IETF
pals@jabber.ietf.org
Friday, March 12, 2021< ^ >
Room Configuration
Room Occupants

GMT+0
[06:39:15] zulipbot leaves the room: Disconnected: closed
[06:41:27] zulipbot joins the room
[06:48:28] zulipbot leaves the room: Disconnected: closed
[06:48:42] zulipbot joins the room
[06:49:31] zulipbot leaves the room: Disconnected: closed
[06:49:48] zulipbot joins the room
[06:52:12] glen joins the room
[06:52:14] glen leaves the room
[06:58:48] zulipbot leaves the room: Disconnected: closed
[07:10:52] zulipbot joins the room
[07:12:26] glen joins the room
[07:12:30] glen leaves the room
[07:29:15] zulipbot leaves the room: Disconnected: closed
[07:29:21] zulipbot joins the room
[11:47:38] Meetecho joins the room
[11:50:03] Tarek Saad_web_906 joins the room
[11:50:03] Tony Li_web_506 joins the room
[11:50:03] Alessandro Toppi_web_269 joins the room
[11:50:03] Richard Li_web_404 joins the room
[11:50:03] Rakesh Gandhi_web_725 joins the room
[11:50:03] Zheng Zhang_web_898 joins the room
[11:50:03] Vishnu Beeram_web_929 joins the room
[11:50:03] Feng Yang_web_165 joins the room
[11:50:03] Daniam Henriques_web_412 joins the room
[11:50:03] Stewart Bryant_web_419 joins the room
[11:50:03] Dhruv Dhody_web_991 joins the room
[11:50:36] Balazs Varga_web_935 joins the room
[11:51:18] Bruno Decraene_web_498 joins the room
[11:52:03] Andy Malis_web_490 joins the room
[11:52:38] dhruvdhody joins the room
[11:52:59] Adrian Farrel_web_734 joins the room
[11:53:25] Xiao Min_web_870 joins the room
[11:54:07] Liu Yao_web_483 joins the room
[11:54:19] Ethan Grossman_web_538 joins the room
[11:54:22] dhruvdhody has set the subject to: IETF-110
[11:54:59] John Drake_web_991 joins the room
[11:55:02] Yisong Liu_web_982 joins the room
[11:55:32] John Drake_web_991 leaves the room
[11:55:35] Joel Halpern_web_862 joins the room
[11:55:36] Jim Guichard_web_413 joins the room
[11:55:37] Martin Vigoureux_web_393 joins the room
[11:56:07] John Drake_web_271 joins the room
[11:56:13] Loa Andersson_web_691 joins the room
[11:56:36] Toerless Eckert_web_998 joins the room
[11:56:42] Dave Sinicrope_web_963 joins the room
[11:57:14] Ville Hallivuori_web_746 joins the room
[11:57:43] Matthew Bocci_web_662 joins the room
[11:58:09] Donald Eastlake_web_452 joins the room
[11:58:19] János Farkas_web_116 joins the room
[11:58:24] Jie Dong_web_409 joins the room
[11:58:38] Sergey Fomin_web_253 joins the room
[11:58:40] Deborah Brungard_web_206 joins the room
[11:58:41] Takuya Miyasaka_web_335 joins the room
[11:58:44] Nicolai Leymann_web_584 joins the room
[11:58:46] Shunwan Zhuang_web_297 joins the room
[11:58:51] Yuji Tochio_web_411 joins the room
[11:58:52] John Scudder_web_552 joins the room
[11:58:56] Yingzhen Qu_web_418 joins the room
[11:59:00] VJ Shah_web_443 joins the room
[11:59:04] Yaakov Stein_web_913 joins the room
[11:59:08] Ruediger Geib_web_942 joins the room
[11:59:08] Geir Egeland_web_349 joins the room
[11:59:17] Matt Green_web_672 joins the room
[11:59:22] Lou Berger_web_618 joins the room
[11:59:23] Bill Fenner_web_358 joins the room
[11:59:29] Boris Khasanov_web_698 joins the room
[11:59:30] Ketan Talaulikar_web_989 joins the room
[11:59:35] Jeong-dong Ryoo_web_529 joins the room
[11:59:46] <Lou Berger_web_618> hello all
[11:59:49] Benjamin Steinert_web_171 joins the room
[11:59:55] Mach Chen_web_837 joins the room
[12:00:00] Zhaohui Zhang_web_175 joins the room
[12:00:01] <Boris Khasanov_web_698> Hi Lou and all!
