0 00:00:00,000 --> 00:00:30,000 Dear viewer, these subtitles were generated by a machine via the service Trint and therefore are (very) buggy. If you are capable, please help us to create good quality subtitles: https://c3subtitles.de/talk/637 Thanks! 1 00:00:14,990 --> 00:00:17,389 So to our next talk, 2 00:00:17,390 --> 00:00:20,359 imagine a world where information 3 00:00:20,360 --> 00:00:22,939 is spread across the whole globe, 4 00:00:22,940 --> 00:00:24,270 even to authorities. 5 00:00:25,310 --> 00:00:27,089 And that's for free. 6 00:00:27,090 --> 00:00:29,689 That's a company that does this, 7 00:00:29,690 --> 00:00:30,919 that's called Ultranet. 8 00:00:30,920 --> 00:00:32,839 They do it through satellites. 9 00:00:32,840 --> 00:00:35,239 Unfortunately, part of their distribution 10 00:00:35,240 --> 00:00:37,039 channel is closed source. 11 00:00:37,040 --> 00:00:39,289 So Danielle Estévez 12 00:00:39,290 --> 00:00:41,779 reverse engineered the system 13 00:00:41,780 --> 00:00:44,269 to signal the protocol. 14 00:00:44,270 --> 00:00:46,399 And the receiver, 15 00:00:46,400 --> 00:00:48,529 please welcome with an anniversary 16 00:00:48,530 --> 00:00:51,049 edition of Plus Daniel Aceves. 17 00:00:58,200 --> 00:00:59,839 Or is it working? 18 00:00:59,840 --> 00:01:01,979 Yeah, so thank you very much for the 19 00:01:01,980 --> 00:01:04,169 introduction. I would also like to thank 20 00:01:04,170 --> 00:01:06,989 the organizers are 21 00:01:06,990 --> 00:01:07,990 on. Let's go. 22 00:01:09,540 --> 00:01:12,089 So first of all, a brief introduction 23 00:01:12,090 --> 00:01:14,069 about myself, because it's the first time 24 00:01:14,070 --> 00:01:16,559 I speak into this conference, so 25 00:01:16,560 --> 00:01:18,869 I come from mathematics, I'm finishing my 26 00:01:18,870 --> 00:01:20,639 Ph.D. It's pure mathematics. 27 00:01:20,640 --> 00:01:22,409 It doesn't have anything to do with 28 00:01:22,410 --> 00:01:24,959 AlterNet or radio stuff. 29 00:01:24,960 --> 00:01:26,849 But I also have a background in computer 30 00:01:26,850 --> 00:01:27,899 science. 31 00:01:27,900 --> 00:01:29,969 And since 2.5 32 00:01:29,970 --> 00:01:32,189 years ago, I took amateur radio 33 00:01:32,190 --> 00:01:34,529 as a hobby, and that's 34 00:01:34,530 --> 00:01:36,779 my Spanish callsign for anyone into 35 00:01:36,780 --> 00:01:37,829 amateur radio. 36 00:01:37,830 --> 00:01:40,019 And I also have a UK callsign 37 00:01:40,020 --> 00:01:41,999 because I've been living in the UK for a 38 00:01:42,000 --> 00:01:43,329 while. 39 00:01:43,330 --> 00:01:45,449 So all this is 40 00:01:45,450 --> 00:01:46,949 related to the talk. 41 00:01:46,950 --> 00:01:49,559 In my amateur radio activities, 42 00:01:49,560 --> 00:01:51,929 I started to decode some 43 00:01:51,930 --> 00:01:54,269 satellites which are orbiting low 44 00:01:54,270 --> 00:01:56,669 earth and they transmit into 45 00:01:56,670 --> 00:01:58,409 amateur frequencies. 46 00:01:58,410 --> 00:02:00,659 So what's the deal 47 00:02:00,660 --> 00:02:01,829 behind this? 48 00:02:01,830 --> 00:02:04,109 In amateur radio, encryption 49 00:02:04,110 --> 00:02:05,279 is forbidden. 50 00:02:05,280 --> 00:02:07,289 All this is just mostly because a 51 00:02:07,290 --> 00:02:10,198 security measure and 52 00:02:10,199 --> 00:02:12,269 that's like part of 53 00:02:12,270 --> 00:02:14,489 the community, also because 54 00:02:14,490 --> 00:02:17,309 it's useful for amateur operators 55 00:02:17,310 --> 00:02:19,559 to just listen across the band and 56 00:02:19,560 --> 00:02:21,749 listening to whoever speaking, 57 00:02:21,750 --> 00:02:23,699 having, well, not so private 58 00:02:23,700 --> 00:02:25,829 conversations, but it's 59 00:02:25,830 --> 00:02:27,929 common to go on and listen, 60 00:02:27,930 --> 00:02:30,359 listen, listen and decode whatever 61 00:02:30,360 --> 00:02:31,409 you want. 62 00:02:31,410 --> 00:02:33,389 So that's kind of the amateur radio 63 00:02:33,390 --> 00:02:34,799 philosophy. 