WEBVTT 00:00.000 --> 00:08.000 Okay, you almost made it. 00:08.000 --> 00:13.000 One of the last talks, as is tradition, we are trying to splice these a few jokes in 00:13.000 --> 00:14.000 also as is tradition. 00:14.000 --> 00:17.000 Some things are static and that's awesome. 00:17.000 --> 00:20.000 Without much further ado, let's get started. 00:20.000 --> 00:23.000 So I have a clicker. 00:23.000 --> 00:26.000 That modern technology. 00:26.000 --> 00:32.000 Yeah, so the core infrastructure is static as compared to last year, which is very good, 00:32.000 --> 00:33.000 because it means that's workforce. 00:33.000 --> 00:39.000 I mean, some details like new SSDs and services such, but by and large, everything is 00:39.000 --> 00:41.000 continuous to be static. 00:41.000 --> 00:45.000 We are still considering to rip out the core out of our next year. 00:45.000 --> 00:47.000 We'll see how it goes. 00:47.000 --> 00:54.000 But the nice thing about this is we don't have to spend endless nights on this, and 00:54.000 --> 00:56.000 at least not as of right now. 00:56.000 --> 00:58.000 We have found a place of stability. 00:58.000 --> 01:03.000 Maybe that's the first learning for anyone wanting to replicate a conference, 01:03.000 --> 01:09.000 work towards having stuff you can replicate and take out of a box next year. 01:09.000 --> 01:12.000 Again, don't have to restart from scratch every single time. 01:12.000 --> 01:17.000 It took us honestly too long to realize this kind of dynamic, but once we had it, 01:17.000 --> 01:20.000 we started saving massive amounts of time. 01:21.000 --> 01:22.000 Yeah, and there was a benefit. 01:22.000 --> 01:26.000 You get a lot of sleep back if you implement that, because this is something that was 01:26.000 --> 01:31.000 lacking from the last years, having actually time to recover and reach out to your personal 01:31.000 --> 01:35.000 batteries, because of dealing too much with infrastructure. 01:35.000 --> 01:40.000 That is needed to run such a conference at that scale. 01:40.000 --> 01:45.000 For this out of itself, it's basically just one big router. 01:45.000 --> 01:49.000 We have a handful of switches that are just distributing some 01:50.000 --> 01:54.000 major points, like you've seen the t-shirt sale before for the point of 01:54.000 --> 01:56.000 friendships, and we have some switches for that. 01:56.000 --> 02:00.000 And our core network that we will see later, we have three 02:00.000 --> 02:04.000 bigger switches from Arista that we just bought last year. 02:04.000 --> 02:09.000 They're only 12 years old, and now 13 years old, but they still do. 02:09.000 --> 02:10.000 Very good. 02:10.000 --> 02:16.000 We have some very, very old crappy service sitting in the data center at the 02:16.000 --> 02:20.000 ULB, but they will be thrown out hopefully anytime soon. 02:20.000 --> 02:25.000 But we also have now a cluster that makes this independent of other cloud 02:25.000 --> 02:29.000 service providers or cloud providers to run our workloads that are, 02:29.000 --> 02:34.000 yeah, basically most before the video processing that is done on 02:34.000 --> 02:39.000 side, and yeah, really happy to have made that switch last year. 02:39.000 --> 02:45.000 So we are very, very free to move this year and not rely on any third 02:45.000 --> 02:46.000 parties. 02:46.000 --> 02:48.000 We've changed our monitoring. 02:48.000 --> 02:49.000 It's now done. 02:49.000 --> 02:52.000 We are premises and low key and visualized by the finance. 02:52.000 --> 02:57.000 So we can easily share information with other parts of the team that is 02:57.000 --> 03:02.000 changing over the years, and yeah, we have also public dashboards. 03:02.000 --> 03:05.000 So you can see actually the status, if you go to dashboard that 03:05.000 --> 03:09.000 falls in the dark, you can see most of the stuff that we are seeing. 03:09.000 --> 03:12.000 Like, is the network saturated or not? 03:12.000 --> 03:14.000 Yeah, if you like, just go on it. 03:14.000 --> 03:20.000 It's in another nice version of all the data. 03:20.000 --> 03:24.000 For video, this is main part of what we are doing when we are building this 03:24.000 --> 03:25.000 conference. 03:25.000 --> 03:29.000 I'm not sure if you've seen it or recognized that we have in every room 03:29.000 --> 03:30.000 there's a camera. 03:30.000 --> 03:31.000 I'm just pointing at it. 03:31.000 --> 03:36.000 We have such lovely video boxes that are standing here. 03:36.000 --> 03:39.000 And they're mostly open source. 03:40.000 --> 03:43.000 The camera obviously is not, but the boxes. 