[12:00:13] Greg Mirsky_web_793 joins the room
[12:00:24] Zongpeng Du_web_446 joins the room
[12:00:26] David Pelton_web_151 joins the room
[12:00:27] Zafar Ali_web_476 joins the room
[12:00:34] Haoyu Song_web_658 joins the room
[12:00:36] Junfeng Zhao_web_910 joins the room
[12:00:38] Les Ginsberg_web_192 joins the room
[12:00:42] Fengwei Qin_web_605 joins the room
[12:00:49] Sandra Murphy_web_906 joins the room
[12:01:13] Yeoncheol Ryoo_web_778 joins the room
[12:01:27] Zhenbin Li_web_875 joins the room
[12:01:29] Italo Busi_web_404 joins the room
[12:01:33] Wim Henderickx_web_622 joins the room
[12:02:14] dee3@hot-chilli.net joins the room
[12:02:30] Kireeti Kompella_web_303 joins the room
[12:02:30] Srihari Sangli_web_709 joins the room
[12:02:32] Tomoko Nezu_web_382 joins the room
[12:02:58] Torunn Narvestad_web_322 joins the room
[12:02:59] Tom Hill_web_121 joins the room
[12:03:31] Ruixue Wang_web_570 joins the room
[12:03:39] toerless@jabber.at joins the room
[12:04:00] Jeff Tantsura_web_282 joins the room
[12:04:07] Ran Chen_web_203 joins the room
[12:04:10] Richard Vallee_web_702 joins the room
[12:04:11] Fan Yang_web_113 joins the room
[12:04:33] <toerless@jabber.at> "stewart will finish us off" how subtle ;-)
[12:04:46] <Lou Berger_web_618> https://codimd.ietf.org/notes-ietf-110-pals?both
[12:04:47] Pascal Thubert_web_158 joins the room
[12:04:51] Sergey Fomin_web_253 leaves the room
[12:04:53] Farni Boten_web_927 joins the room
[12:04:57] Sergey Fomin_web_867 joins the room
[12:04:58] Praveen Mada_web_233 joins the room
[12:05:11] Rick Taylor_web_302 joins the room
[12:05:36] Nils Warnke_web_620 joins the room
[12:05:51] Al Rocheleau_web_787 joins the room
[12:06:22] Kiran Makhijani_web_138 joins the room
[12:07:42] Xufeng Liu_web_536 joins the room
[12:07:50] Luis Contreras_web_194 joins the room
[12:07:54] Ram Santhanakrishnan_web_874 joins the room
[12:08:17] Zongpeng Du_web_446 leaves the room
[12:08:24] Zongpeng Du_web_602 joins the room
[12:08:49] Himanshu Shah_web_876 joins the room
[12:09:08] Shunsuke Homma_web_983 joins the room
[12:09:31] <Yaakov Stein_web_913> well, LFIB actually (isn't that officially ILM?)
[12:13:27] Rakesh Gandhi_web_725 leaves the room
[12:13:31] Rakesh Gandhi_web_975 joins the room
[12:14:58] <Lou Berger_web_618> Minor point. this slide is missing the EVPN CW
[12:15:41] Haibo Wang_web_688 joins the room
[12:16:23] Merike Kaeo_web_656 joins the room
[12:16:27] Kamran Raza_web_387 joins the room
[12:16:43] <Kireeti Kompella_web_303> the path identifier (also beyond eos) is 2-3 years old, now expired
[12:16:46] Luis Contreras_web_194 leaves the room
[12:16:52] Luis Contreras_web_764 joins the room
[12:17:34] <Yaakov Stein_web_913> if you can pick under the ToS it isn't a STACK is it?
[12:17:43] <Yaakov Stein_web_913> s/pick/peek/
[12:18:21] <Tony Li_web_506> It's still a stack, just an out-of--bounds reference.
[12:19:06] <Kireeti Kompella_web_303> path id draft missing
[12:19:06] <Yaakov Stein_web_913> I meant "walking" down the stack to see the first nibble after it (no problem with popping the whole stack and THEN finding additional data)
[12:19:46] Ian Duncan_web_499 joins the room
[12:19:49] <Tony Li_web_506> Searching is obviously a giant pain in the forwarding plane.
[12:21:36] Daniele Ceccarelli_web_716 joins the room
[12:21:51] <toerless@jabber.at> we have a lot of headroom to grow MPLS size and complexity:
[12:21:58] <Tony Li_web_506> Kireeti, your volume is very low.
[12:21:59] <toerless@jabber.at> 1 bit less complex/long as SRv6 ;-)
[12:22:43] LouBerger joins the room
[12:22:44] Fengwei Qin_web_605 leaves the room
[12:22:48] Fengwei Qin_web_161 joins the room
[12:23:11] <Lou Berger_web_618> @all, please add @mic if you want your comment repeated by one of the chairs
[12:23:33] Ruixue Wang_web_570 leaves the room
[12:23:42] Ian Duncan_web_499 leaves the room
[12:23:55] Ian Duncan_web_546 joins the room
[12:25:03] Ruixue Wang_web_988 joins the room
[12:27:17] Fengwei Qin_web_161 leaves the room
[12:27:20] Sandra Murphy_web_906 leaves the room
[12:27:31] Fengwei Qin_web_668 joins the room
[12:28:25] Eve Schooler_web_272 joins the room
[12:30:01] Pascal Thubert_web_158 leaves the room
[12:30:07] Pascal Thubert_web_337 joins the room
[12:30:20] Andy Malis_web_490 leaves the room
[12:30:25] Andy Malis_web_832 joins the room
[12:30:30] <Yaakov Stein_web_913> slides?
[12:31:25] Liu Yao_web_483 leaves the room
[12:31:29] <Kireeti Kompella_web_303> nondeterministic slides talking about deterministic behavior?
[12:31:29] Liu Yao_web_981 joins the room
[12:31:39] Uma Chunduri_web_610 joins the room
[12:32:27] <Yaakov Stein_web_913> determinism enforced by sequence number on each slide
[12:33:52] <Jie Dong_web_409> may need replicated slides?
[12:34:18] Les Ginsberg_web_192 leaves the room
[12:34:21] <Yaakov Stein_web_913> At least there is spare time at the end so we won't need to eliminate any slides
[12:34:23] Les Ginsberg_web_613 joins the room
[12:36:13] james welch_web_725 joins the room
[12:36:17] <Yaakov Stein_web_913> in the previous slide each node was EITHER R or E. Why can't a node eliminate extra copies and replicate going forward?
