[14:58:08] Rob Wilton joins the room
[14:59:32] <mcr> ?! we are the only ones?
[14:59:40] mcr has set the subject to: https://etherpad.ietf.org:9009/p/notes-ietf-interim-2020-anima-01
[15:04:04] EskoDijk joins the room
[15:05:55] EskoDijk leaves the room
[15:06:27] esko joins the room
[15:07:29] <mcr> morning.
[15:08:56] <Rob Wilton> Probably don't need many of us in Jabber :-)
[15:11:41] <mcr> 13.1.  Normative References
   [I-D.ietf-anima-autonomic-control-plane]
[15:14:35] <mcr> http://rootbeerbarrel.com/2008/01/14/jack-astors-brewski-root-beer/  and https://www.google.com/search?newwindow=1&tbm=isch&sxsrf=ALeKk02hHUIjm7tOvgsxMYohjq_TYLMv0g%3A1586445242921&source=hp&biw=1129&bih=972&ei=ujuPXsOiNcyGsAWu2r_4Cg&q=jack+astor+brewski&oq=jack+astor+brewski&gs_lcp=CgNpbWcQAzoFCAAQgwE6AggAOgQIABAeOgYIABAIEB46BAgAEBhKJQgXEiE0LTIxOGcyMDlnMzAxZzI0MmcyMTRnMjAyZzE4OGcxOTdKFQgYEhE0LTFnMWcxZzFnMWcxZzdnNFAPWOYdYLQfaAJwAHgAgAGqAogB8RqSAQUwLjguOZgBAKABAaoBC2d3cy13aXotaW1n&sclient=img&ved=0ahUKEwjD56CJ0dvoAhVMA6wKHS7tD68Q4dUDCAc&uact=5#imgrc=F5-Isni5-4odHM
[15:14:43] <mcr> I was going to buy a couple of cases to Vancouver.
[15:15:06] <mcr> http://www.root-beer.org/modules.php?name=Brands&rbop=Brand&bid=2250
[15:15:32] Toerless Eckert joins the room
[15:16:23] Eric Vyncke joins the room
[15:17:15] <Toerless Eckert> Eric: ACP is at end, other presentations going on now. Given how everybody is virtual, not many of the webex joined users on jabber...
[15:18:03] <Eric Vyncke> Indeed, jabber rooms are usually light during those interims
[15:18:35] <Eric Vyncke> Thank you for putting ACP at the end even if the IESG telechat is over, so, could do the ACP status at any time.
[15:18:47] <Eric Vyncke> Enjoying listening for updates on ANIMA :)
[15:19:16] <mcr> we did the brief verbal update already on BRSKI.  Also see links above for celebratory Root Beer.
[15:19:21] <mcr> why is jabber so uncool?
[15:19:58] <Eric Vyncke> @mcr: let's not start a XMPP discussion ;-) (on my side, I am fine with it)
[15:29:56] <Toerless Eckert> oh no, a second person that likes to say "yes and no"
[15:29:58] <Toerless Eckert> ;-)
[15:30:18] <Eric Vyncke> Quantum person ?
[15:30:18] <Toerless Eckert> mcr: can you check on the one preso missing ?
[15:30:34] <Toerless Eckert> eric: Steffen was just saying it, Sheng also likes to say it ;-)
[15:30:45] <Toerless Eckert> quaint habbit ;-)
[15:32:56] <mcr> Toerless, I see one one slide proposal still waiting for your attention.
[15:33:15] <mcr> Your proposed slides awaiting chair approval
Operational Considerations for BRSKI Registrar (2020-04-09 07:14 PDT)
[15:35:19] <Toerless Eckert> now i see it. done. thanks!
[15:35:37] <mcr> apparently, Jie couldn't upload more than 1MB.
[15:36:02] <Toerless Eckert> (head-banging)
[15:38:13] <mcr> great firewall issue, I think.
[15:38:44] <mcr> I noticed, last fall, that I could get interactive SSH through the firewall, but the moment I started to rsync (or git over SSH), that it slowed down to the point of being useless.
[15:38:59] <mcr> (the great firewall of china, that is)
[15:52:16] <mcr> higher gain, but not with mic against lips.  
[15:55:26] <Toerless Eckert> am i showing the right slides ?
[15:55:54] <mcr> yes, you are.
[15:55:58] <Toerless Eckert> ok.. now i am matching text and slides ;-)
[15:56:10] <mcr> there is just a big latency to transmit slide this way.