03:43.000 --> 03:47.000 And we have everything that we build. 03:47.000 --> 03:49.000 You need to build for such a video setup. 03:49.000 --> 03:51.000 It's on GitHub. 03:51.000 --> 03:54.000 We have those boxes here that are sending all of the video 03:54.000 --> 03:57.000 streams to a rendering form that we have at the campus. 03:57.000 --> 04:01.000 We send these streams off-site for live streaming and distributed 04:01.000 --> 04:02.000 ourselves. 04:02.000 --> 04:07.000 We're not using anything like YouTube or any other big companies 04:07.000 --> 04:08.000 out there. 04:08.000 --> 04:12.000 And we also have our semi-automatic review and cutting process 04:12.000 --> 04:13.000 that we implement. 04:13.000 --> 04:17.000 So the speakers can do the cutting in the browser after they have 04:17.000 --> 04:18.000 their talk. 04:18.000 --> 04:22.000 And that's how we try to publish all of the plus 04:22.000 --> 04:26.000 thousand talks that are given at Fossdom at this year. 04:26.000 --> 04:29.000 I'm sure if you've seen the videos the years before, 04:29.000 --> 04:32.000 this is all done by the video infrastructure that we have here. 04:32.000 --> 04:35.000 And hopefully we will serve the purpose that we 04:35.000 --> 04:37.000 have to do. 04:37.000 --> 04:40.000 So I've got a bit into the video boxes. 04:40.000 --> 04:43.000 The years they have changed. 04:43.000 --> 04:44.000 Not sure if it. 04:44.000 --> 04:45.000 Yeah. 04:45.000 --> 04:47.000 We have some proprietary stuff going on here. 04:47.000 --> 04:50.000 We removed all of that in the last years because this 04:50.000 --> 04:52.000 was closed source. 04:52.000 --> 04:56.000 And we completely read it all of the video boxes. 04:56.000 --> 05:01.000 We started one and a half years ago to design the new 05:01.000 --> 05:04.000 box with all open components. 05:04.000 --> 05:06.000 And we attached the laptop to that. 05:06.000 --> 05:09.000 And then we had problems with the USB-C and firmware issues 05:09.000 --> 05:10.000 in USB-C. 05:10.000 --> 05:17.000 So we made a final video box with yet another component that was not fully 05:17.000 --> 05:20.000 open source, but helped us getting our path to that. 05:20.000 --> 05:23.000 And this was the final video box that we had for last year. 05:23.000 --> 05:27.000 It's basically its dump switch that we have here to connect everything 05:27.000 --> 05:28.000 together. 05:28.000 --> 05:32.000 And this is a USB hub. 05:32.000 --> 05:37.000 And this is basically something that will emit impact data that is 05:37.000 --> 05:40.000 grabbed from the HDMI sources. 05:40.000 --> 05:43.000 And this is the final video box, the final design. 05:43.000 --> 05:48.000 When you do your Thesis, there's always a final of something. 05:48.000 --> 05:51.000 V2.x or something like that. 05:51.000 --> 05:56.000 So for this year we have the next edition, which you can also see here. 05:56.000 --> 05:59.000 To tear down the ring, you can have a look at that. 05:59.000 --> 06:03.000 This is now the final design for this year. 06:03.000 --> 06:06.000 Probably it will get some enhancements in the future. 06:06.000 --> 06:10.000 My colleagues that are built this boxes are just giving a talk about how 06:10.000 --> 06:13.000 they built this boxes over at the age building. 06:13.000 --> 06:16.000 So if you run there, you will be too late because it talks already 06:16.000 --> 06:18.000 only 15 minutes. 06:18.000 --> 06:21.000 But basically everything inside there has been designed 06:21.000 --> 06:23.000 constantly for us. 06:23.000 --> 06:26.000 We've added some diagnosis on the front. 06:26.000 --> 06:28.000 And I'm really like to see that because when I'm speaking, 06:28.000 --> 06:32.000 I can see all of the VU meters going up and down. 06:32.000 --> 06:41.000 And we also made it easy to adapt to all of the usual audio technology 06:41.000 --> 06:42.000 that you have at a conference. 06:42.000 --> 06:44.000 We use XLR a lot. 06:44.000 --> 06:45.000 So that's also embedded. 06:45.000 --> 06:47.000 Something that we haven't had for years. 06:47.000 --> 06:50.000 Because we always needed to go through the XLR to the camera. 06:50.000 --> 06:51.000 That's back there. 06:51.000 --> 06:55.000 Which is quite shitty for all the folks that are running the video. 06:55.000 --> 06:56.000 Because they have to go a lot. 06:56.000 --> 06:58.000 I've had a lot in crowded rooms. 06:58.000 --> 06:59.000 This room is in the crowded. 