iccrg - Monday, 30 July 2012

iccrg

iccrg@jabber.ietf.org

Monday, 30 July 2012< ^ >

jishac has set the subject to: Internet Congestion Control Research Group

Room Configuration

GMT+0
[12:31:20] Michael Tüxen joins the room
[19:08:14] Michael Tüxen leaves the room
[19:08:28] Michael Tüxen joins the room
[20:00:43] Ingemar Johansson joins the room
[20:00:56] <Ingemar Johansson> Is audio up yet ?
[20:03:28] <Michael Tüxen> I think so.
[20:04:03] <Ingemar Johansson> Ok, now it is working
[20:05:14] Andrew McGregor joins the room
[20:05:38] <Andrew McGregor> Introduction… we have a relaxed agenda
[20:06:03] wesley.m.eddy joins the room
[20:07:13] bob.briscoe joins the room
[20:07:13] <Andrew McGregor> Anyone remote getting presenter audio?
[20:07:19] <Michael Tüxen> yes
[20:07:26] cabo joins the room
[20:07:34] <Ingemar Johansson> Yes, good audio
[20:07:53] <Ingemar Johansson> Any slides for this presenation ?
[20:09:05] <Andrew McGregor> I'll look...
[20:09:46] Zahed joins the room
[20:10:03] <Andrew McGregor> Yuchung Cheng: "The GreatSnipe TCP congestion manager" <http://www.ietf.org/proceedings/84/slides/slides-84-iccrg-3.pdf>
[20:10:19] <Andrew McGregor> Now on slide 'A smart (or not-so-dumb) transport should...
[20:10:22] <Andrew McGregor> '
[20:10:35] <Ingemar Johansson> Thanks!
[20:11:04] <Andrew McGregor> Sender-side CM approach
[20:13:35] <Andrew McGregor> Can the client help?
[20:15:46] <Andrew McGregor> Great Snipe
[20:15:47] YJS joins the room
[20:17:54] kencarlberg joins the room
[20:18:07] <Andrew McGregor> Client-based congestion control
[20:18:12] <Andrew McGregor> Implementing standard AIMD
[20:21:11] <Andrew McGregor> Research issues / opportunities
[20:22:19] Ole Troan joins the room
[20:24:49] <Andrew McGregor> Conclusion
[20:25:46] <Andrew McGregor> Questions?
[20:26:34] <Andrew McGregor> next presentation's slides at http://www.ietf.org/proceedings/84/slides/slides-84-iccrg-2.pdf
[20:30:29] Zahed leaves the room
[20:35:29] <Andrew McGregor> Why here, why now?
[20:35:30] bob.briscoe leaves the room
[20:35:47] ena2413 joins the room
[20:37:04] <Andrew McGregor> Background: “newtcp” project at CAIA
[20:37:33] <Andrew McGregor> Why delay-based TCP?
[20:38:31] <Andrew McGregor> Induced queuing delay – “Collateral damage”
[20:39:20] <Andrew McGregor> Our first foray – “delay threshold” TCP
[20:40:05] <Andrew McGregor> Delay-gradient TCP: Less configuration
[20:41:15] <Andrew McGregor> CDG induces far less queuing delay
[20:42:04] <Andrew McGregor> CDG tolerant of intrinsic packet losses
[20:42:57] <Andrew McGregor> Delay gradient as a congestion signal
[20:45:09] <Andrew McGregor> Inferring queue states
[20:46:22] <Andrew McGregor> Inferring queue states (2)
[20:46:47] <Andrew McGregor> Window Progression when no packet loss
[20:47:57] <Andrew McGregor> Probabilistic 'back-off'
[20:49:10] <Andrew McGregor> Competing with loss-based flows #1
[20:51:43] <Andrew McGregor> Competing with loss based flows #2
[20:52:10] <Andrew McGregor> An aside: our testbed
[20:52:41] <Andrew McGregor> CDG achieves a reasonable goodput
[20:54:05] <Andrew McGregor> CDG achieves better goodput with loss
[20:54:36] <Andrew McGregor> Sharing between CDG and NewReno
[20:55:42] <Andrew McGregor> Sharing between flows with different RTT
[20:56:51] <Andrew McGregor> Conclusions (regarding CDG)
[20:57:50] <Andrew McGregor> Next presentation's slides at http://www.ietf.org/proceedings/84/slides/slides-84-iccrg-0.pdf
