAKER: Pete rest Nick. I might have
[13:17:25] <scribe> missed something. Did Spencer ask and did you answer that
[13:17:28] <scribe> IESG is going to look at the non com issues? NEW SPEAKER:
[13:17:31] <scribe> We're going to discuss the documents and what we think ought to
[13:17:34] <scribe> be done about it. I certainly didn't say we were changing.
[13:17:39] <scribe> But we're hoping we can come forward with some recommendations.
[13:17:40] <scribe> To the community.
[13:17:45] <scribe> I'm sure there's another, there's lots of people who are thinking
[13:17:50] <scribe> about these things.
[13:17:56] <scribe> NEW SPEAKER: Oh, come on. With well, we must just be back
[13:17:59] <scribe> on schedule. NEW SPEAKER: Ahead of schedule,
[13:18:00] <scribe> actually. NEW SPEAKER: Move on. NEW SPEAKER:
[13:18:02] <scribe> Bring it on. NEW SPEAKER: Take a chance. NEW SPEAKER: We
[13:18:05] <scribe> have to stay for another 12 minutes. NEW SPEAKER: At
[13:18:08] <scribe> this point, we're going to give a little bit of additional
[13:18:11] <scribe> time to part two. But, during the scene change, I have
[13:18:16] <scribe> arranged for a little bit of entertainment for you. While the,
[13:18:22] <scribe> while the scene is being changed from the plenary to the BOF,
[13:18:26] <scribe> the BOF will be run by the routing and internet A Ds, so this rest of us are
[13:18:28] --- csp has left
[13:18:31] <scribe> leaving the stage. But Burt is going to come up here and give
[13:18:34] <scribe> us a little bit of entertainment while that is happening and
[13:18:36] --- Jack3010 has joined
[13:18:37] <scribe> anyone who is not interested in part two, do not be embarrassed
[13:18:38] <scribe> if you want to exit.
[13:19:06] <scribe> Do you want to sij with me. This is about fair retails much we have
[13:19:10] <scribe> many fair retails you can dream about, and at some point in
[13:19:14] <scribe> time, you will go the meeting places and we will have to do
[13:19:19] <scribe> this all day long. Fair retails. There is fair re,
[13:19:23] <scribe> tails that documents the IETF that will be available one year in
[13:19:26] <scribe> advance so we have all the time to read them and all that. So
[13:19:30] <scribe> those things will continue to exist on fair ee tails and that's
[13:19:31] <scribe> what my little things is about.
[13:19:34] <scribe> Here we go.
[13:19:34] <scribe> ( Singing.)
[13:20:59] <scribe> NEW SPEAKER: We have an extra routing guy. Anyway, so the next
[13:21:03] <scribe> hour or so, we'll talk about routing and addressing. And we
[13:21:08] <scribe> asked the I S Gs to give us this hour because of the I A B
[13:21:10] --- guido.franceschini has joined
[13:21:12] <scribe> workshop last year. Discussion on the San Diego meeting,
[13:21:17] <scribe> lots of e-mails on the mailing lists, various related
[13:21:21] <scribe> meetings in this place, and we thought that the community would
[13:21:26] <scribe> actually like to discuss this topic, and what we're going to
[13:21:31] <scribe> do is give a short report on where we believe we are, mark,
[13:21:33] <scribe> come up on stage. NEW SPEAKER: Sorry.
[13:21:35] <scribe> NEW SPEAKER: So we're going to have a report on where we
[13:21:40] --- tlyu has left: Disconnected
[13:21:41] <scribe> are, a little bit of information on what's happening, in the
[13:21:45] --- pk has left
[13:21:45] <scribe> IETF, and in the I R T F. We already had a meeting on
[13:21:50] <scribe> Saturday, a very successful meeting of the routing group and
[13:21:54] <scribe> there's going to be more meetings tomorrow. And after the
[13:22:00] <scribe> report, there will be a open Mike session, and but before
[13:22:04] <scribe> that, let's get to the report. And for the report, we
[13:22:10] <scribe> actually asked carpenter to give the report. So Brian. You can
[13:22:19] <scribe> escape for too long from here. NEW SPEAKER:
[13:22:28] <scribe> NEW SPEAKER: So, no dots. Right. I just, I'm just
[13:22:30] <scribe> this guy who has been interested had the future of the internet
[13:22:33] --- gih900 has joined
[13:22:35] <scribe> for a number of years, and I found myself dragged into a
[13:22:38] <scribe> conspiracy to prepare a presentation for this session.
[13:22:43] <scribe> And you know, some of the con spear tors are up here, but I
[13:22:46] <scribe> better not say who the others are. Maybe I should.
[13:22:52] <scribe> You know, the word we occurs in the presentation. I guess
[13:22:55] <scribe> first would be working on putting together. So Thomas who is
[13:23:04] <scribe> actually in the conspiracy, Peter, who else was in?
[13:23:06] <scribe> And there may be others. NEW SPEAKER: And others. NEW SPEAKER:
[13:23:11] <scribe> Yes. I know a lot of people. Including I A D members. NEW SPEAKER:
[13:23:15] <scribe> Yes, a lot of people have seen these slides and have of them have
[13:23:21] <scribe> chosen to reply. But there's an important cover slide which you'll see
[13:23:25] <scribe> when I find the presentation. Which is not set up.
[13:23:40] <scribe> Okay. So this is supposed to be a status report, and you know,
[13:23:44] --- csp has joined
[13:23:45] <scribe> if you disagree with it, that's okay. There's a wide range of
[13:23:49] <scribe> views on the questions we're going to discuss. Err there are
[13:23:53] <scribe> people who looked at these slides who disagree with them, very
[13:23:57] <scribe> respected members of the community. So there's no claim this
[13:24:00] <scribe> is a authoritative. We're going to talk about the recent
[13:24:03] <scribe> activity in a certain context. Talk about this week's
[13:24:09] <scribe> activity, about the analysis that's being done since the
[13:24:13] <scribe> Amsterdam I A B workshop, and this is not in any shape or
[13:24:18] <scribe> form an I B, and views that's being presented here. And a bit
[13:24:22] <scribe> of a summary. And then after that, we can go to the open Mike.
[13:24:29] <scribe> And thanks to rus's superb chairman ship, we have extra time.
[13:24:37] <scribe> So, the context of the recent activity, you know, I, I'm
[13:24:41] <scribe> sort of interested in the history of computing. I always have
[13:24:44] <scribe> been. And you have to realize, we're working in a very new
[13:24:48] <scribe> technology. It was only invented starting in 19 six2
[13:24:54] <scribe> when packet switching was invented. As generations of
[13:24:57] <scribe> technology go, that's not long ago. The concept of the
[13:25:01] <scribe> internet was invented around 1974. IP was
[13:25:06] <scribe> designed late '70s. E G F was gee signed late '80s. Side
[13:25:09] <scribe> der was designed I think 1992, fairly accurate date.
[13:25:18] <scribe> 1996 was designed in '95 or 94 to '95. And on the
[13:25:20] <scribe> history of technology revolutions, this
[13:25:21] <scribe> is very recent.
[13:25:29] <scribe> So, during the ten or 11 years leading up to the am sister damn
[13:25:35] <scribe> workshop, there's been growing concern about scaling growing
[13:25:38] <scribe> concern, traffic engineering and of course recently the
[13:25:41] <scribe> impact of IP P v6 on all this if any.
[13:25:49] <scribe> And so, the I B convened its workshop in 2006 in
[13:25:49] <scribe> Amsterdam.
[13:25:53] <scribe> Second piece of context, and I will probably say this wrong,
[13:25:56] <scribe> but there are people in the room who wrote some of these words who
[13:26:00] <scribe> can come up to the mic phone and say it properly. But
[13:26:06] <scribe> there does seem to be an architectural principle here, which
[13:26:10] <scribe> is that a network, an individual network in the internet, should
[13:26:15] <scribe> be ail to implement reasonable internet working choices without
[13:26:19] <scribe> unduly impacting another network's operations. In other
[13:26:22] <scribe> words, we don't want the situation where something happening
[13:26:29] <scribe> in network A or I S P A or whatever you want to call it, will
[13:26:33] <scribe> cause unacceptable impact on the way some other network on the other
[13:26:36] <scribe> side of the globe is working.
[13:26:43] <scribe> By the way, assist you with getting transit to the research
[13:26:45] <scribe> and education networks that we have here, you know, on the
[13:26:51] <scribe> first day, is pretty precise limitation of not respecting that
[13:26:51] <scribe> principle.
[13:26:58] <scribe> As an architectural level, there seem to be some of the things
[13:27:04] <scribe> we need to do in today's internet working that can only
[13:27:07] <scribe> implemented in ways that threaten this principle. Like
[13:27:09] <scribe> transit policy and three GP P
[13:27:16] <scribe> And it is asserted that that's a root cause of many I S P
[13:27:22] <scribe> problems and dissatisfaction by end users sites. Such as
[13:27:25] <scribe> dissatisfaction here when you couldn't reach certain machines on
[13:27:28] <scribe> Sunday. The question is what can be done to harm Monday niz
[13:27:31] <scribe> the network to the architectural principle. And for those of
[13:27:35] <scribe> you who know the story, the tragedy of the
[13:27:40] <scribe> comments where if too many cows eat the same patch of
[13:27:44] <scribe> grass, there's no grass left and all the cows die,
[13:27:47] <scribe> unless you have some sort of grass management program, in
[13:27:53] <scribe> effect. And the similar phenomena that we see in the
[13:27:56] <scribe> internet and which mainly is centerd around the
[13:27:59] <scribe> usage and provider independent address space, and if we do
[13:28:03] <scribe> nothing, we end up with no provider independent address space left.
[13:28:13] <scribe> Third point of context is technical goals, that we need to solve
[13:28:18] <scribe> user problems and we need to solve service provider problems.
[13:28:21] <scribe> End users want to connect to multiple I S Ps while
[13:28:25] <scribe> maintaining support for the kind of traffic engineering
[13:28:29] <scribe> capabilities that they want. That can be things like
[13:28:32] <scribe> redundancy, and more subtle traffic engineering requirements.
[13:28:37] <scribe> They want to be able to change I S Ps without excessive
[13:28:42] <scribe> switching costs, for what the xhiss call
[13:28:45] <scribe> switching economists call switching from one supplier to another.
[13:28:49] <scribe> And the application, we want to support end to end
[13:28:53] <scribe> transparency, so that sessions work end to end without
[13:28:54] <scribe> unexpected problems.
[13:29:00] <scribe> The I S P space, I S Ps are motivated to keep the table size
[13:29:04] <scribe> and dynamics within operational capability of the route terse
[13:29:10] <scribe> that they actually have. And we want to provide I S Ps with the
[13:29:13] <scribe> ability to engineer traffic flows to match their business
[13:29:14] <scribe> needs.