06:59.000 --> 07:03.000 But if you have been to the one of the lower age rooms yesterday, 07:03.000 --> 07:06.000 for example, it's been as a volunteer that needs to take care 07:06.000 --> 07:08.000 that the batteries are charged. 07:08.000 --> 07:09.000 And so it's a nightmare. 07:09.000 --> 07:12.000 And so no everything at one place. 07:12.000 --> 07:14.000 And don't need to do with that. 07:14.000 --> 07:17.000 And not sure if you can see it if you zoom in. 07:17.000 --> 07:19.000 But these are all custom-made stuff. 07:19.000 --> 07:22.000 This is a switch that we designed. 07:22.000 --> 07:25.000 And you can see that logo here on the printed silk screen. 07:25.000 --> 07:28.000 It's a phosdom logo with a switch inside. 07:28.000 --> 07:32.000 And we have the same logo here, which is our USB component. 07:32.000 --> 07:35.000 That will do all the handling of the outside USB connections. 07:35.000 --> 07:41.000 Mainly it's used for charging all of those belt devices. 07:41.000 --> 07:44.000 And so you can charge it directly from the box now. 07:44.000 --> 07:47.000 And we have the little LCD screen. 07:47.000 --> 07:52.000 We have an audio mixer that is fully controllable via software. 07:52.000 --> 07:57.000 So every audio that gets in can now be managed from a central place. 07:57.000 --> 08:01.000 We have a red X4. 08:01.000 --> 08:03.000 I hope it pronounced it right. 08:03.000 --> 08:09.000 X86 CPU running here with Debian. 08:09.000 --> 08:11.000 That will manage everything. 08:11.000 --> 08:16.000 And we have the HDMI video grabber that we will hopefully replace 08:16.000 --> 08:19.000 in the version that we can also put in. 08:19.000 --> 08:21.000 Not only the laptop here, but also the camera. 08:21.000 --> 08:23.000 So have it all in one box. 08:23.000 --> 08:26.000 And don't need two of those boxes per room anymore. 08:26.000 --> 08:27.000 Yes. 08:27.000 --> 08:31.000 I'm not sure. 08:31.000 --> 08:34.000 I have a link to the talk that my colleagues are giving. 08:34.000 --> 08:37.000 And they're explaining everything they've done. 08:37.000 --> 08:40.000 Everything that's went inside there. 08:40.000 --> 08:43.000 We have the key cut. 08:43.000 --> 08:46.000 And then specific that you need and everything that you. 08:46.000 --> 08:49.000 If you want to build it yourself, you can have it. 08:49.000 --> 08:52.000 I don't have all the schematics in my head. 08:52.000 --> 08:53.000 I'm just. 08:53.000 --> 08:55.000 But they should all be available. 08:55.000 --> 09:00.000 If they're not file an issue, because open source and open hardware. 09:00.000 --> 09:03.000 Yeah. 09:03.000 --> 09:05.000 So the final video box. 09:05.000 --> 09:09.000 If you want to have more details like this guy in the front sitting in the front. 09:09.000 --> 09:13.000 There's a talk from marching in angle that actually designed all of that. 09:13.000 --> 09:14.000 The PCB design. 09:14.000 --> 09:20.000 And then getting it done at the manufacturer and running into the sweatshops. 09:20.000 --> 09:25.000 That we're soldering that over the weekends in December. 09:25.000 --> 09:28.000 They give a talk that's happening right now. 09:28.000 --> 09:31.000 There's the QR code to the talk. 09:31.000 --> 09:35.000 It will be released hopefully tomorrow when it's been cut. 09:35.000 --> 09:41.000 You can then see what actually went into that system. 09:41.000 --> 09:43.000 Running such a conference. 09:43.000 --> 09:45.000 You can think of, oh yeah. 09:45.000 --> 09:46.000 It's not that hard. 09:46.000 --> 09:47.000 You have a box here. 09:47.000 --> 09:48.000 You have a laptop. 09:48.000 --> 09:49.000 You have mains. 09:49.000 --> 09:50.000 And then you have a camera. 09:50.000 --> 09:52.000 It can't be that hard. 09:52.000 --> 09:59.000 Actually, it is that hard because you need to think of things like we all do in this in our free time. 09:59.000 --> 10:00.000 We're volunteers. 10:00.000 --> 10:01.000 We don't get paid for that. 10:01.000 --> 10:03.000 We don't have any paid stuff like at other conferences. 10:03.000 --> 10:06.000 So it needs to be bulletproof from a design perspective. 10:06.000 --> 10:09.000 That's why we have schematics for everything. 10:09.000 --> 10:11.000 Again, this is also on the GitHub. 10:11.000 --> 10:14.000 So you can clone it or notice. 10:14.000 --> 10:18.000 And this is what we hand out to our volunteers when they start on Friday in the morning. 10:18.000 --> 10:21.000 Building up all the rooms setting up the cameras. 10:21.000 --> 10:25.000 So they know what is connected where and what to set. 10:25.000 --> 10:32.000 So everything can be reproduced and set to a known state with a known working state as we have defined it. 10:33.000 --> 10:35.000 What's going on there? 