[20:59:26] <Andrew McGregor> Oops, early on that. Part 3 of this one to come
[21:00:03] <kencarlberg> question: would per-packet load balancing also be problematic for CDG in possibly contributing to fluctuating RTT
[21:00:25] <Andrew McGregor> will channel
[21:00:39] PasiS joins the room
[21:01:31] ena2413 leaves the room
[21:02:18] ena2413 joins the room
[21:06:57] <Ingemar Johansson> LTE simulations also give the same observation, essentially very small difference between NewReno and Cubic in terms of goodput but Cubic cause considerably more congestion
[21:11:31] shep joins the room
[21:13:20] <Andrew McGregor> Lawrence, this is still http://www.ietf.org/proceedings/84/slides/slides-84-iccrg-2.pdf
[21:13:25] dtaht joins the room
[21:13:33] <Andrew McGregor> FreeBSD As A R&D Platform
[21:14:54] <Andrew McGregor> FreeBSD As A R&D Platform (2)
[21:15:49] <Andrew McGregor> Modular CC
[21:17:07] <Andrew McGregor> SIFTR
[21:17:48] <Andrew McGregor> Khelp/Hhook
[21:18:11] <Michael Tüxen> Is the NS-3 port of the FreeBSD stack limited to TCP or does it include also SCTP?
[21:18:26] <dtaht> so when will we see codel in bsd?
[21:19:03] <Andrew McGregor> Links
[21:20:57] <Andrew McGregor> Now http://www.ietf.org/proceedings/84/slides/slides-84-iccrg-0.pdf
[21:21:27] <Andrew McGregor> No NDA or GPL'd content
[21:21:55] <Andrew McGregor> Current status
[21:23:37] <Andrew McGregor> Why partition the patch?
[21:24:28] <Andrew McGregor> Possible Laminar Addons
[21:30:05] <Andrew McGregor> Planned new mailing list
[21:36:07] <Andrew McGregor> http://www.ietf.org/proceedings/84/slides/slides-84-iccrg-1.pdf
[21:36:37] <Andrew McGregor> Netalyzr
[21:37:35] <Andrew McGregor> Global research topics
[21:38:55] shep leaves the room
[21:39:04] shep joins the room
[21:40:42] <Andrew McGregor> Netflix streaming
[21:41:06] <Andrew McGregor> Algorithmic questions
[21:41:30] <Ingemar Johansson> The netflix behavior has the benefit however that it saves battery in wireless terminals
[21:44:23] <Andrew McGregor> Data Acquisition & Analysis
[21:46:32] <Andrew McGregor> Tools
[21:48:07] ena2413 leaves the room
[21:48:43] ena2413 joins the room
[21:50:21] <Andrew McGregor> Home Router Disaster
[21:53:03] <Andrew McGregor> In Disaster, There is Opportunity
[22:00:08] PasiS leaves the room
[22:00:15] Andrew McGregor leaves the room
[22:00:37] kencarlberg leaves the room
[22:01:52] wesley.m.eddy leaves the room
[22:02:12] ena2413 leaves the room
[22:03:18] dtaht leaves the room
[22:03:24] Ingemar Johansson leaves the room
[22:13:52] YJS leaves the room
[22:17:04] shep leaves the room
[22:22:47] ena2413 joins the room
[22:27:23] YJS joins the room
[22:28:52] cabo leaves the room
[22:29:22] Ole Troan leaves the room
[22:29:28] ena2413 leaves the room
[22:30:17] YJS leaves the room
[22:33:05] Michael Tüxen leaves the room
[22:35:18] Ole Troan joins the room
[22:35:28] Andrew McGregor joins the room
[22:35:59] Ole Troan leaves the room
[22:36:33] wesley.m.eddy joins the room
[22:37:05] dtaht joins the room
[22:37:22] Ole Troan joins the room
[22:38:28] dtaht leaves the room
[22:40:49] wesley.m.eddy leaves the room
[22:46:15] cabo joins the room
[22:46:15] dtaht joins the room
[22:47:16] wmtownsley joins the room
[23:21:04] Ole Troan leaves the room
[23:41:43] Ole Troan joins the room
[23:45:47] Andrew McGregor leaves the room
[23:47:41] Andrew McGregor joins the room
[23:53:06] wmtownsley leaves the room