Thursday, November 11, 2021< ^ >
meetecho-alexamirante has set the subject to: IETF 111
Room Configuration
Room Occupants

[15:34:15] Meetecho joins the room
[15:40:13] alexamirante joins the room
[15:43:25] alexamirante has set the subject to: IETF 112
[15:45:03] Nicolas Kuhn_web_371 joins the room
[15:45:03] Alessandro Amirante_web_456 joins the room
[15:45:03] Gorry Fairhurst_web_444 joins the room
[15:45:03] Ana Custura_web_370 joins the room
[15:45:03] Mohamed Boucadair_web_972 joins the room
[15:45:03] Spencer Dawkins_web_964 joins the room
[15:46:39] Jen Linkova_web_345 joins the room
[15:51:14] Theresa Enghardt_web_928 joins the room
[15:53:01] Carsten Bormann_web_600 joins the room
[15:54:02] John Border_web_384 joins the room
[15:55:07] Brian Trammell_web_431 joins the room
[15:56:04] Antoine Fressancourt_web_326 joins the room
[15:56:09] Bhavit Shah_web_223 joins the room
[15:56:12] Tianji Jiang_web_239 joins the room
[15:56:47] Cyrill Krähenbühl_web_104 joins the room
[15:56:54] Colin Perkins_web_865 joins the room
[15:57:21] Julien Maisonneuve_web_843 joins the room
[15:57:42] Erik Kline_web_889 joins the room
[15:58:01] Hirochika Asai_web_815 joins the room
[15:58:18] <Erik Kline_web_889> "the matrix"  :-)
[15:58:32] Mirja Kühlewind_web_660 joins the room
[15:58:59] Luigi Iannone_web_137 joins the room
[15:59:15] <Jen Linkova_web_345> It was the best description I can come up with ;)
[15:59:37] Kireeti Kompella_web_465 joins the room
[15:59:43] Adam Wiethuechter_web_978 joins the room
[16:00:05] Lixia Zhang_web_696 joins the room
[16:00:08] <Erik Kline_web_889> "Madrix"?
[16:00:19] <Jen Linkova_web_345> lol
[16:00:24] <Jen Linkova_web_345> That's good one
[16:00:37] <Antoine Fressancourt_web_326> @Erik lol
[16:01:13] Sabine Randriamasy_web_992 joins the room
[16:01:18] Kazuaki Ueda_web_942 joins the room
[16:01:58] Rayhaan Jaufeerally_web_631 joins the room
[16:02:27] Toerless Eckert_web_946 joins the room
[16:02:33] Roland Bless_web_697 joins the room
[16:02:37] Tom Jones_web_983 joins the room
[16:02:55] Mihail Zverev_web_347 joins the room
[16:03:01] Tom Jones_web_983 leaves the room
[16:03:05] Tom Jones_web_273 joins the room
[16:03:05] Philip Eardley_web_768 joins the room
[16:03:24] Tom Jones_web_273 leaves the room
[16:03:28] Tom Jones_web_814 joins the room
[16:04:24] Zaheduzzaman Sarker_web_375 joins the room
[16:04:24] Lijun Dong_web_510 joins the room
[16:04:42] <Tom Jones_web_814> I am not getting any audio at all, can someone confirm its my problem?
[16:04:50] <Colin Perkins_web_865> I can hear the audio
[16:04:52] <Theresa Enghardt_web_928> I can hear the chairs
[16:05:07] <Tom Jones_web_814> thanks
[16:05:13] <Spencer Dawkins_web_964> Tom, do you have a VPN going?
[16:05:18] <Jen Linkova_web_345> Tom: are you on VPN by any chance?
[16:05:19] Luigi Iannone_web_137 leaves the room
[16:05:23] Lijun Dong_web_510 leaves the room
[16:05:31] Xavier de Foy_web_392 joins the room
[16:05:45] <Tom Jones_web_814> nope
[16:05:50] <Tom Jones_web_814> seems to be chrome
[16:06:14] Kireeti Kompella_web_465 leaves the room
[16:06:16] Anna Brunstrom_web_865 joins the room
[16:06:18] Kireeti Kompella_web_930 joins the room
[16:06:19] <Jen Linkova_web_345> Tom: I've seen cases this week when the browser restart helped
[16:06:24] John Kaippallimalil_web_552 joins the room
[16:06:27] Tom Jones_web_814 leaves the room
[16:06:31] Tom Jones_web_216 joins the room
[16:07:28] <Tom Jones_web_216> speakers stopped working in my laptop entirely
[16:07:30] Stuart Card_web_416 joins the room
[16:07:33] <Tom Jones_web_216> headphones are fine
[16:09:20] Hannu Flinck_web_636 joins the room
[16:09:33] Kireeti Kompella_web_930 leaves the room
[16:09:37] Kireeti Kompella_web_944 joins the room
[16:11:02] Luis Contreras_web_551 joins the room
[16:13:04] Nigel Davis_web_986 joins the room
[16:13:38] Umberto Fattore_web_562 joins the room
[16:17:24] <Colin Perkins_web_865> Some things don't need to be defined?