10:35.000 --> 10:38.000 Is it more interesting? 10:38.000 --> 10:45.000 If you want to know this is me trying to meet with a member of the European Commission, or one of the offices of the European Commission, 10:45.000 --> 10:51.000 to get them to send a letter to you to thank them for fostering some sort of political support. 10:51.000 --> 10:52.000 Yes, it's important. 10:52.000 --> 10:55.000 So how's the whole conference managed? 10:55.000 --> 10:59.000 We have one big video operating center. 10:59.000 --> 11:01.000 It's not like minority reports. 11:01.000 --> 11:04.000 It's crappy HTML. 11:04.000 --> 11:05.000 But it works. 11:05.000 --> 11:07.000 You can see it. 11:07.000 --> 11:10.000 We have a preview of what's going on in every room. 11:10.000 --> 11:13.000 We have the audio levels that you can. 11:13.000 --> 11:14.000 Where do you see it? 11:14.000 --> 11:15.000 We have the audio levels. 11:15.000 --> 11:23.000 If they're working and we can all see this from the remote and then can maybe change it if it's not working as expected. 11:23.000 --> 11:26.000 So we have everything in the K building. 11:26.000 --> 11:27.000 We're sitting there. 11:28.000 --> 11:31.000 Volunteers that are actually listening into the streams. 11:31.000 --> 11:33.000 If the video signal is okay. 11:33.000 --> 11:35.000 If the audio signal is okay. 11:35.000 --> 11:37.000 It's somewhat synchronized with my lips. 11:37.000 --> 11:38.000 What I'm saying. 11:38.000 --> 11:43.000 So they can see everything from a central place and everyone here in the room. 11:43.000 --> 11:46.000 Not in the light and talk room, but in every deaf room. 11:46.000 --> 11:52.000 We have deaf room managers that get their room and they can control for example, which is shown in the video stream. 11:52.000 --> 11:56.000 And this is all handled by the deaf room managers. 11:56.000 --> 11:58.000 How does it look like? 11:58.000 --> 12:00.000 This is from yesterday. 12:00.000 --> 12:02.000 When you have such a room. 12:02.000 --> 12:04.000 They can actually see life. 12:04.000 --> 12:08.000 The view meters of the audio levels on this side. 12:08.000 --> 12:13.000 They could also temper them and just change things. 12:13.000 --> 12:17.000 They can set the output for the video stream and also for the recording. 12:17.000 --> 12:21.000 Should it be side by side presenter or should it be the slides that were shown. 12:21.000 --> 12:25.000 And all of those things and it's very, very basic HTML. 12:25.000 --> 12:27.000 So it work work on any device. 12:27.000 --> 12:29.000 You can also have it on a smartphone. 12:29.000 --> 12:37.000 I saw a lot of deaf room managers just using their smartphone to control the video stream, which is quite convenient for them. 12:37.000 --> 12:45.000 And this is actually how the conference is running and how the video is managed. 12:46.000 --> 12:55.000 All of these videos that are produced here during the conference are sent to a rendering farm, which is looking quite messy. 12:55.000 --> 12:57.000 If you look over here. 12:57.000 --> 12:58.000 This is our wreck. 12:58.000 --> 12:59.000 We have some service in there. 12:59.000 --> 13:03.000 We have a lot of cables and we've had a lot of laptops. 13:03.000 --> 13:09.000 Since we started adding more and more development rooms that are running at the same time, 13:09.000 --> 13:11.000 we needed to upgrade the laptop. 13:11.000 --> 13:15.000 That's why we have chosen to change the position. 13:15.000 --> 13:17.000 It's also better from the cooling perspective. 13:17.000 --> 13:23.000 And we have now placed for a switch that has bigger uplinks. 13:23.000 --> 13:28.000 That also is one of the core networks, which is where all of the rendering is going on. 13:28.000 --> 13:35.000 This year we have 32 laptops that are doing all of that. 13:35.000 --> 13:40.000 And yeah, basically it's quite fine. 13:40.000 --> 13:48.000 We still use the same Ikea racks to stack everything in cooling is better. 13:48.000 --> 13:51.000 Yeah, that's it basically. 13:51.000 --> 14:02.000 And from what I said before, like getting back sleep or not being too sleepy when arriving here and dealing with the conference. 14:02.000 --> 14:08.000 Like Richard said, when you have want to reach a level of stability, 14:08.000 --> 14:14.000 installation of the router is one of the core things that we need to do to maintain the access here. 14:14.000 --> 14:18.000 And also get the routes up and working. 