[12:36:19] Zafar Ali_web_476 leaves the room
[12:36:25] Zafar Ali_web_852 joins the room
[12:36:55] <John Scudder_web_552> Too hard to fit in the ASCII art?
[12:37:23] Zafar Ali_web_852 leaves the room
[12:37:26] <LouBerger> @yakov -- it can
[12:37:31] <LouBerger> it was just an example
[12:37:38] <Rick Taylor_web_302> I think they are "logical" functions, rather than an indication of physical location.  Nothing to stop the functions being co-located
[12:37:53] Ram Santhanakrishnan_web_874 leaves the room
[12:38:00] Ram Santhanakrishnan_web_514 joins the room
[12:39:01] <Yaakov Stein_web_913> Just that in the example there was no node with both.
[12:39:04] Xueyan Song_web_555 joins the room
[12:39:06] Ram Santhanakrishnan_web_514 leaves the room
[12:39:10] Ram Santhanakrishnan_web_585 joins the room
[12:39:11] Ram Santhanakrishnan_web_585 leaves the room
[12:39:17] Ram Santhanakrishnan_web_598 joins the room
[12:39:44] Eve Schooler_web_272 leaves the room
[12:40:01] <Yaakov Stein_web_913> To allay misunderstanding there should have been at least one E+R node.
[12:40:35] Fengwei Qin_web_668 leaves the room
[12:40:42] <Kiran Makhijani_web_138> and i guess it was for one direction. opposite direction flow should reverse the roles.
[12:40:43] Fengwei Qin_web_400 joins the room
[12:40:50] <Yaakov Stein_web_913> Have allocated Ethernet MAC addresses gotten up to 4 in the first nibble yet?
[12:41:37] Guangpeng Li_web_271 joins the room
[12:41:49] <Bill Fenner_web_358> 44-D7-91   (hex)                HUAWEI TECHNOLOGIES CO.,LTD
[12:42:02] <Mach Chen_web_837> Similar to PW, single PW, there will no node with both, MS-PW, the S-PE will be with R+E
[12:42:43] <Bill Fenner_web_358> 2051 entries in oui.txt that begin with 4 or 6
[12:42:58] Yuehua Wei_web_934 joins the room
[12:44:04] <Yaakov Stein_web_913> @Bill - wow! They're headed for address exhaustion as well. Next stop EUI64.
[12:44:32] <Tony Li_web_506> This is not news.
[12:45:13] <Yaakov Stein_web_913> I thought they still had many years...
[12:45:13] <Tony Li_web_506> We are more likely to see MAC reuse before EUI64.
[12:45:27] <Bill Fenner_web_358> less than 30k OUIs in the file, though, allocation is not sequential
[12:45:48] <Tony Li_web_506> Those are the public ones, right?
[12:46:01] <Rick Taylor_web_302> How much code out there has "char mac[6]" and how much to fix? ;)
[12:46:03] <John Scudder_web_552> MAC reuse sounds like a movie that won't end well.
[12:46:23] <Rick Taylor_web_302> Wasn't there a MAC-NAT draft out there?
[12:47:09] <Yaakov Stein_web_913> You can double the number by redefining the U bit
[12:47:12] <Jeff Tantsura_web_282> MAC uniqueness is not there for years
[12:47:18] <Bill Fenner_web_358> the public numbers are pretty sequential until 00-30-FF, then there are varying size gaps
[12:48:20] <Tony Li_web_506> The failure of EUI64 to catch on kinda makes reuse inevitable.
[12:48:40] Ram Santhanakrishnan_web_598 leaves the room
[12:48:41] <Yaakov Stein_web_913> You could have said that about IPv6 a few years ago
[12:48:45] Weiqiang Cheng_web_525 joins the room
[12:48:45] Ram Santhanakrishnan_web_517 joins the room
[12:48:48] Matthew Bocci_web_662 leaves the room
[12:48:49] Ram Santhanakrishnan_web_517 leaves the room
[12:48:54] Matthew Bocci_web_407 joins the room
[12:48:54] Ram Santhanakrishnan_web_656 joins the room
[12:49:04] Fengwei Qin_web_400 leaves the room
[12:49:10] Fengwei Qin_web_339 joins the room
[12:49:28] <Tony Li_web_506> And we have reuse through NAT. I'm just predicting that OUIs follow a similar path.
[12:50:42] Liu Yao_web_981 leaves the room
[12:50:47] Liu Yao_web_277 joins the room
[12:51:23] <Yaakov Stein_web_913> Had WiFi used EUI64 ...
[12:52:14] <John Scudder_web_552> NATting MACs in the enterprise, while bletcherous, seems practicable. But I don't even know how many MACs I have in my home. A lot.
[12:52:42] <John Scudder_web_552> I suppose we've already trained home users to assume none of this technology works right anyway, and that they should just throw away nonfunctional devices and play again.
[12:53:17] Srihari Sangli_web_709 leaves the room
[12:53:24] <Yaakov Stein_web_913> The problem is that unlike IP addresses, physical machines need an allocated MAC address. To DHCP MAC addresses you'll need yet another lower layer.
[12:53:27] Ram Santhanakrishnan_web_656 leaves the room
[12:53:29] Srihari Sangli_web_550 joins the room
[12:53:40] Ram Santhanakrishnan_web_384 joins the room
[12:53:52] Dhruv Dhody_web_991 leaves the room
[12:53:57] Dhruv Dhody_web_595 joins the room
[12:54:00] <Tony Li_web_506> I expect that rather than a MAC NAT, we'll see resale of old OUI.