64 00:02:34,800 --> 00:02:37,199 And there are these small satellites, 65 00:02:37,200 --> 00:02:40,079 most of them all launched by universities 66 00:02:40,080 --> 00:02:41,999 or educational projects. 67 00:02:42,000 --> 00:02:44,429 They transmit some data telemetry 68 00:02:44,430 --> 00:02:45,569 mostly. 69 00:02:45,570 --> 00:02:47,829 And the thing is that 70 00:02:47,830 --> 00:02:49,799 while these specifications for how to 71 00:02:49,800 --> 00:02:52,439 decode the telemetry, 72 00:02:52,440 --> 00:02:54,839 many times they they release 73 00:02:54,840 --> 00:02:56,879 incomplete specifications. 74 00:02:56,880 --> 00:02:59,159 So you have to 75 00:02:59,160 --> 00:03:01,449 to do some work in order to get 76 00:03:01,450 --> 00:03:02,999 a working decoder. 77 00:03:03,000 --> 00:03:05,399 So I have been doing this for 78 00:03:05,400 --> 00:03:07,619 like two years ago and 79 00:03:07,620 --> 00:03:09,719 that gave me some experience when 80 00:03:09,720 --> 00:03:12,509 I started working with AlterNet. 81 00:03:12,510 --> 00:03:15,119 So enough of the introduction. 82 00:03:15,120 --> 00:03:17,369 Let's first go with what is all 83 00:03:17,370 --> 00:03:18,569 about Oughton, that 84 00:03:19,650 --> 00:03:22,139 essentially you have a geostationary 85 00:03:22,140 --> 00:03:24,389 satellite, so it's fixed in 86 00:03:24,390 --> 00:03:26,369 one part of the sky as you look at it 87 00:03:26,370 --> 00:03:29,489 from Earth, and that's transmitting 88 00:03:29,490 --> 00:03:30,939 files from the Internet. 89 00:03:30,940 --> 00:03:33,149 I think Wikipedia files 90 00:03:33,150 --> 00:03:36,239 or weather information or 91 00:03:36,240 --> 00:03:38,309 or miscellaneous information, small 92 00:03:38,310 --> 00:03:40,439 files, because the bandwidth 93 00:03:40,440 --> 00:03:42,419 is not so much. 94 00:03:42,420 --> 00:03:45,119 So then you have this signal and 95 00:03:45,120 --> 00:03:47,099 well, you have these the from the 96 00:03:47,100 --> 00:03:49,859 alternate, but the specifications 97 00:03:49,860 --> 00:03:50,789 are not open. 98 00:03:50,790 --> 00:03:52,949 So we want to construct a fully working 99 00:03:52,950 --> 00:03:53,939 decoder. 100 00:03:53,940 --> 00:03:56,129 And I'm going to explain how it works 101 00:03:56,130 --> 00:03:57,119 in two parts. 102 00:03:57,120 --> 00:03:59,429 First, we have our RF signal and 103 00:03:59,430 --> 00:04:01,349 we're going to turn that into bits or 104 00:04:01,350 --> 00:04:03,119 more precisely, frames. 105 00:04:03,120 --> 00:04:05,399 And then what do we do with this 106 00:04:05,400 --> 00:04:07,529 frames? We want to turn those into 107 00:04:07,530 --> 00:04:09,839 files and after that, some 108 00:04:09,840 --> 00:04:10,840 miscellaneous. 109 00:04:12,990 --> 00:04:15,599 OK, so what is your thoughts on that? 110 00:04:15,600 --> 00:04:17,699 It's a start up company and 111 00:04:17,700 --> 00:04:19,889 its goal is, well, there are many 112 00:04:19,890 --> 00:04:22,139 places in the world where it's not so 113 00:04:22,140 --> 00:04:24,779 easy to get Internet access. 114 00:04:24,780 --> 00:04:27,119 And even if you could have 115 00:04:27,120 --> 00:04:29,309 a downlink, only access to some 116 00:04:29,310 --> 00:04:31,409 Internet content, that would 117 00:04:31,410 --> 00:04:32,369 be interesting. 