[16:20:15] <Brian Trammell_web_431> colin: that's my hope
[16:20:38] Chris Box_web_830 joins the room
[16:20:51] <Brian Trammell_web_431> OTOH it took us a while to get to something generic and useful for Node and Entity here
[16:21:02] <Mirja Kühlewind_web_660> no you have to define everything always in research
[16:21:24] <Mohamed Boucadair_web_972> The doc is really in good shape. I don't think that it does need to be too exhaustive.
[16:21:31] Carlos Bernardos_web_513 joins the room
[16:21:32] <Jen Linkova_web_345> well, I guess the part of the problem is we need to define what needs to be defined and what doesn't
[16:21:46] <Brian Trammell_web_431> feeling a little bad for derailing by saying the E word now
[16:22:06] <Mohamed Boucadair_web_972> What would be great is to see some arch-like I-Ds that make use of these terms.
[16:22:14] <Nicolas Kuhn_web_371> IMHO, whenever possible, pointing to existing RFC would help to define the terms (eg RFC8095 for Protocol Features available)
[16:24:27] Stewart Bryant_web_426 joins the room
[16:24:38] Lin Han_web_188 joins the room
[16:25:16] <Mirja Kühlewind_web_660> endpoint is hard to define
[16:25:31] <Mirja Kühlewind_web_660> depends on the layer/protocol
[16:25:50] Roland Bless_web_697 leaves the room
[16:26:06] <Brian Trammell_web_431> med: agreed... a high-level e.g. SCION arch draft referencing this one would be very useful (both as an intro to SCION and an exercise testing the vocabulary)
[16:27:45] <Nicolas Kuhn_web_371> +1 on the architectural document
[16:27:47] Stephan Emile_web_861 joins the room
[16:28:10] <Lixia Zhang_web_696> I would like to second Sabine's comment (hope I got it right): if a word is used in multiple protocol layers, lets make a unique term for each appearance of that word by attaching the context to it to make a unique term for that specific use.  A terminology document is meant to define terms clearly.
[16:29:33] Tarek Saad_web_843 joins the room
[16:30:09] <Spencer Dawkins_web_964> I'm going to mention this from the mike -
[16:30:22] Kireeti Kompella_web_944 leaves the room
[16:30:24] Frode Kileng_web_279 joins the room
[16:30:26] Kireeti Kompella_web_409 joins the room
[16:32:22] Michael Welzl_web_616 joins the room
[16:32:43] Taiji Kimura_web_194 joins the room
[16:32:52] Tarek Saad_web_843 leaves the room
[16:32:56] Tarek Saad_web_332 joins the room
[16:33:31] Tarek Saad_web_332 leaves the room
[16:33:35] Tarek Saad_web_200 joins the room
[16:37:59] frodek joins the room
[16:39:56] <Mohamed Boucadair_web_972> @nico: A "proxy" is a "service function" as per the properties I-D ;-)
[16:42:23] <Mohamed Boucadair_web_972> Agree with Theresa
[16:42:36] Taiji Kimura_web_194 leaves the room
[16:44:22] Ana Custura_web_370 leaves the room
[16:44:45] Lixia Zhang_web_696 leaves the room
[16:45:13] Nigel Davis_web_986 leaves the room
[16:45:39] <Tom Jones_web_216> first place to send it
[16:47:54] <Mohamed Boucadair_web_972> The sat doc does not need to answer to all the questions
[16:48:57] <Michael Welzl_web_616> @Gorry Is this the hat from the texas IETF?  I also have one of those!
[16:49:05] <Mirja Kühlewind_web_660> the British accent with that hat doesn't work...
[16:49:41] <Michael Welzl_web_616> Scottish!
[16:49:49] <Nicolas Kuhn_web_371> Thanks!
[16:49:52] <Tom Jones_web_216> English!