[12:54:13] <Joel Halpern_web_862> I would not be surprised to see enterprise move to setting all the Ethernet Interfaces they control to local OUIs that they can manage.  Then at least those will not collide with each other or with BYOD.
[12:54:34] <Tony Li_web_506> And then we'll start seeing devices that dynamically probe the network for a MAC and grab one that appears to be unused.
[12:54:37] Ran Chen_web_203 leaves the room
[12:54:41] Ran Chen_web_437 joins the room
[12:54:51] <Kireeti Kompella_web_303> DHCP for MAC?
[12:55:04] <Joel Halpern_web_862> The case I hear about that is messy is a massive data center with arbitrary layer 2 interconnect and many virtual entities each needing an Ethernet address.  All not an IETF problem.
[12:55:05] <Tony Li_web_506> AppleTalk for MACs.
[12:55:23] <John Scudder_web_552> @tli for sure. That will work OK in small networks.
[12:55:29] <John Scudder_web_552> Like AppleTalk.
[12:55:47] <Boris Khasanov_web_698> What's happening with MACs which were assigned to non-existing companies?  Could those to be re-used again?
[12:56:32] <Tony Li_web_506> Yes. Old OUI will be resold, just like IPv4 prefixes.
[12:56:50] <dee3@hot-chilli.net> 64 bit MACs used in InfiniBand, 802.15.4 (ZigBee), and IEEE 1394 (Firewire/i.Link).
[12:57:05] yangrulan_web_865 joins the room
[12:57:25] <Yaakov Stein_web_913> @John - thanks for teaching me the word "bletcherous". I'll have to start using that!
[12:57:43] <Kireeti Kompella_web_303> what does Bluetooth use? how many bits?
[12:58:29] Daniele Ceccarelli_web_716 leaves the room
[12:58:33] <Yaakov Stein_web_913> Bluetooth uses EUI48 like Ethernet
[12:58:47] <dee3@hot-chilli.net> DHCP6 for MACs in RFC 8947.
[12:58:48] <Kireeti Kompella_web_303> a cross between bletch and treacherous?
[12:58:58] <Yaakov Stein_web_913> Firewire chose correctly, but died.
[12:59:17] <John Scudder_web_552> I did not originate "bletcherous" but I'm pleased to promote its use.
[12:59:24] yangrulan_web_865 leaves the room
[12:59:33] <Boris Khasanov_web_698> :)
[12:59:37] <dee3@hot-chilli.net> I think Bluetooth (802.15.1) uses 48 bit…
[12:59:38] <Yaakov Stein_web_913> etymology?
[12:59:49] Weiqiang Cheng_web_525 leaves the room
[12:59:59] Weiqiang Cheng_web_847 joins the room
[13:00:18] <John Scudder_web_552> https://en.wiktionary.org/wiki/bletcherous
[13:00:59] <Yaakov Stein_web_913> OK, says it sound like other bletcherous words.
[13:01:05] <John Scudder_web_552> http://www.zvon.org/comp/r/ref-Jargon_file.html#Terms~~bletcherous
[13:01:07] <Yaakov Stein_web_913> But no credit given
[13:01:24] <John Scudder_web_552> doesn't answer your question but adds more color
[13:01:29] <Tony Li_web_506> Yiddish per wordsense.eu
[13:02:29] <Yaakov Stein_web_913> That site is blocked for me. But it doesn't ring any bells (admittedly my Yiddish is not very good - mostly slang and curses)
[13:02:31] János Farkas_web_116 leaves the room
[13:02:39] <Kireeti Kompella_web_303> it's somewhat onomatopoeic (if I spelt that right) esp when accompanied by an appropriate facial expression
[13:03:25] <Yaakov Stein_web_913> you mean from blech (pronounced with a guttural ch) ?
[13:03:25] János Farkas_web_611 joins the room
[13:03:51] <John Scudder_web_552> it seems plausible anyway
[13:03:56] <Kireeti Kompella_web_303> either guttural ch or english ch
[13:04:00] <Balazs Varga_web_935> Thanks Greg for the comment. Fully agree, slide 19 would break DetNet
[13:04:08] <Yaakov Stein_web_913> Were it from Yiddish it would be dreckerous
[13:04:17] <Tony Li_web_506> @Yaakov yes.
[13:04:43] <LouBerger> this would limit this soultion to the DetNet forwarding sub-layer
[13:05:19] Zhenbin Li_web_875 leaves the room
[13:05:24] Zhenbin Li_web_483 joins the room
[13:05:25] Jingrong Xie_web_779 joins the room
[13:05:30] <LouBerger> i.e., would be limited to working with the f-label
[13:07:47] Richard Vallee_web_702 leaves the room
[13:07:54] Richard Vallee_web_526 joins the room
[13:09:50] Ahmed Bashandy_web_851 joins the room
[13:10:21] <Jim Guichard_web_413> great slide!
[13:12:27] Zafar Ali_web_292 joins the room
[13:12:27] Gyan Mishra_web_715 joins the room
[13:14:57] Jingrong Xie_web_779 leaves the room
[13:14:59] Sandra Murphy_web_595 joins the room
[13:15:03] Rakesh Gandhi_web_975 leaves the room
[13:15:07] Rakesh Gandhi_web_970 joins the room
[13:15:26] LucAndré Burdet_web_773 joins the room
[13:17:14] <Yaakov Stein_web_913> If our processing is so much more powerful, then why not put generic opcodes in the header?
[13:17:29] <Yaakov Stein_web_913> (and I don't mean SIDs)
[13:18:14] <Tony Li_web_506> You could.