14:18.000 --> 14:24.000 This has been always tricky because we build up everything here on Friday. 14:24.000 --> 14:30.000 We are allowed to open up the doors of the ULB on Friday in the morning, not earlier. 14:31.000 --> 14:35.000 And since most of the equipment is sitting in the rack anyways. 14:35.000 --> 14:45.000 And we try to pull out that pull forward this task to just come here and have hopefully everything working. 14:45.000 --> 14:49.000 And as you can see by the timeline, it has come better. 14:49.000 --> 14:55.000 It's a selection of the router and also of the routes that need to be propagated for that conference. 14:55.000 --> 15:00.000 And still, it has improved a lot. 15:00.000 --> 15:15.000 Now this year, I think it was the earliest possible date to do the on January 6 because we just get a slash 17 announced from right and see for a month that they're handing out for events. 15:15.000 --> 15:18.000 Oh, a sign, not handed out. 15:19.000 --> 15:26.000 But we can announce, we could announce it on January 6 and this was the earliest thing possible because we will just tear it down tomorrow. 15:26.000 --> 15:34.000 And yeah, this was actually great because we had a lot of time to work on everything and see if it's still working like last year. 15:34.000 --> 15:43.000 And also from the network that was up and running and we've had it the first time that we had it on Wednesday already up and running. 15:43.000 --> 15:49.000 This is due to me being dump and ordering the wrong SSDs and we need to swap out the SSDs. 15:49.000 --> 15:56.000 The consumer grade SSDs to enterprise SSDs in the servers and we needed to do that before the first event. 15:56.000 --> 16:02.000 So it was on Wednesday so we could actually bring up the SSDs here and see if everything's working. 16:03.000 --> 16:22.000 Later pickups on the network were fixed due to some persons unplugging cables in wild situation other situations but it was quite okay and no one needed to stay up for a long and everyone got enough sleep. 16:23.000 --> 16:30.000 The monitoring of everything which helped us to stay sane to see if everything's in the desired state. 16:30.000 --> 16:37.000 We have basically not changed it since 2023 and just kept it running. 16:37.000 --> 16:47.000 We just adapt to maybe a new host name or a new damn distance spun up and this really helped us controlling if everything's in the desired state. 16:47.000 --> 16:52.000 And the video this year was a bit late, 2023. 16:52.000 --> 16:55.000 This is when all of the video cameras have been tested. 16:55.000 --> 17:00.000 We know that every data point that we want to see in our monitoring systems is there. 17:00.000 --> 17:05.000 It should have been done by five o'clock this year that was the plan. 17:05.000 --> 17:12.000 They had a bit of issues with one room because we always tend to break the switches. 17:12.000 --> 17:21.000 No has it break us in the power distributions because we overload them and that cost us roughly two hours. 17:21.000 --> 17:36.000 But we still we still okay with the time of 2023 because no one needed to stay here for long and the securities would like us to have here. 17:36.000 --> 17:41.000 Use more bandwidth and this is where it comes to our sponsors. 17:41.000 --> 17:49.000 So maybe to to somewhere is one of the reasons why we focused on the video box so much right now. 17:49.000 --> 17:54.000 It's because this was the only major component which we actually changed out. 17:54.000 --> 17:59.000 The rest again try and make something which you can reuse and just keep running ideally. 17:59.000 --> 18:03.000 To lower your workload if you try to do something like this yourself. 18:03.000 --> 18:14.000 Also be aware that lots of things where we just have a line have literally a dozen people or several dozen people working on things continuously like you just get the high level overview here. 18:14.000 --> 18:17.000 So I mean yes use more bandwidth. 18:17.000 --> 18:23.000 We deliberately don't provide any wired access anymore because you don't really need it because everyone has Wi-Fi. 18:23.000 --> 18:32.000 But still if you want to do some weird experiment if you want to run this one server out of your pocket during the thing and make a nice tip out which it replies. 18:32.000 --> 18:36.000 Actually here I forgot the one for 2012. 18:36.000 --> 18:41.000 Yeah just just to it like use more stuff at the same time. 18:44.000 --> 18:51.000 Yeah those are the two sponsors who we like to to think because without them we wouldn't have any internet here. 18:51.000 --> 18:55.000 On the topic of sponsors I don't know if you have this right now or not. 18:56.000 --> 19:03.000 Okay one way. On the topic of sponsors this is also something which I'm going to say in the closing keynote in a few minutes. 