IETF
manet
manet@jabber.ietf.org
Thursday, November 13, 2014< ^ >
bebemaster has set the subject to: End
Room Configuration
Room Occupants

GMT+0
[18:10:46] stanratliff joins the room
[18:11:19] stanratliff leaves the room
[18:21:53] Ulrich Herberg joins the room
[18:37:39] hrogge joins the room
[18:37:57] <hrogge> Good evening from Germany!
[18:46:12] <Ulrich Herberg> good morning :-)
[18:46:39] <Ulrich Herberg> from California
[18:48:12] <hrogge> no Honolulu for you too?
[18:48:21] <Ulrich Herberg> no :-(
[18:48:45] <hrogge> damned
[18:48:59] <hrogge> first IETF where we get 2.5 hours and we are both not there
[18:49:19] <hrogge> okay, first IETF in years with 2.5 hours for MANET ^^
[18:51:01] <hrogge> I hope the audio stream will work well
[18:52:38] John Dowdell joins the room
[18:53:24] <John Dowdell> Good good morning all, I will be your scribe today. Anyone here?
[18:53:41] <hrogge> Hi John... greetings from Germany
[18:54:29] <John Dowdell> Hi Henning, good evening to you!
[18:54:40] hrogge is trying to change his alias... brb
[18:54:42] hrogge leaves the room
[18:54:45] hrogge joins the room
[18:54:56] <hrogge> *sigh* no, doesn't work
[18:55:17] <Ulrich Herberg> hi John. I am also here
[18:56:04] hrogge leaves the room
[18:56:06] Henning Rogge joins the room
[18:56:17] <Henning Rogge> okay, that is better
[18:57:07] stanratliff joins the room
[18:57:18] <Henning Rogge> Hi Stan, good to see you!
[18:57:25] <John Dowdell> Hi Ulrich
[18:57:26] <stanratliff> Hi Henning!
[18:57:42] Henning Rogge leaves the room: Disconnected: closed
[18:57:47] hrogge joins the room
[18:57:48] <Ulrich Herberg> is the audio working?
[18:57:55] <John Dowdell> Good afternoon Stan
[18:58:07] <stanratliff> I don't have audio, but that might be my system (damned Ubuntu)... ;-)
[18:58:17] <stanratliff> Hello John
[18:58:21] <John Dowdell> Audio should be working, Justin says
[18:58:31] <Ulrich Herberg> I get a connection error
[18:59:44] <stanratliff> I get Rhythmbox.... Don't want to listen to tunes right now... ;-)
[18:59:48] <John Dowdell> We are working on the audio
[18:59:57] Axel Colin de Verdière joins the room
[19:00:36] <hrogge> I tried to load the m3u file with Amarok... and got an error
[19:00:50] <John Dowdell> Henning do you have audio please?
[19:01:38] adrianfarrel joins the room
[19:01:41] <Ulrich Herberg> now it seems to work
[19:01:50] <adrianfarrel> for all of you?
[19:02:16] <stanratliff> I've got audio
[19:02:24] <John Dowdell> Just starting the meeting
[19:02:41] <Ulrich Herberg> had to use my cell phone... seems like my company blocks the port ;-)
[19:03:09] badamson joins the room
[19:06:06] <stanratliff> I hear Justin, then someone waaaayyy in the background. Please have them step to the mic.