[13:18:38] <Tony Li_web_506> Heck, you could put an entire algorithm in p-code in your stack.
[13:18:45] <Yaakov Stein_web_913> and we wouldn't need P4!
[13:19:17] <Tony Li_web_506> Of course, it won't work for 99.5% of the gear that's out there.
[13:19:19] <Yaakov Stein_web_913> that is what I meant. The first opcode could say jump 12 bytes read the next 2 bytes, etc. A full Turing machine.
[13:19:44] <Tony Li_web_506> As an Evil Greedy Vendor, I would love to obsolete all gear in the field.
[13:20:08] <Tony Li_web_506> Or, we could be less complex...
[13:20:38] <Yaakov Stein_web_913> <blush> I'll stay silent </blush>
[13:21:03] <Boris Khasanov_web_698> SRv6 is already close to it with Locator+ Function+Arg encoding in SID
[13:21:36] <Yaakov Stein_web_913> I don't buy that SRv6 network programming story.
[13:22:08] <Jeff Tantsura_web_282> BIER pushed serious changes in HW, when the value is obvious...
[13:22:33] <Yaakov Stein_web_913> and unlike SR-MPLS, SRv6 already obsoletes all gear in the field
[13:23:03] <Tony Li_web_506> And creates an adoption deterrent.
[13:25:03] Eric Gray_web_327 joins the room
[13:26:42] Jingrong Xie_web_867 joins the room
[13:26:58] Sandra Murphy_web_595 leaves the room
[13:27:02] Sandra Murphy_web_112 joins the room
[13:28:17] <Boris Khasanov_web_698> As example, It is hard to predict how these changes for SPL will influence old gear too. So that is much broader issue anyway.
[13:29:00] Ian Duncan_web_546 leaves the room
[13:30:28] <Jeff Tantsura_web_282> any MPLS FW code treats 0-15 differently
[13:31:08] <Tony Li_web_506> There is chunk of old gear that is hard-coded. All of this will break things.
[13:31:28] <Yaakov Stein_web_913> we'll need a bit like in LDP to tell us when not supported whether to ignore or discard
[13:31:59] <Tony Li_web_506> Some gear is programmable. Stuff can be added, but the more complicated we make things, the longer the adoption cycle.
[13:32:05] Praveen Mada_web_233 leaves the room
[13:32:06] <Adrian Farrel_web_734> The dodgy question has always been that *if* the SPL reaches the top of stack, will the TC and TTL be looked at before the label value? AFAICS this only matters if TTL is 0 because it might cause a different discard path from if the label is SPL
[13:32:09] Praveen Mada_web_887 joins the room
[13:32:18] <Jeff Tantsura_web_282> No free lunch indeed
[13:32:34] Weiqiang Cheng_web_847 leaves the room
[13:32:45] Weiqiang Cheng_web_516 joins the room
[13:32:59] <Adrian Farrel_web_734> Of course, if any use of the Kireeti-label *requires* processing at a transit node, then the challenge is to ensure that the right nodes can process correctly
[13:33:44] Ketan Talaulikar_web_989 leaves the room
[13:33:48] Ketan Talaulikar_web_804 joins the room
[13:35:02] Ahmed Bashandy_web_851 leaves the room
[13:35:04] Matthew Bocci_web_407 leaves the room
[13:35:06] Ahmed Bashandy_web_185 joins the room
[13:35:14] Matthew Bocci_web_592 joins the room
[13:35:53] John Drake_web_271 leaves the room
[13:36:08] John Drake_web_534 joins the room
[13:36:22] John Drake_web_534 leaves the room
[13:36:26] John Drake_web_528 joins the room
[13:36:34] Richard Vallee_web_526 leaves the room
[13:37:12] <Vishnu Beeram_web_929> Like Jeff said -- no free lunch; https://datatracker.ietf.org/doc/html/draft-kompella-mpls-mspl4fa-00#section-3.1 -- the key is ensure that the FAI doesn't show up at the Top Of Stack..
[13:37:15] Weiqiang Cheng_web_516 leaves the room
[13:38:01] <Kireeti Kompella_web_303> @adrian 1 -- yes, that's why we were paranoid that the ELI/EL both had TTL = 0
[13:38:13] Weiqiang Cheng_web_328 joins the room
[13:38:38] John Drake_web_528 leaves the room
[13:38:42] <Adrian Farrel_web_734> Yes, Pavan, but saying router X must not let the LSE reach top of stack, doesn't help router Y process the packet when router X makes an error
[13:38:43] John Drake_web_210 joins the room
[13:38:53] János Farkas_web_611 leaves the room
[13:39:44] <Kireeti Kompella_web_303> @adrian 2 -- we already have this, like it or not.  So far, it's a nice to have, not a must have (as in ELI)
[13:39:45] <Adrian Farrel_web_734> IIRC the most risky case was pop-and-go forwarding
[13:39:54] Weiqiang Cheng_web_328 leaves the room
[13:39:57] <Jeff Tantsura_web_282> @adrian as long as the behavior is deterministic
[13:40:02] Weiqiang Cheng_web_565 joins the room
[13:40:06] John Drake_web_210 leaves the room
[13:40:53] Yahya_web_354 joins the room
[13:41:40] <Kireeti Kompella_web_303> Pavan/Adrian: putting verbiage in case of PFE misbehavior would be good.  but there lie dragons
[13:41:43] <Yaakov Stein_web_913> Here he's saying I what said above - a generic Turing machine!