[15:56:36] <Toerless Eckert> i didn't dare first time around to upload slides to webex directly.
[15:56:47] <mcr> webex tried to do this business with sending slides out in smart format, but they picked PPT rather than PDF.  Really, the high-res slides should have been sent to all viewers.
[15:56:49] <Toerless Eckert> its a lot better, but nobody has done it in ietf so far, so i went the microsoft way
[15:56:55] <mcr> uploading slides is not impleemnted in webrtc version.
[15:57:05] <Toerless Eckert> oh..
[15:57:14] <Toerless Eckert> good to know. alas, good reason not to do it then *sigh*
[15:57:16] <mcr> nor is viewing uploaded slides.
[15:57:31] <mcr> that's why all chairs should use webrtc version, so that they know what it does/does not do.
[15:57:38] <Toerless Eckert> yeah.
[15:57:50] <Toerless Eckert> well, at least i have  a second notebook here to check
[15:58:00] <mcr> that was smart of you.
[15:59:08] <Eric Vyncke> I must be stupid but what would be the use of ACP if you have a SDN controller ?
[15:59:18] <mcr> damn, I was sure that I had numbered the slides.
[15:59:27] <mcr> Eric, so the SDN controller has to manage things in-band.
[15:59:32] <Toerless Eckert> Eric: i didn't have time to read, so i need to shut up ;-))
[15:59:42] <mcr> there aren't wires to every switch.
[15:59:48] <mcr> from the SDN controller.
[16:00:04] <mcr> the ACP *IS* the SDN channel.
[16:00:33] <Eric Vyncke> ack
[16:00:34] <mcr> @Eric, does this make sense to you?  I am very interested in your feedback here.
[16:00:37] <Toerless Eckert> just chatting on second notebook and the presentation notebook has a life of its own.
[16:00:54] <Eric Vyncke> I am failing to follow everything to be honest. I will need to read.
[16:01:17] <Eric Vyncke> AFAIK, LLDP is not forwarded by switches (and may be it is the reason why LLDP is used)
[16:01:45] <Toerless Eckert> if the use-case is still rfc8368, i am of course supporting the idea ;)
[16:01:56] <Toerless Eckert> the choice of protocol details is what i am not cleaar about right now
[16:02:42] <mcr> LLDP is an encapsulation for IPv6 (ACP), whose desireable feature is that it is neither L2 nor L3 forwarded, so it's always hop-by-hop, between switches.
[16:03:35] <Eric Vyncke> But this is linked to a specific MAC address AFAIK
[16:03:57] <Eric Vyncke> which could be used to send plain IPv6 packets
[16:04:08] <Eric Vyncke> using Ethertype as distinguished
[16:10:28] <Toerless Eckert> Eric: Norm Finn told me they (IEEE) are working on an LLDP hop-by-hop forwarded signaling protocol to do for TSN what e.g.: RSVP would do at L3, thats the only thing where i heard LLDP traffic "forwarded" across switches. But it may be an app on top of LLDP, not an LLDP extension.
[16:10:47] <Eric Vyncke> @mcr: for 'punting' the control plane, I wonder what the ASICs are doing: based on MAC address or the combo MAC address + ethertype ?
[16:11:04] <mcr> (but, we don't WANT LLDP forwarded.  We will decapsulate IPv6-LL, pull off IPsec, and then get IPv6 that we might forward)
[16:11:08] <Eric Vyncke> @toerless nice to know
[16:11:11] <Toerless Eckert> any switch supporting 802.1x will have MAC punt filters
[16:11:38] <mcr> @Eric, I don't know, but I'd be happy to chat with your ASIC designers.  I used to play an ASIC designer on TV, back in 2001.
[16:11:54] <Toerless Eckert> whole 8021x filtering architecture is on ethertype (simplified)
[16:12:26] <mcr> right, so I assumed it was all ethertype based, and I know that LLDP goes to the control plane already.
[16:13:31] <Toerless Eckert> but LLDP packets would very likely always be punted, so not a good choice if you wanted to have ACP hardware forwarded.
[16:13:54] <mcr> I, personally, don't want the ACP hardware forwarded. I think that's a major bug.
[16:14:07] <mcr> We disagree on that.
[16:14:17] <Toerless Eckert> SDN controllers need to send and even more receive a lot of bursty traffic
[16:14:49] <mcr> define "bursty" traffic in an era when control plane CPUs are quad-core Xeons with 64G of ram.