[16:50:15] <Michael Welzl_web_616> :rolling_on_the_floor_laughing:
[16:53:04] Carsten Bormann_web_600 leaves the room
[16:53:08] Carsten Bormann_web_375 joins the room
[16:53:34] <Mohamed Boucadair_web_972> MPTCP is in Standards Track since 2020
[16:54:09] <Gorry Fairhurst_web_444> @MW: Yes my "Dallas" hat :-)
[16:54:17] <Jen Linkova_web_345> I told Spencer that his slides MUST have pictures of  kittens - but he got us puppies audio!
[16:54:22] <Stuart Card_web_416> Actually using, in the field, concurrent multipath, for both failover and load balancing (not just flow by flow but packet by packet), since ~~2005. ;-)
[16:54:33] <Tom Jones_web_216> We need to get Gorry a profoundly scottish hat to wear
[16:54:59] <Brian Trammell_web_431> godzilla represents the fact that even though it's been in production for more than a decade, "multipath" is still the easiest way to end a TSV meeting
[16:55:22] <Jen Linkova_web_345> (I've just googled 'a profoundly scottish hat')
[16:55:31] Lars Eggert_web_931 joins the room
[16:55:58] <Mirja Kühlewind_web_660> scottish hat...?
[16:55:58] <Tom Jones_web_216> I think a Tam o' Shanter would be quite fetching
[16:56:06] <Mohamed Boucadair_web_972> I would say "basic path management" in slide 5
[16:56:07] <Tom Jones_web_216> or a glengarry in a pinch
[16:56:49] Ken Takayama_web_830 joins the room
[16:56:54] <Stuart Card_web_416> Also need a way to temporarily suspend use of paths that are currently poor but likely to recover.
[16:57:14] <Mohamed Boucadair_web_972> @stu: there is no such "hold" in the current version
[16:57:23] <Mohamed Boucadair_web_972> they started with basic features
[16:57:35] <Mohamed Boucadair_web_972> that's a good approach to see progress this work
[16:58:01] <Stuart Card_web_416> It happens a lot in aviation networking: plane banks, antenna points in bad direction, once plane levels out, link (and thus path) recovers.
[16:59:17] <Mohamed Boucadair_web_972> The second bullet would be OK if you don't depend on the remote peer
[16:59:27] Carlos Bernardos_web_513 leaves the room
[16:59:39] <Mohamed Boucadair_web_972> That's why we had RFC8803 for MPTCP
[17:01:20] <Mohamed Boucadair_web_972> I would not tag "benefits for early adopters" as "no problem".
[17:03:18] Karen O'Donoghue_web_127 joins the room
[17:03:39] <Mirja Kühlewind_web_660> if you own both endpoints (however that defined) in our deployment model, then it's okay
[17:03:59] <Mirja Kühlewind_web_660> s/our/your/
[17:04:17] <Anna Brunstrom_web_865> Item four should be comparing to single-path QUIC in this context I guess
[17:04:19] <Gorry Fairhurst_web_444> "Outperforming" is a long word to type, but the reality depends on what you hope for - resilliance; lower latency; more robustness to continuously changing paths; lower cost; etc.
[17:06:30] Stewart Bryant_web_426 leaves the room
[17:07:22] <Mohamed Boucadair_web_972> If MP-QUIC is enabled a la ATSSS, then the endpoint trussting midponts and vice versa become relevant
[17:08:33] <Mohamed Boucadair_web_972> for ATSSS:
[17:08:54] Brian Sipos_web_797 joins the room
[17:10:28] Toerless Eckert_web_946 leaves the room
[17:10:32] Toerless Eckert_web_369 joins the room
[17:11:09] John Kaippallimalil_web_552 leaves the room
[17:12:01] Michael zhou_web_308 joins the room
[17:12:20] Toerless Eckert_web_369 leaves the room
[17:13:52] <Cyrill Krähenbühl_web_104> I'll type my question
[17:14:31] Lin Han_web_188 leaves the room
[17:15:14] <Cyrill Krähenbühl_web_104> I would say that partial deployment heavily depends on the existence of different high-quality paths. If there are no good alternative paths then there is essentially no benefit for MPQUIC.
[17:15:22] Adrian Farrel_web_185 joins the room
[17:18:59] <Carsten Bormann_web_375> (Where high-quality includes "resilient")
[17:19:12] Erik Kline_web_889 leaves the room
[17:20:32] Éric Vyncke_web_511 joins the room
[17:20:35] Hannu Flinck_web_636 leaves the room
[17:20:39] Hannu Flinck_web_361 joins the room
[17:21:10] <Jen Linkova_web_345> I'd say 'equally quality pathS'
[17:21:31] <Stuart Card_web_416> Pareto dominance applies.
[17:21:59] <Mohamed Boucadair_web_972> Agree with Gorry except for resiliency: path migration would be just fine.