[13:42:12] Jim Guichard_web_413 leaves the room
[13:42:35] <Kireeti Kompella_web_303> if a PFE popped two labels in a regular label stack (or SR stack), who knows what will happen?
[13:42:52] Ram Santhanakrishnan_web_384 leaves the room
[13:42:58] John Drake_web_833 joins the room
[13:43:07] <Kireeti Kompella_web_303> (when it was supposed to pop just one).  Or it swapped to a random label ...
[13:43:10] <John Scudder_web_552> We could add a mandatory "don't write bugs" section to all our other mandatory sections. Is that the essence of the suggestion?
[13:43:22] <Jeff Tantsura_web_282> :)
[13:43:35] <Mach Chen_web_837> :joy:
[13:43:42] <Adrian Farrel_web_734> Bugs are OK so long as they are kept within a limited domain
[13:43:43] <Tony Li_web_506> Bugs are proportional to complexity.
[13:44:06] Gyan Mishra_web_715 leaves the room
[13:44:10] Gyan Mishra_web_109 joins the room
[13:44:18] <Jeff Tantsura_web_282> BugAbstractioLayer ;-)
[13:44:24] <Kireeti Kompella_web_303> @jgs not quite.  but a short description of how bad things could get with certain PFE misbehaviors might be useful
[13:44:30] <Yaakov Stein_web_913> "don't write bugs" is like an "evil bit" ?
[13:44:43] Geng-Da Tsai_web_352 joins the room
[13:44:55] <Adrian Farrel_web_734> Bottom of Sin flag?
[13:45:08] <Yahya_web_354> Bugs are unintended features
[13:45:52] <Kireeti Kompella_web_303> "Here are the kinds of things that could go wrong if such-and-such label comes to the top of stack"
[13:45:54] <Yaakov Stein_web_913> I use a "debugger", but it isn't a program that automatically removes bugs!
[13:46:36] <Jeff Tantsura_web_282> If TOS==SPL -> drop -  always safe
[13:46:59] <Kireeti Kompella_web_303> agreed.
[13:47:00] Zhenbin Li_web_483 leaves the room
[13:47:35] Zhenbin Li_web_157 joins the room
[13:47:43] <Bruno Decraene_web_498> bug in top label forwarding leads to VPN breach
[13:47:46] <Kireeti Kompella_web_303> i take it back.  TOS = 0 or 2 has a well-defined behaviour
[13:48:06] <Jeff Tantsura_web_282> Yes
[13:48:38] Ahmed Bashandy_web_185 leaves the room
[13:48:38] Jingrong Xie_web_867 leaves the room
[13:48:52] <Kireeti Kompella_web_303> if TOS is unrecognized SPL, drop; if recognized, process per spec.
[13:49:00] <Boris Khasanov_web_698> How to check that TOS  label does not equal SPl? Special range?
[13:49:10] Jingrong Xie_web_894 joins the room
[13:49:14] Jingrong Xie_web_894 leaves the room
[13:49:15] Ahmed Bashandy_web_940 joins the room
[13:49:15] Jingrong Xie_web_868 joins the room
[13:49:26] <Jeff Tantsura_web_282> Yes, SPLs are sspecial
[13:49:27] <Kireeti Kompella_web_303> @bruno: bug in forwarding can be many things, among the worst being a breach of VPN
[13:49:30] Weiqiang Cheng_web_565 leaves the room
[13:49:34] Weiqiang Cheng_web_470 joins the room
[13:49:41] Weiqiang Cheng_web_470 leaves the room
[13:49:54] Weiqiang Cheng_web_226 joins the room
[13:49:56] <Jeff Tantsura_web_282> +1 kireeti
[13:50:02] <Kireeti Kompella_web_303> SPL = [0, 16)
[13:50:12] Fengwei Qin_web_339 leaves the room
[13:50:16] Fengwei Qin_web_294 joins the room
[13:50:31] Fengwei Qin_web_294 leaves the room
[13:50:35] Al Rocheleau_web_787 leaves the room
[13:50:45] Fengwei Qin_web_898 joins the room
[13:51:04] <Kireeti Kompella_web_303> (in answer to Boris)
[13:52:03] <Zhenbin Li_web_157> I think General Delivery Function is to boil the ocean. It is too challenging.
[13:52:12] Fengwei Qin_web_898 leaves the room
[13:52:37] <Jeff Tantsura_web_282> 16 is MLI
[13:52:38] Jingrong Xie_web_868 leaves the room
[13:52:49] Fengwei Qin_web_907 joins the room
[13:52:52] <dee3@hot-chilli.net> Greg: What was that RFC #?
[13:53:12] Jingrong Xie_web_531 joins the room
[13:53:50] <Deborah Brungard_web_206> Thanks Stewart et.
[13:54:30] <Adrian Farrel_web_734> @greg 8595
[13:54:40] <Deborah Brungard_web_206> and all, really useful session, we should do more often
[13:54:49] <Adrian Farrel_web_734> but it is an Extended SPL
[13:55:06] <Jeff Tantsura_web_282> @robin - that’s exactly what had been said about MPLS just a few year ago
[13:55:56] <Tony Li_web_506> Jeff do you recall how many YEARS it took to get consistent behavior?