[19:06:41] <John Dowdell> You should be able to hear Thomas presenting now
[19:06:43] <hrogge> I am still looking for the Fedora packet to decode MP3
[19:09:10] Lorenzo Miniero joins the room
[19:09:38] <hrogge> great, it seems Redhat removed MP3 support from all packages for Fedora
[19:09:38] <hrogge> -.-
[19:09:54] Axel Colin de Verdière leaves the room
[19:11:33] <John Dowdell> Now discussing nhdp optimisation
[19:13:02] <Ulrich Herberg> maybe use Ubuntu instead? or a real OS, such as Mac OS? :-p
[19:13:43] <John Dowdell> now discussing rfc6779bis
[19:13:48] <hrogge> Ubuntu left me without a (available) fix for the graphics drivers for 6 months
[19:13:53] <Lorenzo Miniero> rpmfusion is the way ;) (fedora/redhat)
[19:13:58] Axel Colin de Verdière joins the room
[19:14:40] Meetecho joins the room
[19:15:08] <hrogge> I am working on it
[19:15:51] <John Dowdell> discussing olsrv2 management snapshot
[19:16:02] <hrogge> got it
[19:19:09] <John Dowdell> discussing olsrv2 multitopology
[19:19:22] akatlas joins the room
[19:22:34] <John Dowdell> discussing tlv naming
[19:24:25] akatlas joins the room
[19:24:37] <John Dowdell> anyone using meetecho? the techs seem to be troubleshooting
[19:24:52] <stanratliff> I tried, it's not working.
[19:24:59] akatlas leaves the room
[19:25:08] <hrogge> works for me now... just by using the m3u file
[19:25:43] <hrogge> ahh sorry, have not seen the meetecho comment
[19:26:04] akatlas leaves the room
[19:26:25] <John Dowdell> Rick Taylor stepping to the mic, discussing DLEP
[19:28:12] akatlas joins the room
[19:30:42] Morten V. Pedersen joins the room
[19:30:59] <hrogge> @Room: We also talked about an extension draft/rfc for additional standardized metric types
[19:31:43] Morten V. Pedersen leaves the room
[19:31:49] <adrianfarrel> @hrogge: is that for the mic?
[19:31:55] <hrogge> yes
[19:32:50] Lorenzo Miniero leaves the room
[19:32:50] <stanratliff> For the mic - Yes, I think the extension draft is the right way to go moving forward.
[19:33:52] <stanratliff> For now, the MCAST discovery is 1-hop only. Authentication is based on the transport - e.g., TLS, DTLS.
[19:35:09] <stanratliff> Actually, the auth would be TLS, since we're now on TCP for the session.
[19:37:56] <stanratliff> There's a difference between *stopping* a vendor from doing something stupid, and *helping* them do something stupid...
[19:38:13] <hrogge> For the mic: that is also why the extension drafts/rfcs are important... you can put "you must be RFC xyz compliant" into a requirement for your radios
[19:38:28] <stanratliff> Agree with Henning
[19:39:07] <stanratliff> For the mic - the default for vendor extensions is what is documented in DLEP-07.
[19:39:50] <stanratliff> If we get no input from the WG be Dec 19, we'll go with what is in DLEP-07
[19:40:06] <stanratliff> *by* dec 19...
[19:41:41] <stanratliff> To Thomas - then please make some suggestions on the list. ;-)
[19:42:14] <stanratliff> I believe we're ready for editorial revisions now.
[19:44:03] <stanratliff> Yes, please, Thomas.
[19:44:26] <stanratliff> We're not going to re-write. It's just the correct insertion points for extensions, and we're done (technically).
[19:44:32] <stanratliff> Name?
[19:44:42] <John Dowdell> James Gettys
[19:45:34] <John Dowdell> Charlie Perkins at the mic, AODVv2
[19:49:46] james.huy.nguyen joins the room
[19:51:11] <hrogge> For the mic - it is even unnecessary to specify what kind of RFC5444 index should be generated, because a RFC5444 compliant protocol should be able to understand all of them anyways
[19:53:38] <hrogge> NO
[19:55:35] <stanratliff> I for one would almost like to see a specification of "north-bound" (e.g. client-based) formats into a 5444 parser.
[19:56:07] <stanratliff> Henning? Your thoughts on that?
[19:56:20] <John Dowdell> @Stan, completely agree
[19:56:32] <hrogge> not sure I know what "north-bound" formars mean ^^
[19:57:43] <hrogge> I could in theory give a presentation at a future IETF how my parser/generator API is working... what the clients have to deal with and what the library below does for them
[19:57:47] <stanratliff> The 5444 spec pretty much implies that there is a stand-alone, generic packet parser. 5444 itself describes the format of the packets on the wire. But if you're going to treat the 5444 parser as a stand-alone entity, then let's define the inputs to it - that's really the important part.