[17:22:08] <Brian Trammell_web_431> step one: get a time machine
[17:22:23] Michael zhou_web_308 leaves the room
[17:22:27] <Stuart Card_web_416> (fast, reliable) dominates all; (slow, unreliable) is dominated by all; (fast, unreliable) & (slow, reliable) are in same Pareto rank.
[17:23:45] <Mirja Kühlewind_web_660> This is the github link for comments as well:
[17:25:00] <Colin Perkins_web_865> It's much easier for the app to pass requirements to the transport, when the transport is closely integrated with the app
[17:25:58] <Gorry Fairhurst_web_444> Colin: Might be so, especially if there are a few simple choices.
[17:28:14] Toerless Eckert_web_997 joins the room
[17:30:40] Rayhaan Jaufeerally_web_631 leaves the room
[17:31:13] <Chris Box_web_830> We can certainly expect a complex set of network behaviours if we have endpoints using MPQUIC at the same time as part of the network using ATSSS using MPDCCP. Definitely one that needs thinking about.
[17:31:50] Éric Vyncke_web_511 leaves the room
[17:33:51] Kireeti Kompella_web_409 leaves the room
[17:35:07] <Gorry Fairhurst_web_444> Chris - why not say that at the Mic?
[17:35:41] <Gorry Fairhurst_web_444> .... two sets of policy working at the same time:-).
[17:36:47] <Gorry Fairhurst_web_444> ... or even one "application" being aided by two layers that have the same policy, but a different viewpoint.
[17:36:55] <Mohamed Boucadair_web_972> three actually as MPTCP is already the base ATSSS protocol
[17:37:03] Toerless Eckert_web_997 leaves the room
[17:37:03] <Gorry Fairhurst_web_444> :-)
[17:37:07] Toerless Eckert_web_621 joins the room
[17:37:09] <Mohamed Boucadair_web_972> The steering policies are provisionned by th network
[17:37:18] <Mohamed Boucadair_web_972> not by the app themselves
[17:37:52] <Mohamed Boucadair_web_972> there are polcies to bypass the ATSSS gateways if you discover that the remote peer is MPTCP-aware
[17:38:55] Toerless Eckert_web_621 leaves the room
[17:41:47] <Anna Brunstrom_web_865> Med already put the link here :)
[17:42:20] <Brian Trammell_web_431> ah, there it is. thanks
[17:42:24] <Brian Trammell_web_431> opening :)
[17:42:51] Rurick Kellermann_web_906 joins the room
[17:43:06] <Michael Welzl_web_616> Why worry. Roll out all the MP-flavors with and without proxies and see what happens. Yes there may be odd interactions, but surely it won't give us terrible congestion collapse style scenarios. Later, people can measure, do research, come back to the IETF and fine-tune - the wheel keeps turning like it always did.
[17:44:17] Tom Jones_web_216 leaves the room
[17:44:18] Erik Nordmark_web_201 joins the room
[17:44:50] Toerless Eckert_web_264 joins the room
[17:45:01] <Gorry Fairhurst_web_444> Just pasted into the notes ... it might be useful.
[17:45:11] <Gorry Fairhurst_web_444> The URL.
[17:45:21] Erik Nordmark_web_201 leaves the room
[17:45:30] <Michael Welzl_web_616> I thought my rambling, and was surprised!  :rolling_on_the_floor_laughing:
[17:46:04] <Anna Brunstrom_web_865> :rolling_on_the_floor_laughing:
[17:46:28] Rurick Kellermann_web_906 leaves the room
[17:46:39] Toerless Eckert_web_264 leaves the room
[17:46:58] <Nicolas Kuhn_web_371> there is indeed interesting research activity with multipath over satellite accesses
[17:47:43] Tarek Saad_web_200 leaves the room
[17:47:47] Tarek Saad_web_675 joins the room
[17:47:56] <Nicolas Kuhn_web_371>
[17:51:26] Lixia Zhang_web_497 joins the room
[17:54:59] Nicolas Kuhn_web_371 leaves the room
[17:55:00] Tarek Saad_web_675 leaves the room
[17:55:04] Tarek Saad_web_771 joins the room
[17:56:05] Erik Nordmark_web_791 joins the room
[17:56:08] <Chris Box_web_830> Wondering how Colin has enabled a blurred background over WebRTC
[17:56:18] Mohamed Boucadair_web_972 leaves the room
[17:56:22] Mohamed Boucadair_web_478 joins the room
[17:56:31] <Mirja Kühlewind_web_660> by having a real camera
[17:56:34] <Colin Perkins_web_865> fast lens
[17:56:47] Erik Nordmark_web_791 leaves the room
[17:56:50] Toerless Eckert_web_892 joins the room
[17:56:56] <Chris Box_web_830> Large aperture - the traditional way
[17:57:15] Tarek Saad_web_771 leaves the room
[17:57:19] Tarek Saad_web_306 joins the room
[17:58:04] Bhavit Shah_web_223 leaves the room
[17:58:05] Mohamed Boucadair_web_478 leaves the room
[17:58:06] Tarek Saad_web_306 leaves the room
[17:58:09] Mohamed Boucadair_web_782 joins the room
[17:58:10] Tarek Saad_web_804 joins the room
[17:58:11] <Zaheduzzaman Sarker_web_375> +1 to Brian.. as I said we need to have the conversation..