19:03.000 --> 19:10.000 We are currently at a fragile balance where it comes to our finances. 19:10.000 --> 19:17.000 So if any of you work at companies might potentially want to sponsor us for them in the future. 19:17.000 --> 19:24.000 Your internal marketing department will roughly close the planning for next year towards middle of this year. 19:24.000 --> 19:28.000 That's when they roughly stop their financial planning. 19:28.000 --> 19:43.000 So if any of you find yourself into a position where you can talk to your boss, you're to your boss's boss to whatever that might also be good because we have a pulsating red zero roughly for this year. 19:43.000 --> 19:47.000 And it's not a super comfortable place to be in. 19:47.000 --> 19:53.000 Again, I'm also going to repeat this during the closing keynote because it's a little bit of sounding of the click rules. 19:53.000 --> 19:58.000 But we will still do our next foster next year I hope. 19:58.000 --> 20:01.000 Yeah I think we are. 20:01.000 --> 20:05.000 If you want to clone first them we highly encourage you to. 20:05.000 --> 20:11.000 For those who don't know for Asia some years ago started working with us that we helped them. 20:11.000 --> 20:15.000 We are also going to try to help. 20:15.000 --> 20:19.000 I forgot the complete name of the Indian but it's like the little largest Indian conference. 20:19.000 --> 20:26.000 We are also trying to to support them to help and kickstart more of a of a repeatable process. 20:26.000 --> 20:33.000 Kind of thing Bulgaria is also using for open fest like we want to to and aid others to run their own events. 20:33.000 --> 20:36.000 Their meetups, their local conferences, their whatever. 20:36.000 --> 20:40.000 And not have to reinvent everything from scratch. 20:40.000 --> 20:49.000 We actively invite you to take our stuff and just run with it without you having to do all the heavy lifting which we did over the years. 20:49.000 --> 20:53.000 And now if anyone has questions I'm going to see this microphone. 20:53.000 --> 20:57.000 And we're going to pay help or take it with this one. 20:57.000 --> 21:26.000 All right I have seen multiple infrastructure reviews from the past years and there was never no mention of how the streaming to the world is actually handled because I suppose you guys are not sending the streams from here from the ULB to the clients that are requesting them that you have some sort of CDN or something. 21:26.000 --> 21:28.000 How does that work? 21:28.000 --> 21:30.000 Yeah excellent question thanks. 21:30.000 --> 21:38.000 Actually we rent out some VMs that cloud providers like Hatsner that are acting as the endpoints. 21:38.000 --> 21:42.000 When you see stream.phosdom.org, pointing you to some of those instances. 21:42.000 --> 21:48.000 Basically it's just an engine axe that's relaying the stream that's coming from from the rendering farm. 21:48.000 --> 21:54.000 So if you're requesting a file, usually it will go through one of those VMs at Hatsner. 21:54.000 --> 22:02.000 And that will hit the back end on our rendering farm and since the streaming mechanisms allow some sort of caching at the engine axe is on the thing. 22:02.000 --> 22:08.000 So that's how we actually streaming through the world. 22:08.000 --> 22:15.000 There's another question. 22:15.000 --> 22:21.000 Can you determine from your data how many people have been here? 22:21.000 --> 22:25.000 Sorry can you repeat and put the menu in front of your mouth? 22:25.000 --> 22:31.000 Can you determine from the data you've got how many people have been here roughly? 22:31.000 --> 22:37.000 Not really ever since privacy extensions have started to come and effect. 22:37.000 --> 22:40.000 We stopped having any good numbers. 22:40.000 --> 22:46.000 10k plus is an easy, very, very comfortable guess but we don't have any exact numbers. 22:46.000 --> 22:56.000 It's a very deliberate choice to not require tickets, to not require registration, to not require payment or anything, to not require anything. 22:56.000 --> 23:04.000 And one of the direct effects which we take into account is basically our inability to count total amount of people. 23:04.000 --> 23:08.000 But we're basically filling as you probably noticed your B2 capacity. 23:08.000 --> 23:12.000 What we can tell you just from looking on the hallways. 