[13:29:21] <scribe> And I guess we could have gone into problems for the protocol
[13:29:24] <scribe> designs inventors, but in a sense, the
[13:29:28] <scribe> protocol designers and vendors are subsidiary to the end users
[13:29:32] <scribe> and our technology. So we should be worrying about problems to the
[13:29:35] <scribe> end points and end users of our technology, end
[13:29:37] <scribe> users including service providers.
[13:29:44] <scribe> Then the scaling, it's well known there are about 200,000 V GP
[13:29:49] <scribe> routes in the default free zone. What is interesting is that
[13:29:53] <scribe> most interesting single factor I learned in the Amsterdam
[13:29:57] <scribe> workshop stully is that many I S Ps have several times more
[13:30:02] <scribe> routes internally, due to, you know, customer routing and
[13:30:03] <scribe> VP N routing.
[13:30:08] <scribe> So that gets you to a million. Quite a big number.
[13:30:13] <scribe> The next bullet is an interesting one and had quite a bit of discussion
[13:30:18] <scribe> about whether these numbers mean anything. One possible goal is
[13:30:23] <scribe> to say we need to be able to support ten billion end nodes by some
[13:30:28] <scribe> date. And we need to be able to support you know, ten
[13:30:33] <scribe> million multiple kors by some date. I can give you
[13:30:35] <scribe> justifications from those numbers.
[13:30:38] <scribe> And they indicate that we have to go to the a different scale
[13:30:41] <scribe> from what we have today. Does anybody know how many nodes we
[13:30:44] <scribe> have today? NEW SPEAKER: Less than a billion anyway. NEW SPEAKER:
[13:30:49] <scribe> And we certainly don't have ten million multi home customers.
[13:30:55] <scribe> And the question is, can we get from here reasonable costs
[13:30:58] <scribe> from the vendors and user sites. We have to talk about
[13:31:01] <scribe> prices in the IETF, with the again, sense of trust for
[13:31:04] <scribe> anything else to do so. But the fact is, the way we do our
[13:31:07] <scribe> engineering affects the cost structures of the
[13:31:12] <scribe> vendors, the I S Ps and users. And come being back a little
[13:31:16] <scribe> closer to reality, what should be our goals for 5 years from now, in the
[13:31:17] <scribe> scaling space?
[13:31:24] <scribe> So, moving onto recent activity in that context, the I Bs
[13:31:28] <scribe> and the workshop at ten Dees are being refined in the
[13:31:31] <scribe> workshop report. I'm not up to date with the mail on that
[13:31:35] <scribe> topic, but there's been a lot of it in the last three days.
[13:31:39] <scribe> People have been analyzing the concerns it raised. People have been
[13:31:44] <scribe> looking at solutions. There have been a bunch of meetings,
[13:31:46] <scribe> there was routing and addressing workshops.
[13:31:54] <scribe> In February. There was a session what I something to
[13:31:59] <scribe> February. The European education research networking association
[13:32:03] <scribe> held a workshop on, you know, future of the internet on February
[13:32:07] <scribe> 22nd, and this was certainly a topic. And the N S F and the O
[13:32:12] <scribe> A CD held a joint workshop on January 31st, which was also
[13:32:15] <scribe> future of the internet and this was certainly a topic. And
[13:32:17] <scribe> those are certainly not all the meetings that have taken place.
[13:32:19] --- tlyu has joined
[13:32:22] <scribe> So there's been a lot of activity, whether we should blame,
[13:32:24] <scribe> congratulate or whatever, the I A B for that. But there
[13:32:29] <scribe> has been a lot of activity. I think we should con grade lat
[13:32:31] <scribe> them for stimulating the debate.
[13:32:36] <scribe> In our corner of the Woods, we established, we, when I
[13:32:43] <scribe> have a Dot, Leslie when she, lightly different color Dot,
[13:32:49] <scribe> establish did directorate routing address and directorate. It's
[13:32:50] <scribe> been announced.
[13:32:55] <scribe> The clear intention of that was established to expand after the
[13:33:02] <scribe> IETF, and come to the mic phone, Leslie. Fortunately.
[13:33:06] <scribe> To expand that after this meeting and review it in six months,
[13:33:11] <scribe> but hey, that's not my problem or Leslie's problem. And
[13:33:16] <scribe> it's all this communication and statement in the charter that
[13:33:24] <scribe> it is going to work oo refining the problem statement. As my
[13:33:29] <scribe> dots haven't completely faded away, I can tell you that they
[13:33:32] <scribe> have had their first meeting, in this hotel, I presume,
[13:33:38] <scribe> and things are starting in the directory and we can ask meem who
[13:33:41] <scribe> are in the directorate afterwards, if we want to, what they
[13:33:41] <scribe> are doing.
[13:33:45] <scribe> The routing research group is re charterd and
[13:33:50] <scribe> there's a discussion list which is very active around the I A B dot
[13:33:55] <scribe> org and the A Ds have been preparing for this the IETF.
[13:34:03] <scribe> This week's activity, this report, it's an area tomorrow, is that the only
[13:34:06] <scribe> topic on the agenda tomorrow, guys? NEW SPEAKER: Yes. NEW SPEAKER:
[13:34:12] <scribe> So the only topic on the agenda st the routing and addressing
[13:34:12] --- yangwoo has joined
[13:34:15] <scribe> problem. NEW SPEAKER: Do we focus on IETF work, on
[13:34:18] <scribe> identify locater separation and multi level locator
[13:34:19] <scribe> approaches.
[13:34:25] <scribe> rus referred to this evening as a BOF, we thought about
[13:34:32] <scribe> officially make it a BOF but trying to run a bovr for this many people in the
[13:34:36] <scribe> room is a challenge. So it's really intended I think that, you
[13:34:40] <scribe> know, the more focus BOF style discussion will take place in
[13:34:44] <scribe> the internet area, but obviously, that doesn't mean we
[13:34:47] <scribe> can't discuss anything we want to tonight. But you know, in
[13:34:52] <scribe> terms of summarizing in the end and saying that's where the working group
[13:34:56] <scribe> is, that's not where we are going to be tonight.
[13:35:02] <scribe> The routing group tomorrow is going to focus on extensions and
[13:35:04] <scribe> practices that might help in the problem space.
[13:35:08] <scribe> The routing research group has just been re
[13:35:12] <scribe> charterd as I said, it met on Saturday, and I wasn't there
[13:35:16] <scribe> but apparently it was a very lively and active meeting. So that's good news.
[13:35:21] --- TJ has joined
[13:35:21] <scribe> And so moving to the analysis that has been made, and I would
[13:35:25] <scribe> say, it's hard to say exactly who made this analysis, because
[13:35:27] <scribe> there have been a lot of people contributing and lively
[13:35:31] <scribe> discussion on the mailing list, discussion within the I S G
[13:35:35] <scribe> and I E P sub sets discussions have occurred, threads have
[13:35:38] <scribe> forkd off from other threads. But a bunch of people have been
[13:35:42] <scribe> working on this. The observation is the V G four has
[13:35:47] <scribe> been used for 13 years with little fundamental change. There's
[13:35:50] <scribe> mounting concern about the growth on the table. And the size
[13:35:54] <scribe> of the table, the rate of updates and the impacts of multi
[13:35:57] <scribe> homing on the table. And one observation, I think
[13:36:01] <scribe> everybody who has thought about this deeply has made, is that this is
[13:36:06] <scribe> largely orthagonal to IP V four, versus IP v6, because the
[13:36:13] <scribe> problems here are equivalent. Many IP v6 prefixes,
[13:36:15] <scribe> we would have exactly the same problems.
[13:36:20] <scribe> The workshop concluded there is a problem and more work is
[13:36:24] <scribe> needed to refine it. The workshop report has been evolving.
[13:36:29] <scribe> I said, just now, I'm not up to date on the thread so I don't
[13:36:33] <scribe> know what the greatest sent to the printers or not yet.
[13:36:37] <scribe> Lots of confusing discussions, because this is a very complex
[13:36:39] <scribe> topics knowledge different people have different
[13:36:44] <scribe> views. We need to clarify what we know, what we don't know
[13:36:47] <scribe> and how to proceed systematically. And this is difficult
[13:36:49] <scribe> because of the complexity of the topic.
[13:36:54] <scribe> So, I in vent ted the phrase box, when everybody was getting
[13:37:01] <scribe> very excited about, the impact of this growth that we're
[13:37:06] <scribe> seeing on electronics in the route terse.
[13:37:10] <scribe> Routers, the workshop certainly looks at the hardware
[13:37:14] <scribe> trends and they raise some concerns about how long you can go
[13:37:22] <scribe> scaling the fib, which the information database inside a
[13:37:27] <scribe> router, without getting into serious engineering problems.
[13:37:34] <scribe> And, you know, there's a bullet that says, unless the fib
[13:37:39] <scribe> scales sub linearly with the number of web sites, it brakes
[13:37:43] <scribe> the way we're doing addressing, because it would mean the
[13:37:46] <scribe> tables would expand without limit, basically, or at least
[13:37:51] <scribe> about ten million will become the size of the table. I'm
[13:37:55] <scribe> looking at Ross to see if I said something stupid and he doesn't seem
[13:37:58] <scribe> to think -- press the button if you want to speak. NEW SPEAKER:
[13:38:02] <scribe> No, I'm just thinking about what you're saying. I'm not
[13:38:05] <scribe> that scared by ten million myself. But, NEW SPEAKER: There isn't
[13:38:07] --- hartmans@jis.mit.edu/owl has joined
[13:38:09] <scribe> really agreement on the community on this particular question, there's been a
[13:38:13] <scribe> lot of debate. In the opinion the people that put the slide set
[13:38:17] <scribe> together, the recent analysis, I mean new architectures
[13:38:20] <scribe> that people have used to build routers,
[13:38:23] <scribe> suggest, indicate, you know, the verb is a little
[13:38:26] <scribe> unclear, they said it will improve, that for two more
[13:38:31] <scribe> generations of micro electronics, this probably is not an
[13:38:32] --- apetrescu has left: Replaced by new connection
[13:38:35] <scribe> engineering problem. And if you believe what the crystal
[13:38:39] <scribe> balls say about the life cycles of those generations of micro
[13:38:43] --- apetrescu has joined
[13:38:43] <scribe> electronics, that looks like 5 to ten years is entirely possible
[13:38:45] <scribe> without doing anything but engineering.
[13:38:51] <scribe> All right. And the interesting side observation which has
[13:38:54] <scribe> been made since then was not made in the
[13:38:56] <scribe> workshop to my recollection, is that the
[13:38:59] <scribe> routers m the core of the internet, the ones that essentially
[13:39:03] <scribe> become hot boxes in this terminology,
[13:39:07] <scribe> their scaling seems to be dominated by line speed, obviously,
[13:39:10] <scribe> multiplied by functional complexity in the voiding
[13:39:13] <scribe> path. In other words, every time you add a function to
[13:39:21] <scribe> what a router has to do, to every packet, that has mult plate tiff effect
[13:39:26] <scribe> on scaling, and the engineering of scaling up those Dee advisees.