[19:58:00] <John Dowdell> What I was asking at the mic regarding what a client protocol is allowed to specifiy, and what is filled in by the parser
[19:58:51] <stanratliff> As an application (e.g. AODVv2), what do I have to hand to the stand-alone parser? What flag setting should i care about? What flag settings should I be able to expect? Where SHOULD I care?
[19:58:59] <stanratliff> That's what I mean by a "north-bound" API.
[19:59:23] <hrogge> could be a description of the API you use to talk with your parser/generator
[19:59:29] <stanratliff> Yes!
[19:59:56] <hrogge> I have a C API for this done... including multiplexing/demultiplexing
[20:00:18] <stanratliff> Bingo! Write that up as an informational RFC, and I'm really, really happy. :-)
[20:00:35] <John Dowdell> Me too
[20:01:15] <stanratliff> How are we doing on time here? Do we need to move along?
[20:01:42] <hrogge> we are 1 minute over the time, so we are good ^^
[20:01:49] <hrogge> 5+5+15+15+20 = 60
[20:02:06] <stanratliff> Yes, 1 minute over, and growing by the second... ;-)
[20:04:04] <hrogge> :)
[20:08:29] <hrogge> @stan: 500 seconds and counting
[20:09:27] <stanratliff> No, I asked a question - Did the working group want to proceed on a standards-track RFC with NO implementations?
[20:10:10] <stanratliff> Thomas over-states my position, IMO.
[20:10:32] <adrianfarrel> I'm holding that comment for my preso at the end of the meeting
[20:10:39] <John Dowdell> Jiazi Yi at the mic, multipath extensions for OLSRv2
[20:21:17] <stanratliff> Let's take that offline and keep moving on.
[20:23:35] <stanratliff> For the mic, please - We're over time, let's move on.
[20:24:37] <John Dowdell> @Stan just shared with Justin
[20:26:29] <John Dowdell> Security Threats of SMF is next up
[20:26:52] james.huy.nguyen leaves the room
[20:27:01] <stanratliff> We're almost 1 full presentation over-time.
[20:27:51] <John Dowdell> Not sure Justin is on jabber
[20:27:56] <adrianfarrel> Good job I can talk fast :-)
[20:28:07] <hrogge> ohoh
[20:28:20] <hrogge> "fast talk" about the future of the WG ^^
[20:29:34] <John Dowdell> OLSRv2 security threats
[20:29:41] james.huy.nguyen joins the room
[20:29:41] <adrianfarrel> You know it makes sense!
[20:36:31] Ulrich Herberg leaves the room
[20:36:51] <John Dowdell> Next up ODMRP, Thomas Clausen at the mic
[20:38:19] akatlas leaves the room
[20:39:46] <Axel Colin de Verdière> Looking forward to your review, thanks
[20:44:36] <John Dowdell> Next up: Working Group next steps, Adrian Farrell
[20:50:57] <stanratliff> And we have a "if you don't tell us what to do with vendor extensions, we're going to do what's in DLEP-07" approach.
[20:58:54] <stanratliff> ...and then, the audio stream goes dead.
[20:59:01] <hrogge> good that we still have 30 minutes left (not sure where they came from ^^)
[20:59:07] <hrogge> damned
[20:59:16] <hrogge> someone threw the mike at Thomas
[20:59:27] <stanratliff> :-) That's pretty much what I'm thinking.
[20:59:31] <John Dowdell> really no audio?
[20:59:34] <hrogge> no audio
[20:59:37] <badamson> is it back
[20:59:42] <stanratliff> There it is!
[20:59:44] <hrogge> are there still people alive in the room?
[21:00:05] <John Dowdell> yes
[21:00:12] <badamson> Thomas says it's only a flesh wound
[21:00:20] <stanratliff> Darn.