23:12.000 --> 23:19.000 Last year we were pretty much at pre-pandemic levels but this year is the largest Boston we ever had. 23:19.000 --> 23:36.000 In previous years you sold the laptops that are used in the render farm and blind up this year. 23:36.000 --> 23:41.000 You are a person, probably 10,400. 23:41.000 --> 23:43.000 I know I get the question. 23:43.000 --> 23:45.000 It's a fair question to ask. 23:45.000 --> 23:49.000 For those who don't know the laptops you see in the rendering farm photos, we used to sell it. 23:49.000 --> 23:59.000 We used to bulk by them from refurbishment companies and then reselled them during forced them and then just hand them out for vouchers like you bought a voucher. 23:59.000 --> 24:06.000 And during carried on you literally just got a laptop from the office or the rendering farm or whatever and you just took it and was yours. 24:06.000 --> 24:15.000 We have issues with how the laptops are being charged and how this integrates with both the rendering farm distribution and with the video boxes. 24:15.000 --> 24:24.000 So for now we have decided that a the laptops are powerful enough and b to not have complications with how we run this thing at scale. 24:24.000 --> 24:34.000 It's better to just keep them. But you're far from the only person to ask and it was always super nice to be able to just cycle good laptops into hands of people cheaply. 24:34.000 --> 24:39.000 But for the foreseeable future we will keep them and store them. 24:39.000 --> 24:45.000 We would much prefer to not store them. We much prefer to just sell them off. There was the initial reason because we don't want to store them. 24:45.000 --> 24:48.000 But we kind of have to know. 24:54.000 --> 25:05.000 And we have to share the infrastructure here here at the U.B. 25:05.000 --> 25:16.000 I've proved sometimes troublesome with power, water, etc., water and any major issues this year. 25:16.000 --> 25:24.000 Yes, we managed to successfully break some power breakers especially in the network operation center. 25:24.000 --> 25:36.000 So we're flying blind there because it's like a routine for us to break it every year and we forget over the year and then we are being reminded of that. 25:36.000 --> 25:52.000 And so we have two power distributions in our knock and we're always using the wrong one because there's a special one with more I think 64 amps that we could use and we are still using just the power sockets on the wall. 25:52.000 --> 25:58.000 And we had maybe we learned from that as well. We should write it down. 25:58.000 --> 26:04.000 But we had some rooms also that have issues that they overloaded the circuit. 26:04.000 --> 26:19.000 But nothing like last year where some random person needed to debug power issues due to a failed device and it was constantly standing at the power break and switching it back on to find which device it was. 26:19.000 --> 26:26.000 So nothing has come to my mind. I did it with a broom cause it was high up. 26:26.000 --> 26:40.000 So on the power we are actually purchasing proper power cabling and power sub distributions for next year so that should hopefully be handled as a fun fact which was not really infrastructure but very very exciting on Saturday morning. 26:40.000 --> 26:51.000 We set up in the wrong room and there was a math class going on and they just ignored all the weird cabling and cameras and did their thing so we very very quickly 26:51.000 --> 27:01.000 Toward on this room while they were giving them a math class and build up in the room next door where we were supposed to be in. 27:01.000 --> 27:12.000 So do you have any good statistics about lengths of cables, number of access points, average data, amounts used across the campus? 27:12.000 --> 27:17.000 We have good statistics but not on what you ask. 27:18.000 --> 27:33.000 We actually now do have per room setup instructions with we know which cable length we need to run on here at the wall that you see this like a super Mario level in 2D. 27:33.000 --> 27:41.000 But we actually have those instructions it's also on GitHub to see what every room needs for cabling. 27:41.000 --> 27:50.000 We don't have that actually for our network setup because everything is cable and it's not being torn down and after event. 27:50.000 --> 28:04.000 And for the USB part for the access point and such I mean in previous years we actually built out massive amounts of infrastructure for those who have been around for longer you will remember all of those struts in particular in John saw over there. 