[13:39:31] <scribe> We also had a debate today, rather too late to get a change in the
[13:39:34] <scribe> slides, on whether we should say no need nor panning,
[13:39:39] <scribe> because that makes people start panicking. But I
[13:39:43] <scribe> think the message is here and it's something that will come out of
[13:39:47] <scribe> the discussion I hope. p people who build these route
[13:39:49] <scribe> terse, they have engineering to do but they are not panicking.
[13:39:56] <scribe> NEW SPEAKER: Brian? NEW SPEAKER: Yes. NEW SPEAKER:
[13:40:01] <scribe> Back up one. NEW SPEAKER: David knows what I'm about
[13:40:07] <scribe> to say, the one other thing that one would add there is that the
[13:40:12] <scribe> scaling is also downgraded by the physics at 45 man-hours. We
[13:40:17] <scribe> had a workshop where we got a bunch of us together and we were
[13:40:24] <scribe> told the big thing, the hard concern is to keep the power
[13:40:25] <scribe> something. NEW SPEAKER: Yes, that's the sort of
[13:40:29] <scribe> thing where you can discuss for a long time, you know, and
[13:40:33] <scribe> try to figure out what it means for your product plans. But you
[13:40:36] <scribe> know, you're not going to have that discussion with some of the
[13:40:39] <scribe> other people in this room for obvious reasons.
[13:40:48] <scribe> Okay. This is, somebody forgot this. I think you two are
[13:40:52] <scribe> arguing whether it should be it tall sized today or not.
[13:40:56] <scribe> It does seem to be the case that, you know, in that
[13:41:00] <scribe> statement, that no need for panic statement, we're still not
[13:41:03] <scribe> following the architectural principal pel of avoiding global
[13:41:07] <scribe> impact to global decisions, because that's sort of the result of
[13:41:10] <scribe> how we do things today. And we spent a lot of time arguing
[13:41:14] <scribe> about whether there is a problem, crystal ball predictions, higher
[13:41:18] <scribe> architecture, et cetera. I think the we in this case is
[13:41:23] <scribe> clearly the A Ds, that I'm currently channeling. We'd
[13:41:26] <scribe> like to put that debate behind us, because we don't think it's
[13:41:31] <scribe> productive to go on having that debate and to focus on whether this
[13:41:35] <scribe> is useful work, without breaking backwards compatibility, et
[13:41:39] <scribe> cetera. And I suspect there's a lot hidden in the et cetera.
[13:41:43] <scribe> So if the A Ds want to expand on the statement afterwards,
[13:41:44] <scribe> I'll leave that to them.
[13:41:50] <scribe> Is there a problem, this is a way of looking at the issue of
[13:41:56] <scribe> the update rate of GP. It's often been said that even if we
[13:42:00] <scribe> contain the size of the table, the routing information base.
[13:42:04] <scribe> G four dine mix has to say the number of rate of update
[13:42:09] <scribe> messages will sat rate something. Because there are more and
[13:42:13] <scribe> more messages something is going to get saturated.
[13:42:18] <scribe> There's a lot of activity that's potentially redundant.
[13:42:22] <scribe> Transient routing takes you back to exactly where you were before they
[13:42:23] <scribe> started, for example.
[13:42:28] <scribe> We don't seem to have an analytical model for that, and it
[13:42:32] <scribe> gets problematic, it means again, that the update rate
[13:42:37] <scribe> needs to scale sub linearly, otherwise you end up with hot
[13:42:38] <scribe> wires.
[13:42:41] <scribe> This problem continues to investigation, and this
[13:42:44] <scribe> is the sort of thing the routing area meeting tomorrow is
[13:42:49] <scribe> supposed to be looking at. Transparency problem, so, I
[13:42:53] <scribe> coind this phrase, thing that looks like an address, it's a
[13:43:01] <scribe> six letter acronym, T L.A. Is three letter acronym, for those who
[13:43:06] <scribe> don't know it. So T T L L.A. A. Since you know, the very
[13:43:13] <scribe> beginning of T C P IP when they see something that looks like an
[13:43:18] <scribe> address, three 2 bits long or 8 bits long. Is an address.
[13:43:22] <scribe> And that means that application programmers assume they can
[13:43:23] --- hartmans@jis.mit.edu/owl has left
[13:43:26] <scribe> pass an IP address around to third parties.
[13:43:28] <scribe> That's called, you know, that's called
[13:43:31] <scribe> referrals in the jargon that's built up.
[13:43:35] <scribe> And the reality is that addresses are sometimes only locate
[13:43:39] <scribe> terse today. Your address is ten Dot one Dot one, Dot one.
[13:43:42] <scribe> It's only a location that works within the hotel. If it
[13:43:47] <scribe> gets outside the hotel, it's meaningless. And you know, this
[13:43:50] <scribe> can cause quite a bit of discussion. It's my, one way of
[13:43:55] <scribe> looking at it is that nats are a solution to a problem, and
[13:43:59] <scribe> the problem is that addresses are sometimes just locate tors
[13:44:02] <scribe> and not identifiers. And trouble is
[13:44:06] <scribe> nats cause their own problems, which is where things like stun come
[13:44:09] <scribe> from. Those of you who looked at stun and ice and so on,
[13:44:13] <scribe> know that these are very complex super structures built to deal
[13:44:18] <scribe> with the problem that addresses application level identifiers.
[13:44:22] <scribe> In addition to that, the way we've looked at addresses being
[13:44:25] <scribe> transparent object objects that can be
[13:44:28] <scribe> carried from one end of the network to the other and is still mean
[13:44:32] <scribe> the same thing, does create difficulty in traffic engineering
[13:44:35] <scribe> and codecs. This problem seems to need a solution.
[13:44:41] <scribe> So solution directions, tell me when I need to stop. NEW SPEAKER:
[13:44:45] <scribe> Doing okay. NEW SPEAKER: Okay. So the scaling of
[13:44:50] <scribe> the fib e-mails, the regimen seems to be at the moment
[13:44:56] <scribe> engineering, by micro electronics people and router
[13:44:59] <scribe> designers, and we're supposed to be doing engineering in tivt
[13:45:01] <scribe> buts not that engineering, that's product engineering.
[13:45:09] <scribe> The update dine mix dynamics, may require
[13:45:11] <scribe> adjustment and better operational practices,
[13:45:14] <scribe> those are the sort of things we can tackle in the IETF.
[13:45:19] <scribe> You know, basic requirements like traffic engineering, multi
[13:45:25] <scribe> homing, this end to end issue, mobility, seems like they
[13:45:30] <scribe> would bern fit from architectural change. And the approaches
[13:45:33] <scribe> that are being look at there are either straightforward
[13:45:38] <scribe> identifier locater separation, so they become two completely
[13:45:42] <scribe> separate name spaces, or multi level
[13:45:46] <scribe> locaters, where you, you know, have a different locators
[13:45:47] --- TJ has left
[13:45:50] <scribe> for the same session in the internet from when you're at the end
[13:45:58] <scribe> user sites. And which is actually the purpose of what's being
[13:45:59] <scribe> discussed tomorrow.
[13:46:02] <scribe> And all of those points are definitely oring
[13:46:07] <scribe> national, and exactly the same issues as IP V four and they're
[13:46:09] <scribe> orthagonal to anything you wanl to do at application level
[13:46:13] <scribe> space. You have to be very careful not to confuse the
[13:46:17] <scribe> identifier issue at the network level with the identifier issue
[13:46:20] <scribe> at application level. And you know, when you confuse those two,
[13:46:23] <scribe> you enter trouble.
[13:46:26] <scribe> There's more thinking about some of this in the URLs, that are
[13:46:34] <scribe> listed at the end of this presentation. Three summary slides
[13:46:39] <scribe> and then I'm done and we'll run quickly. Except I need to
[13:46:45] <scribe> get my something back. NEW SPEAKER: The IETF's
[13:46:48] <scribe> role, so, it seems we can provide a forum for problem
[13:46:53] <scribe> analysis and an open way, because we can get vendors, users
[13:46:59] <scribe> and operators together here. Actually, one thing I'm very glad
[13:47:03] <scribe> about personally, is that this whole discussion of the last
[13:47:07] <scribe> ten months has re launched discussion between the IETF and some
[13:47:11] <scribe> of the operations people. Who have been a little bit away from the
[13:47:16] <scribe> IETF. And we can provide a forum for open development of
[13:47:18] <scribe> solutions. Same players.
[13:47:23] <scribe> We can't do research but fortunately, we have some people who
[13:47:27] <scribe> can. On the IETF. Aaron is probably some where in the
[13:47:30] <scribe> room. Yes. He's right in front of me.
[13:47:35] <scribe> And we can, above all, we cannot control economic business
[13:47:39] <scribe> behavior, so although we know that what we do has economic and
[13:47:42] <scribe> business impact, we have no control over that. We cannot
[13:47:46] <scribe> predict you you know, how economic players will
[13:47:47] <scribe> react to particular things we do.
[13:47:55] <scribe> Technically, routing table growth, quote only injurious in the
[13:47:59] <scribe> short term. Routing dynamics needs to be better
[13:48:03] <scribe> understood. But it's like ae an engineering issue, where we
[13:48:07] <scribe> have said, we think we can do things in the I S P community
[13:48:10] <scribe> and we think they're protocol improvements and implementation
[13:48:12] <scribe> improvements.
[13:48:16] <scribe> So there is reason to believe that we do not have a short term
[13:48:19] <scribe> technology problem. There's a lot of hard work ahead. And
[13:48:23] <scribe> once some of this gets translated into products and services,
[13:48:27] <scribe> there are businesses uses which are not our business, there
[13:48:30] <scribe> are architectural problems. We talked about that. We can
[13:48:37] <scribe> help in the short term protocol work such as doing,
[13:48:42] <scribe> adjusting, tuning V GP. And we can help by looking at
[13:48:44] <scribe> architectural changes such as I D lock, separation, which
[13:48:48] <scribe> is how we came to focus in on these two meetings tomorrow.
[13:48:52] <scribe> Because it's the most constructive way to go forward on this
[13:48:53] <scribe> whole problem space.
[13:48:58] <scribe> The overall plan is to divide and conquer plan. We need to
[13:49:01] <scribe> continue talking with the operator community. We need to pursue
[13:49:06] <scribe> the implementation and incremental V GP improvements need to be
[13:49:09] <scribe> evaluated. We can't commit to them before we evaluate them.
[13:49:15] <scribe> We need to encourage research and expert ment taition in I D
[13:49:18] <scribe> lock separation multi level locator
[13:49:26] <scribe> approaches. And if that looks hopeful, we need to define
[13:49:28] <scribe> solution along those lines.
[13:49:35] <scribe> And that's the sort of program, and you can't read
[13:49:41] <scribe> those URLs, because I don't know why some artifact with the
[13:49:46] <scribe> way Gary uses power point. But the presentation in P D F form is
[13:49:50] <scribe> already on the proceedings page so you can find the URLs. And this
[13:49:55] <scribe> is where I run away. NEW SPEAKER: I'm going to take my
[13:49:56] <scribe> computer with me.