118 00:04:32,370 --> 00:04:34,529 So picture yourself in the middle 119 00:04:34,530 --> 00:04:36,719 of Africa. It's impossible to get 120 00:04:36,720 --> 00:04:38,999 a proper Internet connection. 121 00:04:39,000 --> 00:04:40,559 The best thing you can achieve is 122 00:04:40,560 --> 00:04:42,959 satellite. But that's very expensive. 123 00:04:42,960 --> 00:04:45,329 What if you could just put 124 00:04:45,330 --> 00:04:47,609 it down, link only satellite receiver 125 00:04:47,610 --> 00:04:49,829 and start getting information 126 00:04:49,830 --> 00:04:51,989 from Wikipedia, newsworthy 127 00:04:51,990 --> 00:04:53,319 content and so on. 128 00:04:53,320 --> 00:04:55,379 So that's the goal of 129 00:04:55,380 --> 00:04:58,259 this company, which is called out on it. 130 00:04:58,260 --> 00:05:00,449 And they started broadcasting 131 00:05:00,450 --> 00:05:02,759 using digital television 132 00:05:02,760 --> 00:05:04,829 satellites so you could use 133 00:05:04,830 --> 00:05:07,229 all the same equipment 134 00:05:07,230 --> 00:05:09,419 that you would use in Europe, 135 00:05:09,420 --> 00:05:12,029 for instance, to receive 136 00:05:12,030 --> 00:05:14,339 digital TV from geostationary 137 00:05:14,340 --> 00:05:16,019 satellites. 138 00:05:16,020 --> 00:05:18,119 And indeed, you had this set top 139 00:05:18,120 --> 00:05:20,579 box. And this works because in 140 00:05:20,580 --> 00:05:22,889 DVD or DVD 141 00:05:22,890 --> 00:05:25,049 to it, usually you embed 142 00:05:25,050 --> 00:05:27,269 digital video, MBG 143 00:05:27,270 --> 00:05:29,369 or whatever, but you can embed all 144 00:05:29,370 --> 00:05:30,279 sorts of content. 145 00:05:30,280 --> 00:05:32,669 So the embedded or 146 00:05:32,670 --> 00:05:34,919 files into that are. 147 00:05:34,920 --> 00:05:37,229 But the problem with this scheme is that 148 00:05:37,230 --> 00:05:39,539 these satellites use beams 149 00:05:39,540 --> 00:05:41,429 which cover only a region. 150 00:05:41,430 --> 00:05:43,619 For instance, you could have Astra 151 00:05:43,620 --> 00:05:45,779 satellite and the beam only covers part 152 00:05:45,780 --> 00:05:47,099 of Europe. 153 00:05:47,100 --> 00:05:49,229 You have satellites broadcasting to 154 00:05:49,230 --> 00:05:51,329 the Middle East, but 155 00:05:51,330 --> 00:05:53,699 you have you need 156 00:05:53,700 --> 00:05:55,799 many, let's say, 10 or 20 157 00:05:55,800 --> 00:05:58,199 satellites like these with their beams 158 00:05:58,200 --> 00:06:00,449 covering all the world if you want 159 00:06:00,450 --> 00:06:02,849 to have worldwide access. 160 00:06:02,850 --> 00:06:04,949 So that was very difficult for them to 161 00:06:04,950 --> 00:06:06,449 achieve, if not impossible. 162 00:06:06,450 --> 00:06:09,299 And they switch it to Albon broadcast. 163 00:06:09,300 --> 00:06:11,609 So Albon is around one point five 164 00:06:11,610 --> 00:06:14,159 gigahertz, the same as GPS, 165 00:06:14,160 --> 00:06:15,269 more or less. 166 00:06:15,270 --> 00:06:18,449 And they now broadcast exclusively 167 00:06:18,450 --> 00:06:20,939 through el-Banna and they do it 168 00:06:20,940 --> 00:06:23,139 through three. Inmarsat satellite 169 00:06:23,140 --> 00:06:25,259 says this is a really large 170 00:06:25,260 --> 00:06:27,599 satellite company which 171 00:06:27,600 --> 00:06:30,059 also broadcast differential 172 00:06:30,060 --> 00:06:32,519 GPS information, 173 00:06:32,520 --> 00:06:34,919 information for air traffic 174 00:06:34,920 --> 00:06:36,149 and so on. 175 00:06:36,150 --> 00:06:38,459 So they have this satellite in orbit 176 00:06:38,460 --> 00:06:40,829 and it's just a matter of money 177 00:06:40,830 --> 00:06:42,989 to go to Inmarsat and say, well, I want 178 00:06:42,990 --> 00:06:44,819 a really small channel on your satellite 179 00:06:44,820 --> 00:06:46,679 to to broadcast my data. 180 00:06:46,680 --> 00:06:48,119 So that's what they did. 181 00:06:48,120 --> 00:06:50,519 And they have three satellites. 