[16:16:29] <Toerless Eckert> its always fine to have devices that only do SW forwarding, but its impossible to do HW forwarding with wrong encap choices.
[16:16:37] <Toerless Eckert> DC switches have 4 core low end atoms
[16:16:59] <Toerless Eckert> in vendor implementation we could often not get more than 20 mbps through SW ipsec on routers
[16:18:37] <mcr> @Toerless, of course, different encapsulations can be used between routers, as long as the result is IPv6 ACP inside.
[16:20:05] <mcr> so, we have 1-2s latency on slide updates across the planet.
[16:20:06] <Toerless Eckert> NOw that we're getting rid of bufferbloat at L3/L4, we have it at L7
[16:20:15] <Toerless Eckert> i tink Jiu has > 10 seconds RTT
[16:20:36] <Toerless Eckert> i often get that effect too, so its not only china
[16:20:41] <Toerless Eckert> have to stop audio and restart
[16:20:55] <mcr> "click" was good.  Maybe say, "clicking to slide 3"
[16:21:02] <Toerless Eckert> yes
[16:21:16] <mcr> maybe remind speakers that there is a 2-5s delay on slide updates.
[16:21:21] <mcr> I also saw that delay, btw.
[16:21:22] <Toerless Eckert> process not embodied in low level firmware yet, still requires higher brain functions ;-)
[16:22:17] <Toerless Eckert> local two notebooks its 2 seconds.
[16:23:19] <mcr> Awesome real time re-programmable online human protocols.
[16:23:29] <Toerless Eckert> lousy memory
[16:33:11] <mcr> you probably lost the full-screen button on the browser, so it is doing continuous mode.
[16:33:22] <mcr> hover on the right, and you'll see the full-screen button.
[16:33:54] <Toerless Eckert> i tried what i thought is full screen button on mac, and it moved the window to a different desktop and i had to start the browser
[16:34:04] <mcr> that's the wrong key.
[16:34:16] <mcr> it's within the browser window, assuming that's chrome, which it looks to be.
[16:34:25] <Toerless Eckert> was that symbol full screen ?
[16:34:27] <mcr> you want the slide full screen within the browser.
[16:34:31] <mcr> yes!
[16:34:53] <Toerless Eckert> your fault if the universe implodes now ;-)
[16:35:28] <Toerless Eckert> you win!
[16:35:28] <mcr> it worked.
[16:35:33] <Toerless Eckert> thanks
[16:36:21] <mcr> this method, which I wish had a name, is apparently common in qualcom and many other ARM CPU vendors.
[16:36:40] <Toerless Eckert> which SDO was that Laurence slides from ?
[16:37:06] <mcr> they were presented at RATS/TEEP/SUIT virtual interim in Berlin in Feb.
[16:37:06] <Toerless Eckert> didn't we also hear that in the IETF105? ACME/ANIMA side meeting ?
[16:37:40] <Toerless Eckert> hah. please forward invites to such meetings to ANIMA next time, i missed that one.
[16:38:18] <mcr> slides at: https://siot-hackathon.github.io/agenda and https://siot-hackathon.github.io/slides/rats01.pdf, slides 32->
[16:40:08] <mcr> uhm, I'm not sure that the Berlin SUIT/TEEP/RATS Hackathon had ANIMA content, beyond that BRSKI can be used to convey attestations, but there is no ID for that yet, so neither ANIMA nor RATS have linked up yet.
[16:43:31] <Toerless Eckert> ack
[17:01:22] esko leaves the room
[17:01:26] <Toerless Eckert> that wasn't too shabby
[17:01:40] <mcr> I hope you like my limited notes.
[17:01:42] <Eric Vyncke> AND GOOD TIMING !
[17:01:54] <Toerless Eckert> surprise, yes
[17:02:04] <Toerless Eckert> thanks for attending even if just for one crucial uestion
[17:03:28] Toerless Eckert leaves the room
[17:07:16] Toerless Eckert joins the room
[17:13:29] mcr leaves the room
[18:15:45] Toerless Eckert leaves the room
[19:18:51] Toerless Eckert joins the room
[19:44:40] Toerless Eckert leaves the room
[20:13:57] Toerless Eckert joins the room
[20:14:01] Toerless Eckert leaves the room
[20:42:37] Eric Vyncke leaves the room
[21:01:43] Toerless Eckert joins the room