[13:50:06] <scribe> NEW SPEAKER: Thank you, Brian. And at this point, open
[13:50:09] <scribe> up for questions. ted? NEW SPEAKER: Well, first I'd
[13:50:16] <scribe> like to thank citizen carpenter for for the presentation.
[13:50:19] <scribe> ( Applause. ) NEW SPEAKER: It took an interesting
[13:50:24] <scribe> topic and not for a lack of panic, but possibly not into feeling
[13:50:28] <scribe> enough important. So I really feel like that is a useful
[13:50:30] <scribe> presentation but probably way too calm.
[13:50:37] <scribe> I got up at the mic phone last time, to ask the IETF to be a big
[13:50:40] <scribe> boulder in how it tackled this problem, which I
[13:50:44] <scribe> believe is both very serious and very real. And I, I don't
[13:50:47] <scribe> see any evidence that that has been taken into account, even
[13:50:53] <scribe> though I have said it privately to many people. Dotted and not dotted.
[13:50:59] <scribe> Looking at this problem space, a lot of people have the believe
[13:51:04] <scribe> that separating what they believe to be a locater, and
[13:51:08] <scribe> separating what they believe to be an identifier is an easy
[13:51:11] <scribe> problem. Frankly, from everything I've seen, if you
[13:51:15] <scribe> think the current set of directions is hopeful, you are hoping
[13:51:20] <scribe> for the wrong thing. Because you are creating an enormous
[13:51:23] <scribe> rendezvous problem which you have no idea how to solve. That
[13:51:25] <scribe> doesn't help scaling.
[13:51:30] <scribe> We have an enormous set of nodes in our network which are
[13:51:33] <scribe> mobile. We are currently dealing with them through a set of
[13:51:37] <scribe> tunneling technologies which is fragile and which, if you
[13:51:42] <scribe> change this set of technologies, not only might break,
[13:51:44] <scribe> almost certainly will break. The architectural level of this
[13:51:49] <scribe> discussion needs to go up at least a few thousand feet and
[13:51:52] <scribe> maybe a few tens of thousands of feet. I am happy to hear that
[13:51:57] <scribe> the short term engineering problem doesn't need, to force us to
[13:52:01] <scribe> make a decision too fast. But if you don't take that
[13:52:06] <scribe> architectural leap up those tens of thousands of feet, you're going to end up with this problem
[13:52:11] <scribe> biting you in the knees over and over again, until this
[13:52:12] <scribe> kills the internet.
[13:52:20] <scribe> So, once again, be bowled err, be more panicked,
[13:52:27] <scribe> please, and don't assume that any of the current directions are
[13:52:29] <scribe> solving the problem at this point. Because from everything I've
[13:52:35] <scribe> seen, on RAM or the workshop lists, et sets, we're not
[13:52:36] <scribe> even in the right room yet.
[13:52:42] <scribe> NEW SPEAKER: Okay. So, personally, I, I don't want
[13:52:44] <scribe> to be panicked. I want to get to action. So, NEW SPEAKER:
[13:52:48] <scribe> Anything you can do that isn't panicked against people to active
[13:52:54] <scribe> participation is fine. But frankly, around here, I found
[13:52:55] <scribe> panic tends to work. (Several people
[13:52:55] <scribe> talking.)
[13:53:00] <scribe> NEW SPEAKER: There's different forms of panic, right.
[13:53:03] <scribe> There's panic where you just do something just to do
[13:53:06] <scribe> something, and if it's silly, you don't care. And then
[13:53:10] <scribe> there's panic where you really put your mind to it and think
[13:53:13] <scribe> hard and talk to the smart people you know and try to bounce
[13:53:17] <scribe> around ideas, and my own personal belief is that we're not
[13:53:21] <scribe> going to come up with a completely fundamental new solution to
[13:53:24] <scribe> this problem without there being some cost.
[13:53:29] <scribe> And at this point, I don't think we fully understand all of
[13:53:33] <scribe> the trade offs between possible solutions, nor am I
[13:53:36] <scribe> completely convinced that we understand exactly all the
[13:53:40] <scribe> implications of all the possible solutions. So one thing
[13:53:43] <scribe> that's good to know is that we do see a path to keep things
[13:53:46] <scribe> working over the next, you know, ten years. But at the same
[13:53:50] <scribe> time, we understand z that this problem is a hard problem.
[13:53:54] <scribe> And so we need to think hard about it and don't, you know,
[13:53:59] <scribe> drink beer and get fat and ignore it. Because we got ten
[13:54:04] <scribe> years, and we'll worry about it 9 years and 11 months from now. Because
[13:54:09] <scribe> it might take more than a month to completely test and deploy
[13:54:13] <scribe> and spec and test and think through.
[13:54:18] <scribe> And I think there are plenty of people that are panicked enough to
[13:54:20] <scribe> have people thinking very hard about this problem.
[13:54:24] <scribe> I've certainly spent a lot of time since the Amsterdam
[13:54:30] <scribe> workshop talking with a great many people from, you know, multiple companies,
[13:54:35] <scribe> vendors shs service providers, researchers, about posz
[13:54:39] <scribe> solutions, and I'll admit, I feel I'm still
[13:54:42] <scribe> gathering information and filling up my brain. And you know,
[13:54:45] <scribe> when my brain gets full, I'll take a few hours off and get
[13:54:48] <scribe> back to it and get more in later. But, you know,
[13:54:54] <scribe> I think we are thinking very hard about this issue, and I don't think, there might
[13:54:58] <scribe> be some people that are panicked. I don't think we need to panic in
[13:55:01] <scribe> the sense of doing something silly of the but doy think we need
[13:55:04] <scribe> to panic in the sense of taking this very seriously and working
[13:55:09] <scribe> on it. NEW SPEAKER: Yes, and further response to
[13:55:14] <scribe> that on this issue of raising the discussion to hire
[13:55:19] <scribe> architectural level, I agree on that. And also, it's
[13:55:22] <scribe> very clear that solutions we have right now on the table,
[13:55:26] <scribe> just the beginning. We really need to think about this hard and there are
[13:55:32] <scribe> really many traps that we can fall into, you know, so,
[13:55:36] <scribe> yes, we believe that this is a problematic area to find a
[13:55:40] <scribe> solution. So, it's not like we are planning to, you know,
[13:55:51] <scribe> approve the next RFC on this tomorrow. NEW SPEAKER:
[13:55:57] <scribe> Sensor dawk kins. A couple of things. The first one, and this is a question
[13:56:01] <scribe> that you guys don't have to answer now, but probably have to worry about
[13:56:07] <scribe> it a little bit. Is did thing with B GP and the things where it gets
[13:56:12] <scribe> wedged into non routing solutions. I don't know how that
[13:56:16] <scribe> sort of dynamic, that was not a set of
[13:56:19] <scribe> dynamics that we heard in the presentation. But as I say, you
[13:56:24] <scribe> don't have to solve the problem now, or worry about it. Just
[13:56:25] <scribe> wanted to ask.
[13:56:29] <scribe> NEW SPEAKER: I'm sorry. I missed a bit of that. Can you
[13:56:30] <scribe> --
[13:56:35] <scribe> NEW SPEAKER: It's the work that Randy bush was talking
[13:56:40] <scribe> about, for some reason. NEW SPEAKER: V GP. Yes. NEW SPEAKER:
[13:56:44] <scribe> There have been discussions, I've certainly been involved in
[13:56:48] <scribe> discussions relating to B GP dynamics for quite
[13:56:52] <scribe> a long time. Probably, I don't know, probably as long as
[13:56:58] <scribe> there's been V GP, frankly, probably fifteen years. And it's
[13:57:02] <scribe> been clear for quite a while that there are some things you can
[13:57:05] <scribe> do to make things better. And some of those would be along
[13:57:11] <scribe> the lines of operational practices, some of them might be along
[13:57:14] <scribe> the lines of what policies do you allow or not allow. Some might
[13:57:18] <scribe> be along the lines of tweaks in
[13:57:19] <scribe> implementations or tweaks in
[13:57:24] <scribe> protocols. And there are discussions going on now about, you
[13:57:28] <scribe> know, improvements. In fact, that's why I didn't go to the
[13:57:31] <scribe> social last night, is that some of us spent that time thinking
[13:57:33] <scribe> about these issues instead.
[13:57:38] <scribe> Now, understanding there's not yet a lot of public discussion, we
[13:57:41] <scribe> are planning to talk about this at st routing area meeting
[13:57:50] <scribe> tomorrow. And you know, I think it sort of might be a group
[13:57:55] <scribe> of 200 people in a room don't often come up with fundamental original
[13:57:58] <scribe> brand new approaches to solving a problem. And instead, a
[13:58:01] <scribe> number of different people go off and number of different individuals
[13:58:04] <scribe> go off and think about it and various people come up with
[13:58:08] <scribe> various ideas and then you bring them into a bigger group and present
[13:58:12] <scribe> them and discuss them. Then you go home and come back and yell
[13:58:15] <scribe> each other and go back home and think some who are. And
[13:58:19] <scribe> eventually, sensible ideas coalesce. And sh of this also
[13:58:23] <scribe> comes down to, do you have to do it or not? If
[13:58:26] <scribe> there's something that's going to require vendors to implement
[13:58:30] <scribe> stuff, I S Ps to configure something or maybe vendors to try
[13:58:35] <scribe> to make it easier to configure or the IETF to design it so it's
[13:58:38] <scribe> easier to configure, there's still some pain somewhere.
[13:58:40] <scribe> So, you sort of look at the problems that are occurring in the
[13:58:45] <scribe> world, and you do as much as you can motivate people to do to
[13:58:49] <scribe> avoid the pain that's there. But by definitely think that this is a
[13:58:53] <scribe> topic that we should be working on, and there definitely are people thinking
[13:58:57] <scribe> about it and we do have a discussion planned for tomorrow. NEW SPEAKER:
[13:59:03] <scribe> Second thing, so there was this, on the I D
[13:59:07] <scribe> locator's list, I was going to ask if you guys could please
[13:59:11] <scribe> be sure that you get enough input from the applications guys.
[13:59:16] <scribe> It's very easy to go off and do something that should work, but
[13:59:22] <scribe> doesn't. Probably the most recent good example of a good
[13:59:26] <scribe> idea that didn't work at the application level, so, it's
[13:59:30] <scribe> not that these are stupid things. It's just that there needs
[13:59:31] <scribe> to be communication.
[13:59:37] <scribe> I'm already starting to see this big, like I said, don't
[13:59:41] <scribe> panic. I'm already starting to see people that are trying to
[13:59:45] <scribe> use SIP in a way that they were trying to use S U D P 5 years
[13:59:49] <scribe> ago and I'm looking forward to not seeing the RFC z on the use
[13:59:53] <scribe> of SIP as a transport substrate.