[13:56:03] Sandra Murphy_web_112 leaves the room
[13:56:07] Sandra Murphy_web_798 joins the room
[13:56:28] Fengwei Qin_web_907 leaves the room
[13:56:34] Fengwei Qin_web_784 joins the room
[13:56:51] Andy Malis_web_832 leaves the room
[13:56:56] <Jeff Tantsura_web_282> @tony - yes, still PTSDed
[13:56:57] Andy Malis_web_457 joins the room
[13:57:11] <Zhaohui Zhang_web_175> @Zhenbin I suppose it is less challenging than NewIP :-)
[13:57:19] <Jeff Tantsura_web_282> @boris https://www.iana.org/assignments/mpls-label-values/mpls-label-values.xhtml
[13:57:24] <Tony Li_web_506> boiling the ocean == unbounded complexity == infinite bugs
[13:57:39] <Kireeti Kompella_web_303> for stewart's (2), I would like a time horizon
[13:57:46] <dhruvdhody> concussions (?) :-D
[13:57:51] <Jeff Tantsura_web_282> == job security :)
[13:58:07] <Yaakov Stein_web_913> Why can't we forget about 2 and use a new EtherType?
[13:58:30] <Adrian Farrel_web_734> @Yaakov A very valid quesiton
[13:58:33] <Lou Berger_web_618> @yakov interesting point.
[13:58:33] Geng-Da Tsai_web_352 leaves the room
[13:58:42] <John Scudder_web_552> Makes insertion challenging
[13:58:47] Xufeng Liu_web_536 leaves the room
[13:58:50] <Lou Berger_web_618> Some technologies don't have ethertype, but those can be handled too...
[13:58:51] <Mach Chen_web_837> @yaakov, MPLS version 2 :-)
[13:59:08] <Yaakov Stein_web_913> @Lou - or UPI or whatever used to determine it IS MPLS
[13:59:10] <Kireeti Kompella_web_303> NOT 8
[13:59:17] <Jeff Tantsura_web_282> Yes to DT and bi-weekly interims
[13:59:20] <Boris Khasanov_web_698> @Jeff, I know those standard numbers for SPL. I meant more extended way to differentiate them
[14:00:04] <Jeff Tantsura_web_282> @boris, why? The context is the value
[14:00:14] Luis Contreras_web_764 leaves the room
[14:00:16] Eric Gray_web_327 leaves the room
[14:00:45] <Boris Khasanov_web_698> Would be good to see more practical proposals about  coordination between WGs to handle those different proposals... Like mentioned DT and some inter-WG meetings
[14:00:50] Benjamin Steinert_web_171 leaves the room
[14:00:53] <Yaakov Stein_web_913> OK, if using a different EtherType the discussion should be on arch-d not MPLS list
[14:00:58] LucAndré Burdet_web_773 leaves the room
[14:01:02] LucAndré Burdet_web_967 joins the room
[14:01:05] Ruixue Wang_web_988 leaves the room
[14:01:12] Ruixue Wang_web_443 joins the room
[14:01:54] <Adrian Farrel_web_734> Oh! Don't bury this on arch-d :-)
[14:01:59] james welch_web_725 leaves the room
[14:02:01] <Yaakov Stein_web_913> :)
[14:02:12] Ruixue Wang_web_443 leaves the room
[14:02:18] Ruixue Wang_web_689 joins the room
[14:02:58] Les Ginsberg_web_613 leaves the room
[14:03:01] Jingrong Xie_web_531 leaves the room
[14:03:07] Jingrong Xie_web_726 joins the room
[14:03:26] <Adrian Farrel_web_734> Stewart just described interim meetings
[14:03:33] <Kireeti Kompella_web_303> i would see the label stack discussion on MPLS; the stuff beyond BoS in PALS/DetNET
[14:03:35] <Lou Berger_web_618> It wasn't a formal design team
[14:03:53] <Kireeti Kompella_web_303> of course the same people could be in both
[14:03:54] <Boris Khasanov_web_698> @Jeff, I am quite confused - how to assign those range(s) for many different proposals which we saw today because pushing them into one seems unrealistic
[14:04:01] <Adrian Farrel_web_734> Kireeti +1
[14:04:11] <Lou Berger_web_618> the people that wanted to contribute showed up
[14:04:21] <Jeff Tantsura_web_282> +1
[14:04:22] <Lou Berger_web_618> so a self organizing DT ;-)
[14:04:35] Joel Halpern_web_862 leaves the room
[14:04:55] <Kireeti Kompella_web_303> so, why not a design team?