182 00:06:50,520 --> 00:06:52,709 One covers the Americas, the 183 00:06:52,710 --> 00:06:54,869 other one covers Europe and Africa on the 184 00:06:54,870 --> 00:06:57,599 old one conversation, the Pacific. 185 00:06:57,600 --> 00:07:00,419 So this is worldwide coverage except 186 00:07:00,420 --> 00:07:03,299 near the Poles where, of course, 187 00:07:03,300 --> 00:07:05,519 the geostationary satellites, which 188 00:07:05,520 --> 00:07:07,619 are on the equator, are almost 189 00:07:07,620 --> 00:07:09,689 all out of view, but almost 190 00:07:09,690 --> 00:07:11,819 worldwide coverage. 191 00:07:11,820 --> 00:07:14,759 What do you need to receive out on it? 192 00:07:14,760 --> 00:07:16,799 Let's first go with hardware. 193 00:07:16,800 --> 00:07:18,899 You need an 11 antenna, so 194 00:07:18,900 --> 00:07:21,269 typically you would use a patch 195 00:07:21,270 --> 00:07:23,729 antenna. This is like a square 196 00:07:23,730 --> 00:07:26,219 this big. You can also make it slightly 197 00:07:26,220 --> 00:07:28,559 smaller. But picture 198 00:07:28,560 --> 00:07:31,469 a flat square like this size 199 00:07:31,470 --> 00:07:32,999 and that works fine. 200 00:07:33,000 --> 00:07:35,099 But if you have a dish like 201 00:07:35,100 --> 00:07:37,349 a satellite TV dish, 202 00:07:37,350 --> 00:07:39,659 it's one meter 203 00:07:39,660 --> 00:07:41,909 diameter or 60 centimeters 204 00:07:41,910 --> 00:07:44,219 diameter. You can use that. 205 00:07:44,220 --> 00:07:46,439 They will give you a much, much better 206 00:07:46,440 --> 00:07:48,329 signal. There are now people using a dish 207 00:07:48,330 --> 00:07:51,089 and it's absolutely fantastic. 208 00:07:51,090 --> 00:07:53,639 You need an LNA that's a pre amplifier. 209 00:07:53,640 --> 00:07:55,709 So unless your 210 00:07:55,710 --> 00:07:57,809 radio receiver is just 211 00:07:57,810 --> 00:07:59,999 next to your antenna, 212 00:08:00,000 --> 00:08:02,129 you need some sort of amplification 213 00:08:02,130 --> 00:08:04,259 into the signal. And usually 214 00:08:04,260 --> 00:08:06,509 you do it with a small piece of hardware, 215 00:08:06,510 --> 00:08:08,579 which is called LNA and just 216 00:08:08,580 --> 00:08:11,669 fits between the antenna and the receiver 217 00:08:11,670 --> 00:08:12,930 for the receiver. 218 00:08:14,250 --> 00:08:16,829 You can use any as the receiver you 219 00:08:16,830 --> 00:08:19,019 want, but the software from Ultranet 220 00:08:19,020 --> 00:08:21,299 only supports this really popular 221 00:08:21,300 --> 00:08:23,009 artle as the are dongles. 222 00:08:23,010 --> 00:08:26,039 I think many of you in the audience 223 00:08:26,040 --> 00:08:28,829 have known about these 224 00:08:28,830 --> 00:08:30,419 artlessly are dongles. 225 00:08:30,420 --> 00:08:31,769 They're really popular on Chip. 226 00:08:33,110 --> 00:08:35,359 And of course, you could run 227 00:08:35,360 --> 00:08:37,489 the software in any computer 228 00:08:37,490 --> 00:08:39,829 you like, but the software from out 229 00:08:39,830 --> 00:08:42,129 on it is intended 230 00:08:42,130 --> 00:08:44,869 to work on a single board on computers 231 00:08:44,870 --> 00:08:46,969 such as the chip, which 232 00:08:46,970 --> 00:08:49,549 is five dollars or so, I think, 233 00:08:49,550 --> 00:08:51,179 or the Raspberry Pi three. 