[13:59:59] <scribe> But I mean, it's like, oh, well, we got to get through
[14:00:02] <scribe> nats, and we can't mess around, because Brian said, sub
[14:00:07] <scribe> addresses are ambiguous. I think the peer to peer SIP view is that
[14:00:11] <scribe> all of them are. All of the interesting use cases or
[14:00:14] <scribe> virtually all of them are. So, you know, basically,
[14:00:19] <scribe> something like address of record starts a at the application
[14:00:22] <scribe> level, is a really good plan. And it's going to be very
[14:00:24] <scribe> attempting for people to do that.
[14:00:32] <scribe> We had lots of research groups that were can kind of focused on
[14:00:34] <scribe> that. We're having a conversation with peer to peer SIP that say
[14:00:38] <scribe> basically, oh, we can use SIP to get through nats, you
[14:00:43] <scribe> know, be in this room today. So, and like I said, please just
[14:00:50] <scribe> be aware that, of that way for the horse to run out of the barn
[14:00:55] <scribe> when you're thinking of the routing, you know, the router guys are
[14:01:01] <scribe> all so happy. But this may still be before anybody notices. NEW SPEAKER:
[14:01:04] <scribe> One of the primary reasons we, thank you, Spencer, one of the
[14:01:08] <scribe> primary wons we asked for this time at a plenary level
[14:01:14] <scribe> discussion, was to bring in this to all areas, so your point is very
[14:01:19] <scribe> well taken, and we definitely want, we need to take in all
[14:01:22] <scribe> aspects, even if the core of this, we're, the people
[14:01:28] <scribe> focused on one piece of it. But, certainly, we need to take
[14:01:31] <scribe> into that account, and we've talked to some applications people
[14:01:34] <scribe> and we need to talk more. They need to come contribute, et
[14:01:39] <scribe> cetera. There's lots of people. So I probably need to
[14:01:42] <scribe> cut, and move to the next person and load balance. NEW SPEAKER:
[14:01:44] <scribe> I'm only saying, thank you, good job. NEW SPEAKER:
[14:01:46] <scribe> Okay. Thanks Spencer.
[14:01:52] <scribe> And maybe we need to do some load balancing. I don't know who is next. But I
[14:01:55] <scribe> know Dave Meyer was there for a while. Do you want to go? NEW SPEAKER:
[14:02:03] <scribe> So what's let's let Dave go. Dave Meyer, when he was on
[14:02:07] <scribe> the I A B, really was the catalyst for bringing this problem to
[14:02:11] <scribe> light, or bringing the problem of say the last two decades to light
[14:02:13] <scribe> again and realizing that it's never too late to
[14:02:16] <scribe> continue to look at this problem. So thank you very much,
[14:02:18] <scribe> Dave. ( Applause.)
[14:02:23] <scribe> NEW SPEAKER: Now it's your turn. NEW SPEAKER: Dave
[14:02:28] <scribe> Meyer, resident difficulties disdent. A
[14:02:32] <scribe> couple of things, I have to spon respond to what Mr. Moore
[14:02:36] <scribe> just said. I had the lever to pull, but a lot of my com
[14:02:39] <scribe> leagues colleagues were involved in the
[14:02:43] <scribe> formulation of the idea. I just got the community, all of
[14:02:46] <scribe> you or non com or whatever, where I had the lever and I
[14:02:48] <scribe> pulled it. That's all I did.
[14:02:52] <scribe> Secondly, what I really came up here to say is, a couple of
[14:02:55] <scribe> things and then Jim added one so I want to give Jim some air
[14:02:58] <scribe> time on this. But basically, first off, we were having
[14:03:01] <scribe> this discussion about panicking. I would like to
[14:03:05] <scribe> congratulate ted on his incredibly clear thinking on this topic
[14:03:09] <scribe> and now we said it twice. We're fifteen years into this. I don't
[14:03:13] <scribe> think anybody is panicking or panic horizon is fifteen
[14:03:18] <scribe> years. I mean, so are we going to go another fifteen years
[14:03:21] <scribe> guys or are we going to go 20. We've got to do something or
[14:03:25] <scribe> nothing. And I think, by the way, when rus was talking
[14:03:30] <scribe> about incremental improvements to V GP, why weren't we doing that
[14:03:34] <scribe> anyway. That has nothing do do with this. We need
[14:03:38] <scribe> improvements to V GP because that's what we do. That is not what this is
[14:03:38] <scribe> about.
[14:03:44] <scribe> Secondly, Jim just wanted me to mention that we need to be
[14:03:48] <scribe> aware, and I think he's correct, that we need to be aware that
[14:03:51] <scribe> there may be two or three or maybe even more kinds of solution
[14:03:54] <scribe> spaces to this problem that operate rate on different time
[14:03:57] <scribe> scales and we really need to keep that in mind when we're talking
[14:03:57] <scribe> about this.
[14:04:01] <scribe> So, I think those are important points that Jill
[14:04:06] <scribe> wanted me to make of the but the final thing I want to say, I
[14:04:08] <scribe> think this is a controversial topic. It's going to be hard
[14:04:13] <scribe> to get, you know, the classic IETF consensus. So I think
[14:04:17] <scribe> this is not only an issue of our technology, but an issue of our process.
[14:04:21] <scribe> And the first 5 or ten minutes or whatever of this we
[14:04:22] <scribe> talked, we were talking about process.
[14:04:27] <scribe> So, I'm interested in whether or not we have anyway forward in our
[14:04:31] <scribe> process to the deal with this. And I haven't heard anything about
[14:04:35] <scribe> that, other than I've heard about e incremental improvements,
[14:04:38] <scribe> not panicking. I don't know who is
[14:04:41] <scribe> panicking. It's fifteen years later, nobody is
[14:04:43] <scribe> panicking. I've heard about architecture. I've heard nothing
[14:04:47] <scribe> about how solutions are going to come out. And I think that's
[14:04:51] <scribe> squairm squairlly on the IESG. NEW SPEAKER: Dave, I don't agree with
[14:04:54] <scribe> you, that you haven't heard anything about how solutions came
[14:04:57] <scribe> out. I was talking about that about three minutes ago.
[14:05:00] <scribe> 200 or, or I don't know how many people are in this room,
[14:05:02] <scribe> but hundreds of people sit anything the room, we're not going
[14:05:06] <scribe> to sit here in a group of several hundred people and design the
[14:05:08] <scribe> next generation solution. NEW SPEAKER: Nobody
[14:05:10] <scribe> suggested that. NEW SPEAKER: Nobody suggested that. NEW SPEAKER:
[14:05:13] <scribe> We're talking about process issues here, how the IETF is
[14:05:16] <scribe> going to solve this problem. Not tonight. I understand
[14:05:16] --- apetrescu has left: Replaced by new connection
[14:05:18] <scribe> we're not going to do it tonight. NEW SPEAKER: You need to have
[14:05:22] <scribe> ideas, you need to have an understanding of the issues.
[14:05:25] <scribe> It's a complicated issue and a lot of people have to work hard on this.
[14:05:31] <scribe> We have to thank you, really, for applying a cattle prod to
[14:05:36] <scribe> the prod to the proper place. And I think we're moving. I'm
[14:05:39] <scribe> not sure we're all moving in the same direction. But that's
[14:05:42] <scribe> okay. That's part of exploring possibilities, is to move in different
[14:05:47] <scribe> directions. If what you're saying is, we weren't movk
[14:05:51] <scribe> movring for fifteen years, I agree completely, and I thank you for
[14:05:55] <scribe> applying that cattle prod and continue to run behind us, onto
[14:05:56] <scribe> make sure we keep moving.
[14:06:01] <scribe> We have to get a lot of smart people thinking hard about this
[14:06:03] <scribe> problem and thinking through the implications and not jumping to
[14:06:07] <scribe> rapid conclusions until we really understand the trade
[14:06:10] <scribe> offs, but we've got to work on it and think about it and some of
[14:06:13] <scribe> us have to start writing more things down as well and bouncing
[14:06:16] <scribe> them around and so on. NEW SPEAKER: So in addition, one
[14:06:22] <scribe> thing that should be clear, the end game is just not modern
[14:06:27] <scribe> improvements to V GP. If that was over emphasized, sincere
[14:06:30] <scribe> apologies. Those are things that we should be doing and I'm
[14:06:34] <scribe> not going to sit up here as somebody who work on V
[14:06:38] <scribe> GP of why we haven't done a better job of getting that into the
[14:06:38] <scribe> protocol.
[14:06:43] <scribe> With the difficulty that the I S G and I guess I'm the news
[14:06:48] <scribe> member sitting up here newest member sitting here up
[14:06:52] <scribe> here. The process that's been laid out is the, a lot of
[14:06:55] <scribe> the action at this point is going to happen in R O G. That's
[14:06:59] <scribe> basically been effectively for the time being, the selected
[14:07:02] <scribe> place where this discussion is going to occur. And
[14:07:04] <scribe> discussions of design space and design parameters,
[14:07:08] <scribe> bounding the solution space of what we're going to attempt to solve is going to be
[14:07:14] <scribe> discussed. There has been, as Brian mentioned, a, you
[14:07:20] <scribe> know, the director put together to attempt to define the
[14:07:24] <scribe> boundaries of the problem as well as the solution to R O G. We
[14:07:26] <scribe> are having the discussions tomorrow and so a discussion is
[14:07:30] <scribe> cheap. Discussion doesn't necessarily allow us to move
[14:07:31] <scribe> forward.
[14:07:35] <scribe> But given the catalyst of having to fit within the process that
[14:07:39] <scribe> exists, which is very difficult to do, extremely large scale
[14:07:43] <scribe> architectural work within this process, that's what is going to
[14:07:47] <scribe> be put in place to date and it's certainly got to move forward from
[14:07:50] <scribe> there. It's still not the end game. And my understanding
[14:07:55] <scribe> of re chart erring of R O G of recommendations of
[14:07:59] <scribe> solutions is what is going to come out of the R O G and that's
[14:08:03] <scribe> certainly what we're looking for. The design space is incredibly
[14:08:05] <scribe> large. The number of variables in trying
[14:08:08] <scribe> to find a solution is absolutely enormous. And we can talk
[14:08:11] <scribe> about panics or other metaphors
[14:08:16] <scribe> of cancer or global warming, but that's really probably the best
[14:08:18] <scribe> analogy that we can give, because this is a long term problem that
[14:08:22] <scribe> needs to be fixed and it may be just as difficult as others
[14:08:23] --- jclee has joined
[14:08:25] <scribe> have said on the list. NEW SPEAKER: I'll try to make
[14:08:27] <scribe> this brief. Dave, you mentioned you had a lever to pull and you
[14:08:31] <scribe> pulled it. We have a certain set of levers as
[14:08:36] <scribe> well and I think we pulled them pretty much, you know, full
[14:08:39] <scribe> stop over the past six months. Inasmuch as we can do.
[14:08:45] <scribe> Okay. You know, two full working group areas
[14:08:49] <scribe> tomorrow, a plenary discussion here, a directorate formed, a lot of discussion.