[17:58:36] Toerless Eckert_web_892 leaves the room
[17:58:40] Toerless Eckert_web_150 joins the room
[17:59:24] <Mirja Kühlewind_web_660> scheduling is in scope for ICCRG as well (not only congestion control)
[17:59:53] <Colin Perkins_web_865> right
[17:59:56] <Brian Trammell_web_431> true
[18:00:20] Lixia Zhang_web_497 leaves the room
[18:00:36] <Colin Perkins_web_865> and some of the issue touch on programmability of the network and forwarding plane too
[18:01:51] Mohamed Boucadair_web_782 leaves the room
[18:01:55] Mohamed Boucadair_web_686 joins the room
[18:02:33] Mohamed Boucadair_web_686 leaves the room
[18:02:34] Toerless Eckert_web_150 leaves the room
[18:02:37] Mohamed Boucadair_web_365 joins the room
[18:02:38] Toerless Eckert_web_453 joins the room
[18:03:54] Philip Eardley_web_768 leaves the room
[18:04:03] Luis Contreras_web_551 leaves the room
[18:04:03] Adam Wiethuechter_web_978 leaves the room
[18:04:07] <Mirja Kühlewind_web_660> bye! thx!
[18:04:08] Mohamed Boucadair_web_365 leaves the room
[18:04:09] Michael Welzl_web_616 leaves the room
[18:04:09] Zaheduzzaman Sarker_web_375 leaves the room
[18:04:12] John Border_web_384 leaves the room
[18:04:12] Jen Linkova_web_345 leaves the room
[18:04:13] Chris Box_web_830 leaves the room
[18:04:13] <Antoine Fressancourt_web_326> Thank you
[18:04:17] Sabine Randriamasy_web_992 leaves the room
[18:04:17] Colin Perkins_web_865 leaves the room
[18:04:18] Mirja Kühlewind_web_660 leaves the room
[18:04:21] Xavier de Foy_web_392 leaves the room
[18:04:24] Mihail Zverev_web_347 leaves the room
[18:04:28] Spencer Dawkins_web_964 leaves the room
[18:04:28] Gorry Fairhurst_web_444 leaves the room
[18:04:28] Alessandro Amirante_web_456 leaves the room
[18:04:28] Theresa Enghardt_web_928 leaves the room
[18:04:28] Brian Trammell_web_431 leaves the room
[18:04:28] Antoine Fressancourt_web_326 leaves the room
[18:04:28] Tianji Jiang_web_239 leaves the room
[18:04:28] Cyrill Krähenbühl_web_104 leaves the room
[18:04:28] Hirochika Asai_web_815 leaves the room
[18:04:28] Julien Maisonneuve_web_843 leaves the room
[18:04:28] Kazuaki Ueda_web_942 leaves the room
[18:04:28] Stuart Card_web_416 leaves the room
[18:04:28] Umberto Fattore_web_562 leaves the room
[18:04:28] Anna Brunstrom_web_865 leaves the room
[18:04:28] Stephan Emile_web_861 leaves the room
[18:04:28] Frode Kileng_web_279 leaves the room
[18:04:28] Carsten Bormann_web_375 leaves the room
[18:04:28] Lars Eggert_web_931 leaves the room
[18:04:28] Ken Takayama_web_830 leaves the room
[18:04:28] Karen O'Donoghue_web_127 leaves the room
[18:04:28] Brian Sipos_web_797 leaves the room
[18:04:28] Adrian Farrel_web_185 leaves the room
[18:04:28] Tarek Saad_web_804 leaves the room
[18:04:28] Hannu Flinck_web_361 leaves the room
[18:04:28] Toerless Eckert_web_453 leaves the room
[18:04:31] frodek leaves the room
[18:05:31] alexamirante leaves the room
[18:06:05] Meetecho leaves the room