[21:00:22] <stanratliff> ;-)
[21:00:24] <hrogge> sound is back
[21:02:05] <hrogge> for the mic: early DLEP?
[21:02:27] <stanratliff> What do you mean, Henning?
[21:02:35] <hrogge> protocol using RFC5444
[21:02:36] <John Dowdell> @henning in what way?
[21:02:47] <james.huy.nguyen> I thought they are talking about AODVv2
[21:02:49] <stanratliff> Yes, and we apparently screwed it up.
[21:03:11] <james.huy.nguyen> Adrian is talking
[21:03:12] <stanratliff> If you're trying to use early DLEP to show how easy 5444 is, that's a miserable failure.
[21:03:20] <james.huy.nguyen> 3 options for AODVv2
[21:07:57] Meetecho leaves the room
[21:07:57] Meetecho joins the room
[21:07:57] <stanratliff> And the difference in opinion as to whether the document is "usable" or "in really rough shape" is the crux of the concern over how optimistic the schedule is.
[21:09:12] <stanratliff> two. I had a devil of a time with 5444 and DLEP.
[21:10:11] <hrogge> @Stan: do you think an API description for a parser/generator would help you in the future with this (not necessarily with DLEP ^^)?
[21:10:21] <stanratliff> Yes. Absolutely.
[21:10:30] <stanratliff> And I think it would help LOTS of people.
[21:10:48] Jiazi YI joins the room
[21:10:52] <stanratliff> That's exactly the "north-bound interface" document I was talking about earlier.
[21:11:36] <hrogge> the funny thing is that the API of my code has been more stable than the code behind it... had to fix quite a few annoying and subtil things
[21:11:52] <stanratliff> Well, maybe that gets to the point. ;-)
[21:12:35] <hrogge> I implemented an automatic address compression thing in a very generic way... that got quite complex ;)
[21:13:11] <hrogge> but the API does not show this, so we are lucky in this case
[21:13:12] <stanratliff> and some of the really crazy-complex stuff is the address blocks, and how to manipulate them, etc...
[21:14:16] <stanratliff> Remember, there are some of us (and I think Charlie is included in that) are thinking about a routing protocol that *directly* emits 5444 packets - it does NOT necessarily go through a generic, stand-alone parser.
[21:15:21] <hrogge> if you don't need multiplexing of independent code parts and no address compression beyond "known network prefix" thats not that difficult to do ^^
[21:15:47] <stanratliff> Trust me. From my DLEP experience - that *IS* really difficult to do... ;-)
[21:16:03] <hrogge> maybe I worked to much with RFC5444 to appreciate this ;)
[21:16:05] <stanratliff> ... to directly emit "proper" 5444 packets, that is.
[21:19:21] <stanratliff> *THIS IS FOR THE MIC* If we're going to do the report MIB, we need the same type of schedule that DLEP and AODVv2 were required to provide. And it needs to be enforced.
[21:21:29] <stanratliff> So let it be done.... ;-)
[21:26:36] <hrogge> For the mic: doing IPv6 source-specific routing in OLSRv2 should not be that difficult
[21:26:47] stanratliff leaves the room
[21:27:31] <John Dowdell> That's all folks!
[21:27:43] <hrogge> sounds like everyone survived...
[21:27:48] <james.huy.nguyen> :)
[21:27:59] james.huy.nguyen leaves the room
[21:28:08] adrianfarrel leaves the room
[21:28:18] John Dowdell leaves the room
[21:28:20] hrogge leaves the room
[21:28:21] Axel Colin de Verdière leaves the room
[21:32:35] Jiazi YI leaves the room
[21:32:54] badamson leaves the room
[21:34:04] badamson joins the room
[21:36:05] Meetecho leaves the room
[21:39:29] badamson leaves the room
[22:02:59] adrianfarrel joins the room
[22:55:09] adrianfarrel leaves the room
[22:59:28] badamson joins the room
[23:00:18] adrianfarrel joins the room
[23:02:18] adrianfarrel leaves the room
[23:11:18] badamson leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!