[14:08:55] <scribe> Like you said, execution is cheap. But there's also been work
[14:08:59] <scribe> spawnd off. We've seen it and we've tried to comprehend it.
[14:09:06] <scribe> But, the intention for very hard work is certainly here and we're
[14:09:09] <scribe> pulling all the levers that we have. And well, if
[14:09:14] <scribe> there's anything else we can do, we could certainly use the advice.
[14:09:22] <scribe> NEW SPEAKER: Okay. Come back random engineer, very
[14:09:26] <scribe> distant past, I used to be a packet pusher, I have lots of
[14:09:32] <scribe> gray hair to prove it. I just wanted to insert a sort of ten
[14:09:37] <scribe> kilometers of suggestions here to really think of the people
[14:09:41] <scribe> involved in this, think a little bit and to learn from our
[14:09:46] <scribe> experiences. In this way, what comes to mind, No. 1 that
[14:09:51] <scribe> comes to my mind is side err. You know, there was some panic about
[14:09:55] <scribe> writing table growth at some point. And we needed a quick fix,
[14:10:01] <scribe> so we sd a quick fix. Try to think, try to think back how that
[14:10:05] <scribe> happened. Maybe ask people who were around at the time. NEW SPEAKER:
[14:10:08] <scribe> I think the talk about that, if you want to. NEW SPEAKER:
[14:10:10] <scribe> No, hang on. I was there. NEW SPEAKER: I'm not
[14:10:15] <scribe> finished. I'm just saying, think what you want to learn
[14:10:18] <scribe> from that. Think from other
[14:10:20] <scribe> experiences, like there was a panic about address space
[14:10:24] <scribe> running out. What that did is IP v6. And some of the
[14:10:29] <scribe> features that's call them features, of IP v6 is because we
[14:10:33] <scribe> rushed them. To my mind. So think what we can learn from
[14:10:39] <scribe> that. And think really sort of independent mode, the
[14:10:44] <scribe> solution to some of this has, is a bigger space than the one solution.
[14:10:49] <scribe> So you might want to see what one can learn from past
[14:10:52] <scribe> experiences and what we, in order not to repeat all the
[14:10:55] <scribe> mistakes. And that's pretty philosophical and I don't want to
[14:10:59] <scribe> go into the details and start a debate about past
[14:11:00] <scribe> experiences, think about the
[14:11:05] <scribe> experiences and make your own judgments and try not to repeat
[14:11:09] <scribe> all the mistakes. NEW SPEAKER: We've got 20 minutes left
[14:11:12] <scribe> if we're not going to go over time.
[14:11:15] <scribe> So that's a little bit more than a minute per person. I
[14:11:18] <scribe> don't mind staying here all night. But just for everybody. NEW SPEAKER:
[14:11:27] <scribe> Quick. NEW SPEAKER: Tim shepherd. Some thank you.
[14:11:30] <scribe> I'm very encouraged that we're having more and more discussion
[14:11:35] <scribe> about this kind of stuff. There was a time prior to the
[14:11:40] <scribe> IETF, in soul Korea, where it seemed that our talking about po attention shall
[14:11:46] <scribe> issues, and starting at the mren Rees there, I remember being
[14:11:50] <scribe> encouraged and I've seen study ramp off. And think we're
[14:11:51] <scribe> going into the right direction.
[14:11:55] <scribe> I'm not entirely clear that we're actually, that we've
[14:12:00] <scribe> actually figured out what the right process is for deciding to
[14:12:03] <scribe> make architectural changes to the internet. And I think
[14:12:10] <scribe> making those decisions is some in sense, so important that we
[14:12:17] <scribe> can't just leave it to the, at least the existing process.
[14:12:21] <scribe> So, please think carefully about how we go about making that
[14:12:21] <scribe> decision.
[14:12:26] <scribe> It's too important to be left to the IETF. It's too important
[14:12:30] <scribe> to be left to the I A B. I was just difficulties
[14:12:35] <scribe> mayd at the I A B internally cannot make architectural
[14:12:39] <scribe> decisions or they the don't think they have that power anymore.
[14:12:40] --- Jack3010 has left
[14:12:49] <scribe> So, and you know, please go forward.
[14:12:53] <scribe> One minor lesser point, but it's still something I want to
[14:12:58] <scribe> say, all the talks that Brian went through, discussion, and
[14:13:03] <scribe> about how to go forward, I think one of the key issues in
[14:13:05] <scribe> something like, and I'm just trying to use this as the
[14:13:10] <scribe> example of it, of I B locater stuff, I think almost any
[14:13:13] <scribe> architectural change is going to help us in this direction is going to have a
[14:13:17] <scribe> tricky implications in another direction, and in particular,
[14:13:20] <scribe> security. I think there's been a history in the past of
[14:13:24] <scribe> leaving security, security bugs. But in fact,
[14:13:29] <scribe> how you solve the security issues are so sort of ingrained to the
[14:13:33] <scribe> then how you can trust it and use it afterwards, that you
[14:13:37] <scribe> have to make sure that you're not leaving security architecture
[14:13:43] <scribe> you decisions from the security requirements decisions to those
[14:13:45] <scribe> people who spend all of their time in the security working
[14:13:48] <scribe> groups in the IETF. There's some fairly fundamental decisions
[14:13:53] <scribe> about what it means to trust some look up or some resolution or
[14:13:57] <scribe> some referral means something, and if you don't make some
[14:14:02] <scribe> architectural choices early on there, and then some people
[14:14:03] --- DonFedyk has left
[14:14:05] <scribe> come along later and try to figure out how to secure it,
[14:14:09] <scribe> they're going to try to secure it in ways that perhaps
[14:14:10] <scribe> makes it less useful.
[14:14:14] <scribe> There's a history of securing things to get three
[14:14:19] <scribe> protocols being deployed, because they wind up being too hard to
[14:14:23] <scribe> use. So that, that's sort of just, that sort of stuff
[14:14:26] --- DonFedyk has joined
[14:14:27] <scribe> has to come early in the architectural decisions. I should
[14:14:30] <scribe> probably sit down. NEW SPEAKER: Process wise, part of the
[14:14:34] <scribe> meeting area tomorrow, is design space stuff and security,
[14:14:36] <scribe> and it's in the slide set that I've seen. Thank you. NEW SPEAKER:
[14:14:44] <scribe> Jim, engineering architect. I don't need any
[14:14:50] <scribe> comments back from you. What I want to say is that the IETF journal is in
[14:14:52] <scribe> everybody's packet that's here, you're here for probably
[14:14:58] <scribe> wrote. And there's an article that points to a
[14:15:01] <scribe> document that is called new ar sh, and I believe that's the
[14:15:04] <scribe> guiding principle that IETF should use, the conclusions in
[14:15:07] <scribe> that document. Read it for yourself, if you don't agree, that's
[14:15:12] <scribe> fine. I dhi Dave's point was missed by Ross. The point
[14:15:15] <scribe> is, and I want to repeat it again. One of the boundaries
[14:15:19] <scribe> that we need to complete a solution is the bound the solution
[14:15:20] <scribe> space.
[14:15:24] <scribe> There could be a solution that will work in two years, there could
[14:15:27] <scribe> be one that works in 5 and there's one that could work in ten.
[14:15:31] <scribe> And that boundary and a agreement on if we're going to work on
[14:15:36] <scribe> those time slices has to be understood up front so that everybody on the
[14:15:41] <scribe> mailing list moves forward and right now, RAM and I know
[14:15:46] <scribe> Ross, you just joined it. But the RAM width, we've been following it.
[14:15:50] <scribe> Which I really believe and somebody else who was also an
[14:15:53] --- guido.franceschini has left
[14:15:55] <scribe> author of new ar sh jacked it up and we had to Jack this thing
[14:15:59] <scribe> up to get it right. And then we have to ask the operators if
[14:16:02] <scribe> they think we have a clue, or else we don't have a clue. NEW SPEAKER:
[14:16:05] <scribe> Just a quick comment. I also believe that we need,
[14:16:09] <scribe> possibly have multiple types of solutions at different time
[14:16:14] <scribe> frames, and I think one approach to that is that we, you
[14:16:20] <scribe> know, we have things coming out that we try now, and then we
[14:16:27] <scribe> have other things coming out later. We have routing research
[14:16:31] <scribe> group working on experiments. So I think, and actually they
[14:16:36] <scribe> can try several things over the course of, you know, this
[14:16:45] <scribe> whole process. NEW SPEAKER: Baker ver sign. NEW SPEAKER:
[14:16:49] <scribe> Go. NEW SPEAKER: Okay. Your issues may not be new
[14:16:53] <scribe> architectural issues, so much as applying old architectural
[14:16:55] <scribe> principles, that aren't currently being applied. For
[14:16:59] <scribe> example, on security, people talk about security here, they
[14:17:02] <scribe> immediately think about access control and stopping bad
[14:17:06] <scribe> things from happening. You can't for this application. Think
[14:17:11] <scribe> about the accountability. Think about how to fix things when
[14:17:15] <scribe> things go wrong, find out who was responsible for making them
[14:17:21] <scribe> go wrong. And fix it in the future. At your
[14:17:23] <scribe> level, it's in effect a model, particularly because you're talking
[14:17:27] <scribe> about internet working, not network. Another
[14:17:35] <scribe> issue, encapsulation, a large part of your problem with IP
[14:17:38] <scribe> staff, when the standards were written, there was no DNS.
[14:17:43] <scribe> So all of that grips directly into layer 5. So now when you
[14:17:48] <scribe> want to switch out layer 5 to something else, you got a
[14:17:52] <scribe> series of statements that say, here's how you switch from talking
[14:17:58] <scribe> about A records to quadruple A records. What you need to do
[14:18:01] <scribe> is insert a presentation layer. It doesn't need to be much.
[14:18:06] <scribe> It doesn't need to be as much as the on S I presentation layer. You've got the
[14:18:10] <scribe> beginnings of one with the DNS. But you need to make that the
[14:18:15] <scribe> only way the applications talk to the IP layer. Not just a
[14:18:15] <scribe> way.
[14:18:22] <scribe> Encapsulation hat has got to be black box, object oriented, the
[14:18:26] <scribe> only way you talk about. So as I say, old architectural
[14:18:30] <scribe> principles, but applied to ones we're fot doing today. NEW SPEAKER:
[14:18:34] <scribe> Thank you. I don't think we have time for a lot of
[14:18:36] <scribe> comments, so moving on to the next question here. NEW SPEAKER:
[14:18:43] <scribe> Tony, one of the things that struck me is an issue here, and
[14:18:46] <scribe> we're talking about different solutions over different time
[14:18:49] <scribe> frames. I have no issue with that. But there are also different
[14:18:53] <scribe> solutions within the same time frame, that
[14:18:56] <scribe> we've got this perception at least in the presentation here,
[14:18:59] <scribe> that one tool solves all of the prob
[14:19:03] <scribe> lements lems that are on the list. We can't split the list of
[14:19:07] <scribe> requirements and say, well, things like tool X solve this set of
[14:19:11] <scribe> problems and tool Y solves this other set of problems and we
[14:19:13] <scribe> can use them together. You know, prime example, you
[14:19:18] <scribe> know, in the spirit of Brian saying that nat is a locater
[14:19:23] <scribe> split, well, so is N P L S. We have the I D in the IP dpres, we have the
[14:19:28] <scribe> locater in the N P L S locater. Different set of tools
[14:19:34] <scribe> solving a different part of the space. We're trying to, you
[14:19:38] <scribe> know, one tool solving all the problems. Operator problems
[14:19:41] <scribe> are very different from application, security, end user
[14:19:45] <scribe> problems. And they may be mutually exclusive. I'm not
[14:19:45] <scribe> saying they are.