28:04.000 --> 28:17.000 We have basically built out a complete conference and back then Cisco even used this as a test point and low test for their follow commercial hardware because so many people it's so much weird stuff at the same time. 28:17.000 --> 28:23.000 So this was actually like a thing which which the research and development of Cisco appreciated. 28:23.000 --> 28:36.000 These days it's also it's all just commercial stuff which which you will be pre installs and we get to use it's a wild mix of Cisco extreme networks and this year also your big team. 28:36.000 --> 28:51.000 So if you look at the ceiling you can see there's one lonely Cisco access point and in front of it there's an extreme network access point there's an extreme there's an extreme and there's an extreme so we got way more density which is that. 28:51.000 --> 29:03.000 And there's another extreme and we've we know that the ULB is redoing their infrastructure there they are equipping the rooms with more and more access points. 29:03.000 --> 29:26.000 I'm not sure if you heard the 25 years talk today some of the buildings here are around 100 years old over 100 years old and they are there keeping up with the infrastructure like this year in Charleston the biggest room we didn't need to bring our own Wi-Fi which we've done for years because they installed a lot of high density access points that we can just use they just. 29:26.000 --> 29:39.000 We just get a VLAN that we push over to them and they just push our SSID to their controller and then all of a sudden we have Wi-Fi working here so this is one thing that we don't need to take care. 29:39.000 --> 29:46.000 Because you'll be is now handing that for us and that's really nice because we don't have to deal with all that. 29:46.000 --> 29:54.000 There's another question somewhere in the third and in front there's also a question. 29:54.000 --> 29:59.000 Yeah. 29:59.000 --> 30:06.000 Do you guys do dedicated network cable runs to each room or are you using existing cable? 30:07.000 --> 30:22.000 We guys and girls do like largely these days it's it's static and we pay you will be for the use of the premises like it's below market rate, way below market rate which is also why we think them as a sponsor. 30:22.000 --> 30:35.000 But over the years we sometimes get a little bit of say of where the money is being used and network cabling is usually our top ask and over the years they started building out stuff more and more. 30:35.000 --> 30:41.000 So the amount of cable which we have to run is going down and down. 30:42.000 --> 30:59.000 So for long time for them visitors they would have known that there were cables running outside running in network and we would have network from this device patched via wild just fresh installed cat five cables that we just crimped. 30:59.000 --> 31:21.000 We can get to the university land they just put our VLANs here in front of the front desk so we can just plug ourselves in and after the event they just use the normal VLAN on that port and this is also reducing friction that we have when you crimp a network cable it's not working and you start debugging things like that so. 31:21.000 --> 31:25.000 There's one asking in the front row. 31:25.000 --> 31:30.000 First thank you very much put together this amazing event and everyone who's volunteered and spoken at it. 31:30.000 --> 31:41.000 I'm just wondering how many people are in the knuck or like in the do you need do you have like a dedicated team of people who actually manage network infrastructure at the back end side and if so how many. 31:41.000 --> 31:45.000 You're just looking at them. 31:45.000 --> 31:52.000 Well, thank you. 31:52.000 --> 32:11.000 We have we have another two folks Peter and Walter that hopefully can see us and we have a new join volunteer our on all who will hopefully assist us also he used to work here at the you'll be before so he might be also beneficial for for helping out here. 32:11.000 --> 32:18.000 I'm basically most of the active work routing so it's done for us and yeah. 32:18.000 --> 32:29.000 But they have we worked done by the you'll be because they provide the wireless network nowadays and that's really really something we really. 32:29.000 --> 32:37.000 Yeah, we can just can't be any more grateful that they just take all of that. 32:37.000 --> 32:40.000 Anyone else? 32:41.000 --> 32:45.000 Any more questions? 32:45.000 --> 32:46.000 No? 32:46.000 --> 32:48.000 Going once twice okay. 32:48.000 --> 32:50.000 Thank you very much. 32:50.000 --> 32:53.000 Thank you very much.