[14:04:59] VJ Shah_web_443 leaves the room
[14:05:43] <Kireeti Kompella_web_303> +1 to andy's sugg that we use just one list (MPLS)
[14:05:48] <Lou Berger_web_618> so that those who actually do the work can participate
[14:05:49] <Jeff Tantsura_web_282> @boris, 15/eSPL/label
[14:05:50] Matt Green_web_672 leaves the room
[14:05:50] Srihari Sangli_web_550 leaves the room
[14:05:55] Tarek Saad_web_906 leaves the room
[14:05:56] Balazs Varga_web_935 leaves the room
[14:05:56] Greg Mirsky_web_793 leaves the room
[14:05:57] Fengwei Qin_web_784 leaves the room
[14:05:57] Italo Busi_web_404 leaves the room
[14:05:57] Geir Egeland_web_349 leaves the room
[14:05:58] Adrian Farrel_web_734 leaves the room
[14:05:58] Daniam Henriques_web_412 leaves the room
[14:05:58] Ram Santhanakrishnan_web_640 joins the room
[14:05:59] Guangpeng Li_web_271 leaves the room
[14:05:59] Ville Hallivuori_web_746 leaves the room
[14:05:59] Lou Berger_web_618 leaves the room
[14:06:01] Yuji Tochio_web_411 leaves the room
[14:06:01] Kiran Makhijani_web_138 leaves the room
[14:06:01] Torunn Narvestad_web_322 leaves the room
[14:06:02] Shunsuke Homma_web_983 leaves the room
[14:06:02] Tomoko Nezu_web_382 leaves the room
[14:06:02] Yaakov Stein_web_913 leaves the room
[14:06:02] Tom Hill_web_121 leaves the room
[14:06:03] Martin Vigoureux_web_393 leaves the room
[14:06:03] Ethan Grossman_web_538 leaves the room
[14:06:03] Haoyu Song_web_658 leaves the room
[14:06:03] Zafar Ali_web_292 leaves the room
[14:06:04] Dave Sinicrope_web_963 leaves the room
[14:06:04] Deborah Brungard_web_206 leaves the room
[14:06:04] Andy Malis_web_457 leaves the room
[14:06:04] Ketan Talaulikar_web_804 leaves the room
[14:06:05] Uma Chunduri_web_610 leaves the room
[14:06:05] Himanshu Shah_web_876 leaves the room
[14:06:05] Mach Chen_web_837 leaves the room
[14:06:05] David Pelton_web_151 leaves the room
[14:06:05] Bill Fenner_web_358 leaves the room
[14:06:05] Xiao Min_web_870 leaves the room
[14:06:06] Vishnu Beeram_web_929 leaves the room
[14:06:06] Jingrong Xie_web_726 leaves the room
[14:06:07] Xueyan Song_web_555 leaves the room
[14:06:08] Bruno Decraene_web_498 leaves the room
[14:06:08] Nicolai Leymann_web_584 leaves the room
[14:06:08] Donald Eastlake_web_452 leaves the room
[14:06:08] Yisong Liu_web_982 leaves the room
[14:06:09] Takuya Miyasaka_web_335 leaves the room
[14:06:09] Jeong-dong Ryoo_web_529 leaves the room
[14:06:10] Richard Li_web_404 leaves the room
[14:06:13] Kireeti Kompella_web_303 leaves the room
[14:06:14] Stewart Bryant_web_419 leaves the room
[14:06:16] Ruixue Wang_web_689 leaves the room
[14:06:17] Jie Dong_web_409 leaves the room
[14:06:17] John Drake_web_833 leaves the room
[14:06:17] Liu Yao_web_277 leaves the room
[14:06:18] <Tony Li_web_506> @Kireeti talk offline...
[14:06:18] <Boris Khasanov_web_698> @Jeff, okay, it might work.
[14:06:20] Loa Andersson_web_691 leaves the room
[14:06:21] dee3@hot-chilli.net leaves the room
[14:06:21] Rakesh Gandhi_web_970 leaves the room
[14:06:28] Yeoncheol Ryoo_web_778 leaves the room
[14:06:30] Alessandro Toppi_web_269 leaves the room
[14:06:30] Zheng Zhang_web_898 leaves the room
[14:06:30] Toerless Eckert_web_998 leaves the room
[14:06:30] Tony Li_web_506 leaves the room
[14:06:30] Shunwan Zhuang_web_297 leaves the room
[14:06:30] Yingzhen Qu_web_418 leaves the room
[14:06:30] Boris Khasanov_web_698 leaves the room
[14:06:30] Junfeng Zhao_web_910 leaves the room
[14:06:30] Zhaohui Zhang_web_175 leaves the room
[14:06:30] Wim Henderickx_web_622 leaves the room
[14:06:30] Feng Yang_web_165 leaves the room
[14:06:30] John Scudder_web_552 leaves the room
[14:06:30] Sergey Fomin_web_867 leaves the room
[14:06:30] Ruediger Geib_web_942 leaves the room
[14:06:30] Jeff Tantsura_web_282 leaves the room
[14:06:30] Gyan Mishra_web_109 leaves the room
[14:06:30] Rick Taylor_web_302 leaves the room
[14:06:30] Nils Warnke_web_620 leaves the room
[14:06:30] Farni Boten_web_927 leaves the room
[14:06:30] Fan Yang_web_113 leaves the room
[14:06:30] Ran Chen_web_437 leaves the room
[14:06:30] Dhruv Dhody_web_595 leaves the room
[14:06:30] Praveen Mada_web_887 leaves the room
[14:06:30] Matthew Bocci_web_592 leaves the room
[14:06:30] Ahmed Bashandy_web_940 leaves the room
[14:06:30] Zongpeng Du_web_602 leaves the room
[14:06:30] Yahya_web_354 leaves the room
[14:06:30] LucAndré Burdet_web_967 leaves the room
[14:06:30] Merike Kaeo_web_656 leaves the room
[14:06:31] Sandra Murphy_web_798 leaves the room
[14:06:31] Weiqiang Cheng_web_226 leaves the room
[14:06:31] Ram Santhanakrishnan_web_640 leaves the room
[14:06:31] Pascal Thubert_web_337 leaves the room
[14:06:31] Kamran Raza_web_387 leaves the room
[14:06:31] Zhenbin Li_web_157 leaves the room
[14:06:31] Haibo Wang_web_688 leaves the room
[14:06:31] Yuehua Wei_web_934 leaves the room
[14:06:44] Meetecho leaves the room
[14:06:45] slm joins the room
[14:08:53] slm leaves the room
[14:20:47] dhruvdhody leaves the room
[14:27:41] LouBerger leaves the room
[15:04:33] agmalis@jabber.hot-chilli.net joins the room
[15:06:52] agmalis@jabber.hot-chilli.net leaves the room
[19:35:10] toerless@jabber.at leaves the room