[14:19:49] <scribe> That's part of what has to happen here of we have to decide which
[14:19:53] <scribe> tools decide which set of problems. And trying to solve
[14:19:56] <scribe> everything with V GP or its replacement may not be the right
[14:19:57] <scribe> answer. NEW SPEAKER:
[14:20:01] <scribe> I agree and actually, attempt to do this divide and conquer.
[14:20:04] <scribe> So moving on to the next question over there. NEW SPEAKER:
[14:20:10] <scribe> ted seally. Somebody has to buy what you're ped telling.
[14:20:15] <scribe> A couple comments quick. Panic is not the
[14:20:19] <scribe> right word. That was awake up call. In the I S G, hit the
[14:20:24] <scribe> snooze button a few times to many. I think technical
[14:20:28] <scribe> solution slide two had some is verb badge regarding SIP design
[14:20:37] <scribe> changes coming in the future, from 2 nano meter I believe.
[14:20:40] <scribe> And further down was a decision about business issues and Brian
[14:20:42] <scribe> made the comment that those are outside the scope of the IETF.
[14:20:47] <scribe> I know ta typically the IETF doesn't worry about economics,
[14:20:52] <scribe> but I would strongly disagree with Brian's comment. If our
[14:20:53] --- jpope has joined
[14:20:55] <scribe> time scale for upgrade is two years, then the economic
[14:20:59] <scribe> reality of what this thing costs is real. And Mr. Hardie's
[14:21:02] <scribe> comments were kind of spot on there as well. You're
[14:21:03] <scribe> going to break it.
[14:21:07] <scribe> This system is very fragile from an economic perspective of the you
[14:21:11] <scribe> have no idea. If you expect every operator to buy your new chip
[14:21:15] <scribe> set every two years to support this thing, you are gravely mistaken.
[14:21:16] <scribe> ( Applause.)
[14:21:25] <scribe> NEW SPEAKER: I pretty much want to agree with ted, he
[14:21:29] <scribe> stole a little bit of my thunder. NEW SPEAKER: Sorry. NEW SPEAKER:
[14:21:34] <scribe> The reality here is that I'm not panicked at p point but I'm
[14:21:36] <scribe> extremely concerned and it's not only from the economic
[14:21:40] <scribe> perspective but it's also from a process perspective. It's
[14:21:43] <scribe> going to take us time to figure out what is the, what is
[14:21:46] <scribe> going to be a right solution, design it, experiment it,
[14:21:51] <scribe> go back, refine it, design it, refine it and experiment
[14:21:56] <scribe> with it. Look how long it's taken to get v6 out the door.
[14:22:00] <scribe> It's still not even fully deployed in anybody's imagination.
[14:22:04] <scribe> But the fact is, it's going to take us several years if not
[14:22:08] <scribe> decades to get out there. I mean, it's going to take us several
[14:22:11] <scribe> years to buy the equipment, deploy the equipment, train end
[14:22:15] <scribe> users and train application writers how they're going to do this
[14:22:19] <scribe> stuff. So we can't keep hitting the snooze button. We
[14:22:23] <scribe> need to be bowled about the statements we're making. One of the points that was
[14:22:28] <scribe> on the slides that Brian shared was that fib size is not the
[14:22:32] <scribe> issue. That's an issue for micro electronic engineers to go
[14:22:37] <scribe> fix. That's absolutely wrong. Fib size the if fundamental
[14:22:37] <scribe> issue that we need to solve for.
[14:22:42] <scribe> Now, we're going to solve that probably by addressing multi
[14:22:46] <scribe> honing traffic engineering but those are direct inputs to the
[14:22:50] <scribe> hardware engineers that design the chips and then consume power
[14:22:53] <scribe> and cooling inside of my Gateways deploying the
[14:22:58] <scribe> equipment. I don't need any xhens but we need to get bold and take
[14:23:04] <scribe> action and moving forward. NEW SPEAKER: Next. NEW SPEAKER:
[14:23:12] <scribe> One thing looks like I'm very much reflecting what ted said,
[14:23:19] <scribe> but, well, okay. I'm quite happy about the presentation
[14:23:24] <scribe> given. And I totally agree panic would be bad. But
[14:23:30] <scribe> keeping a sense of urgency and seriously focus on creating a
[14:23:34] <scribe> solution that has not started as a work full activity quite
[14:23:36] <scribe> certainly is needed.
[14:23:44] <scribe> The point that the previous speaker and ted already made, and
[14:23:47] <scribe> in kind of adjusting some of the limitations that Brian pointed
[14:23:54] <scribe> out in my wording is that, well, if the architecture requires
[14:23:58] <scribe> leading edge micro electronic for scaling,
[14:24:02] <scribe> this implies risks and high costs and in particular,
[14:24:05] <scribe> potentially very short lifetime of the hardware. That will
[14:24:11] <scribe> go into costs or potential costs, that the operators and the
[14:24:15] <scribe> users out there quite certainly are not willing and probably
[14:24:20] --- yangwoo has left
[14:24:20] <scribe> not able to pay. So, that's something that we have to keep
[14:24:28] <scribe> in mind, even if it's based on areas of expertise that are not our
[14:24:32] <scribe> core issues, and are actually resolved somewhere else.
[14:24:38] <scribe> Note that I started from, what are the architectural
[14:24:43] <scribe> requirements that are to be mapped to something? And we have
[14:24:50] <scribe> to make sure or we have to create some confidence that, well,
[14:24:54] <scribe> okay, we deal with the architecture in ways so that
[14:25:02] <scribe> this, this danger of triggering a path into high
[14:25:06] <scribe> costs and short lifetimes. And high risks, is
[14:25:11] <scribe> actually avoided from the architectural side. And we also have
[14:25:16] <scribe> to consider and control the complexity that is injected from other
[14:25:25] <scribe> functional dimensions. And I've seen that quite obvious.
[14:25:33] <scribe> And actually, preventing complexity from any
[14:25:38] <scribe> dimension that requires different futures is
[14:25:40] <scribe> obviously something different than we've seen over the last
[14:25:45] <scribe> years. NEW SPEAKER: Thank you. I read that urgency is
[14:25:50] <scribe> needed and arc tech tur requirements are going to be discussed
[14:26:04] <scribe> in inter area. NEW SPEAKER: Essentially, I'm afraid
[14:26:06] <scribe> queer going into a fishing expedition. One thing
[14:26:11] <scribe> about government, is that success to good government, not so
[14:26:15] <scribe> much as compared to wealth, but to spread the faith. And I
[14:26:19] <scribe> would like to mention that, keep this in mind when you are sging
[14:26:21] <scribe> to design the solution, to make sure that people who are going
[14:26:24] <scribe> to benefit from the solution are going to be the one who pay
[14:26:30] <scribe> for it. The people who don't support it have any cost.
[14:26:33] <scribe> Because cost is one thing, but those are not the only costs that we
[14:26:37] <scribe> have to suffer as an operator. That's the No. 1 point that I
[14:26:38] <scribe> really wanted to make.
[14:26:43] <scribe> And No. 2 point is, what I've got from a presentation earlier
[14:26:51] <scribe> today is that essentially you're looking at solution space
[14:26:56] <scribe> into, routing and there's other side you're going
[14:27:00] <scribe> to do a huge change going from, going to something. And
[14:27:04] <scribe> I'm a little bit worried about it. NEW SPEAKER: Inter area is
[14:27:09] <scribe> looking at inter locater split, but with as much input as we can
[14:27:12] <scribe> possibly get and we have invited experts on this from the
[14:27:16] <scribe> routing area. So it's just, you know, time management as
[14:27:19] <scribe> much as anything else. You see all of us up here together
[14:27:25] <scribe> because it is, a coordinate the ted problem between two of us
[14:27:35] <scribe> as well as the rest of the IETF. NEW SPEAKER: So, sue
[14:27:43] <scribe> Harris. I'm an internet citizen. So in 20 years with GP cc I've
[14:27:43] --- hardie@jabber.qualcomm.com has joined
[14:27:48] <scribe> seen a lot of work. You know, a lot of research, a lot of
[14:27:53] <scribe> thoughts, a lot of things. Brian probably gave one-tenth of
[14:27:54] --- shep has joined
[14:27:59] <scribe> what I have seen as far as changes. And ideas. And to
[14:28:03] <scribe> make efficient forward progress, I hope that we will be very
[14:28:07] <scribe> careful, because, you know, I'm
[14:28:11] <scribe> probably as passionate as anyone who has seen better and better V
[14:28:14] --- sconte has joined
[14:28:15] <scribe> GP or better and better inter domain routing, but we hit several
[14:28:19] --- gordon.lennox has joined
[14:28:20] <scribe> lumps, and not looking at the lumps, in fact, I recommend
[14:28:24] <scribe> listening to John's talk tomorrow about some
[14:28:27] <scribe> miscontraceptions versus non miscontraceptions in the routing
[14:28:30] <scribe> area. And with V GP.
[14:28:34] <scribe> You must, you know, what it says is those who do not
[14:28:37] <scribe> remember history are doomd to repeat their mistakes.
[14:28:40] <scribe> If we're going after architecture, if we're going after all of
[14:28:44] <scribe> this, I really hope we're going to MIME, you know,
[14:28:48] <scribe> there's the phrase called data mining. There's a lot of data to
[14:28:51] <scribe> mine. I hope we get in there and mine it real quick so we
[14:28:53] <scribe> can get on to the next stage. Thank you.
[14:29:00] <scribe> NEW SPEAKER: I'm maring get was ser man. I agree with the approach
[14:29:05] <scribe> to look at shorter term solutions and longer term solutions.
[14:29:11] <scribe> I would like the message that he said along the lines, where there
[14:29:11] --- mzoo has joined
[14:29:13] <scribe> were more than just two possible time frames.
[14:29:17] <scribe> I think that in the shorter term solution space it makes sense
[14:29:20] <scribe> that we're looking at the routing area and the internet area
[14:29:23] <scribe> and we see you guys up there. We see a director the rat that
[14:29:28] <scribe> is as far as I know, populated with IP and routing people.
[14:29:33] <scribe> But in the longer term solution, when we're talking about the
[14:29:37] <scribe> locator split and application area, we're talking about security
[14:29:39] <scribe> area, we should be talking about the management area, which
[14:29:43] <scribe> hasn't come up yet. I'm not sure there's an area that
[14:29:46] <scribe> doesn't need to be involved, because certainly the transport area
[14:29:50] <scribe> needs to be involved. And I think I'd like to see, I understand
[14:29:53] <scribe> that you're saying you're trying to involve those people, but
[14:29:57] <scribe> I think I would like to see a little bit more wood behind that
[14:30:00] <scribe> arrow, like some people from those areas being represented in
[14:30:02] <scribe> some of the leadership of this process as you expand that
[14:30:06] <scribe> directorate, which I guess is really a message for some other
[14:30:07] <scribe> chairs somewhere.
[14:30:14] <scribe> The other point that I would like to make is that one of the
[14:30:17] <scribe> things I had learned over the past several years, trying to
[14:30:22] <scribe> work on some changes at the IP level, is that you know,
[14:30:25] <scribe> there's a great feeling of don't touch that because you don't know
[14:30:30] <scribe> what it's attached to. Okay. And we have this concern,
[14:30:32] <scribe> this conservatism about changing that areas, because
[14:30:36] <scribe> everything you try to change, some application depended on
[14:30:39] <scribe> it. Some other mek dmis many depends on it. This won't
[14:30:40] <scribe> work, that wonlt work.
[14:30:44] <scribe> It's kind of frustrating, that's what you get for being in the
[14:30:48] <scribe> middle. But I would suggest that the first step in making an
[14:30:51] <scribe> architectural change at the IP layer is to understand what an
[14:30:52] <scribe> architecture is today.
[14:30:57] <scribe> We have it documented what the IP service model is, that
[14:31:01] <scribe> applications expect from IP, so we don't know what it is that
[14:31:05] <scribe> we can't change without higher layer implications. We don't
[14:31:08] <scribe> really have a good understanding of what the IP architecture per
[14:31:11] <scribe> se is. I think some people do have an understanding of
[14:31:14] <scribe> that, but it's knots a common understanding that's been document
[14:31:18] <scribe> Ed in the organization. There's contention about it. And
[14:31:21] <scribe> I think that's something that we have to nail down, before it
[14:31:24] <scribe> really makes sense to talk about whether we can make an IP
[14:31:28] <scribe> locater split mechanism that beats the model. Because we
[14:31:30] <scribe> don't know what that means. There was a discussion
[14:31:33] <scribe> about attitude, can't applications just pass identifiers
[14:31:37] <scribe> around. The problem is you you can't send a packet through an identifier.
[14:31:41] <scribe> So somebody somewhere has to map that to a locater. Okay. We know there
[14:31:46] <scribe> are problems with the fact that the IP address is IV and the IP
[14:31:50] <scribe> address is locater and it gets misused. And layers that
[14:31:52] <scribe> shouldn't be talking about IPs are talking about them. The
[14:31:56] <scribe> thing is, there's also a huge feature in that, which is the
[14:32:00] <scribe> mapping for IP locate stead, it is free, it's the identity. You know, it's
[14:32:02] <scribe> the, they are the same. Okay.
[14:32:06] <scribe> And without understanding what the service model is of who is
[14:32:10] <scribe> using what for what, talking about splitting them is in my mind
[14:32:14] <scribe> premature, because we continually run into the same set of
[14:32:17] <scribe> problems here. And I, I'm not positive that we're not just
[14:32:20] <scribe> moving the problem. Okay. Because I'm not sure we're not
[14:32:24] <scribe> taking the problem out of the core routing, and moving it to
[14:32:29] <scribe> some theoretical I G to locater mapping mechanism that if we
[14:32:33] <scribe> ever manage to actually define and create and scale it away,
[14:32:37] <scribe> might not be able to slip right in there and solve the routing problem
[14:32:42] <scribe> too. Because I don't think we know exactly what we're going
[14:32:45] <scribe> for here. We need to understand that better before we rush
[14:32:54] <scribe> ahead. NEW SPEAKER: Dave cok err. I'd like to thank
[14:32:59] <scribe> you. First, for generating one of the more animated open Mike
[14:33:03] <scribe> sessions that we've had in a while. And the second is for prompting me
[14:33:10] <scribe> into a much more familiar role. Of complaining about lack of
[14:33:11] <scribe> productivity.
[14:33:15] <scribe> I don't think folks heard Dave Meyer adequately. I remember the
[14:33:19] <scribe> discussions about these topics going back fifteen years. I remember
[14:33:24] <scribe> them going back very animatedly 5 years. And things that you are saying
[14:33:26] --- sconte has left
[14:33:27] <scribe> about the status of how much work we have to do to understand
[14:33:31] <scribe> the problem has been said many times before.
[14:33:35] <scribe> The list that Brian put up of the scope of topics to try to deal
[14:33:40] <scribe> with, is too damn large. And as long as it stays
[14:33:44] <scribe> that large, it ain't going to happen. So my suggestions is
[14:33:52] <scribe> find the smallest set that routing has to solve and punt on the
[14:33:57] <scribe> rest. This is either in, complete agreement with Margaret
[14:34:00] <scribe> just said or please completely opposite. I'm not quite sure.
[14:34:05] <scribe> From what I can tell, the routing table size and flow are the
[14:34:10] <scribe> two things that have to be done in or routing, and the rest can be
[14:34:14] <scribe> puntd. Even though you probably don't think so. NEW SPEAKER:
[14:34:21] <scribe> Thanks. Next question over there. marl la. Frontier
[14:34:24] <scribe> communications, I'm actually sent hereby the American
[14:34:29] <scribe> registry for internet numbers. I'm on a advising council and
[14:34:33] <scribe> I'm the chair there. The reason they started sending me here three
[14:34:34] --- admcd has left
[14:34:37] <scribe> IETFs s ago is because of IP v6 issue. The issue
[14:34:40] <scribe> being multi homing and traffic engineering. I have
[14:34:42] <scribe> to say, it's been a little discouraging
[14:34:46] <scribe> because I do hear a lot of contributing and bik
[14:34:49] <scribe> customering that honestly does and need to happen. And there
[14:34:54] <scribe> are network providers that have customers lined up to do that with
[14:34:58] <scribe> v6 and we need to do traffic engineering. It's proprietary
[14:35:01] <scribe> information, so I'm sorry those people have numbers for absolutely
[14:35:05] <scribe> everything. It can't happen. But there are providers that
[14:35:08] <scribe> will tell you, we need it. We are going to do it.
[14:35:12] <scribe> And we came to the IETF for support. And I would love to have
[14:35:18] <scribe> faith. In the IETF, that they will concentrate and focus the
[14:35:21] <scribe> meetings more and invite people that are thereto actually help
[14:35:26] <scribe> create the solution, and not debate whether or not it needs to be
[14:35:30] <scribe> looked at. Because I saw that yet again this time. And if
[14:35:34] <scribe> you don't think it's needed, just don't show up. It's, there
[14:35:38] <scribe> are people that belief they need it and they want to work on
[14:35:42] <scribe> it. And that's where we need to put the focus. If it's a
[14:35:45] <scribe> distract ter, it would be great if I could go back to the
[14:35:51] <scribe> region and say we have a focus group of peep that people that understand the issues
[14:35:57] <scribe> and that's who is showing up and talking. But I still see
[14:36:02] <scribe> quawb link, and you can have your pink, it's great that
[14:36:05] <scribe> you opinions. I would love to just see the people that understand
[14:36:10] <scribe> the issue, or care about the issue to actually work on it.
[14:36:14] <scribe> And it's probably a waste of other people's time if you don't
[14:36:21] <scribe> think it is an issue. NEW SPEAKER: Last comment. NEW SPEAKER:
[14:36:28] <scribe> I'm Bob. So, one thing I want to adhere, is that,
[14:36:31] <scribe> well this is a pretty hard problem and I am quite happy we're starting to talk
[14:36:36] <scribe> about it again. But my guess is, if we're going solve it
[14:36:41] <scribe> in a a long term fundamental way, it can't be done and
[14:36:43] <scribe> preserve all the current business models of all the current
[14:36:48] <scribe> players. I think we have to be real careful to not, to not
[14:36:51] <scribe> go down paths, because someone says, well, that brakes
[14:36:55] <scribe> my business model. And I'm not trying to talk about any
[14:36:59] <scribe> particular one here. We all are mostly here from people,
[14:37:00] <scribe> companies who have business models.
[14:37:06] <scribe> But, you know, the internet got started by sort of breaking the
[14:37:09] <scribe> existing player's business model completely, and they're still not
[14:37:12] --- AWGY has left
[14:37:14] <scribe> very happy about it. But, so, my guess is, and well,
[14:37:16] <scribe> I, we won't talk about that now.
[14:37:20] <scribe> But, so I think it's real important if we're going to work,
[14:37:25] <scribe> solve this long term problem that we be willing to investigate
[14:37:30] <scribe> directions that might require people to do business, you
[14:37:34] <scribe> know, the money part of this, the business model part, in a
[14:37:37] <scribe> different way. It might create new opportunities for new
[14:37:41] <scribe> players or not. But we shouldn't just be constrained by the
[14:37:50] <scribe> way people do it today. NEW SPEAKER: Well, so we're just a few
[14:37:52] <scribe> minutes over. Thank you, everybody for all the
[14:37:57] <scribe> comments, and talk is cheap, but I think I can say, on behalf
[14:37:59] --- mzoo has left: Computer went to sleep
[14:38:00] <scribe> of all of us here, that we understand this problem is very
[14:38:05] <scribe> urgent, and please show up at our inter area and routing area
[14:38:07] --- weiler has left
[14:38:07] <scribe> meetings, that's the time we have scheduled
[14:38:10] --- Stephen Farrell has left
[14:38:10] <scribe> to talk about it here. And then of course back on e-mail and
[14:38:16] <scribe> everything else, and enjoy your time out tonight. Thanks.
[14:38:17] --- scribe has left
[14:38:19] --- csp has left
[14:38:20] --- hardie@jabber.qualcomm.com has left
[14:38:34] --- JeffH has left
[14:38:34] --- gordon.lennox has left
[14:38:48] --- gih900 has left
[14:38:49] --- resnick has left
[14:38:52] --- renee has left
[14:38:57] --- miaofuyou has left
[14:39:08] --- DonFedyk has left
[14:39:15] --- Leslie has left: Logged out
[14:40:04] --- tonyhansen has left
[14:41:19] --- bhoeneis has left: plenary@IETF-68@Praha
[14:41:50] --- mrex has left
[14:43:10] --- jlcjohn has left
[14:44:24] --- shep has left
[14:44:50] --- tlyu has left
[14:47:45] --- jpope has left
[14:48:51] --- dthaler has left: Replaced by new connection
[14:50:54] --- jclee has left
[14:52:19] --- pesherb has left
[14:54:10] --- dthaler has joined
[14:54:23] --- renee has joined
[14:57:11] --- renee has left: Replaced by new connection
[17:53:33] --- guido.franceschini has joined
[18:00:57] --- guido.franceschini has left
[18:10:35] --- dthaler has left