234 00:08:52,310 --> 00:08:54,199 So the idea is that you have a 235 00:08:54,200 --> 00:08:56,269 self-contained kit where you have these 236 00:08:56,270 --> 00:08:58,339 software running on 237 00:08:58,340 --> 00:09:00,529 a dedicated 238 00:09:00,530 --> 00:09:03,229 small computer and then 239 00:09:03,230 --> 00:09:04,559 it does all you want. 240 00:09:04,560 --> 00:09:06,649 So it gets the files from the satellite 241 00:09:06,650 --> 00:09:08,749 and then it acts as a Wi-Fi 242 00:09:08,750 --> 00:09:09,679 hotspot. 243 00:09:09,680 --> 00:09:11,809 You can log in to the Wi-Fi 244 00:09:11,810 --> 00:09:13,849 hotspot and copy the files to your 245 00:09:13,850 --> 00:09:16,009 devices or whatever you want. 246 00:09:16,010 --> 00:09:18,469 And doujin that sells these items 247 00:09:18,470 --> 00:09:20,449 as a kit from the Web page. 248 00:09:20,450 --> 00:09:22,699 Or you can go anywhere online and 249 00:09:22,700 --> 00:09:25,759 buy them because they were really 250 00:09:25,760 --> 00:09:28,009 inexpensive stuff 251 00:09:28,010 --> 00:09:29,209 for the software. 252 00:09:29,210 --> 00:09:31,309 There is these are X 253 00:09:31,310 --> 00:09:32,389 OS. 254 00:09:32,390 --> 00:09:34,759 It's really a Linux IMH image. 255 00:09:34,760 --> 00:09:36,919 You put it on your flashcard 256 00:09:36,920 --> 00:09:39,129 on your computer. 257 00:09:39,130 --> 00:09:41,769 And then he does all the decoding 258 00:09:41,770 --> 00:09:43,989 and as I said, you can log 259 00:09:43,990 --> 00:09:46,599 in via Wi-Fi hotspot 260 00:09:46,600 --> 00:09:48,609 and get your files on whatever. 261 00:09:48,610 --> 00:09:50,889 So most of the software in our 262 00:09:50,890 --> 00:09:53,169 service is open source. 263 00:09:53,170 --> 00:09:55,629 The only problem is that the key parts 264 00:09:55,630 --> 00:09:57,939 of the receiver, the most important 265 00:09:57,940 --> 00:10:00,369 parts that you need to get the files 266 00:10:00,370 --> 00:10:02,739 are closed source and distributed only 267 00:10:02,740 --> 00:10:03,740 as a binary. 268 00:10:04,660 --> 00:10:06,849 Actually, there is kind 269 00:10:06,850 --> 00:10:09,999 of an issue because this binary 270 00:10:10,000 --> 00:10:12,279 link, some GPL libraries, 271 00:10:12,280 --> 00:10:14,079 the ones that had to do with the 272 00:10:14,080 --> 00:10:15,759 Artlessly are a dongle. 273 00:10:15,760 --> 00:10:18,009 So these are GPL only, not L 274 00:10:18,010 --> 00:10:20,349 DPL and 275 00:10:20,350 --> 00:10:22,629 well, I'm not a lawyer, but I think 276 00:10:22,630 --> 00:10:24,879 it's possible GPL violation 277 00:10:24,880 --> 00:10:27,039 to link GPL libraries 278 00:10:27,040 --> 00:10:28,959 from closed source code. 279 00:10:28,960 --> 00:10:31,009 I told AlterNet about this. 280 00:10:31,010 --> 00:10:32,319 They don't seem to care. 281 00:10:32,320 --> 00:10:34,839 So anyway, and 282 00:10:34,840 --> 00:10:37,149 most important, the protocols, 283 00:10:37,150 --> 00:10:39,279 the modulation on the format of 284 00:10:39,280 --> 00:10:42,069 the signals of how everything works 285 00:10:42,070 --> 00:10:44,259 is secret. It's hidden behind 286 00:10:44,260 --> 00:10:46,330 these closed salts receiver. 287 00:10:48,730 --> 00:10:51,219 So why did I decide to 288 00:10:51,220 --> 00:10:53,439 reverse engineering out on that? 289 00:10:53,440 --> 00:10:55,509 Well, first of all, that 290 00:10:55,510 --> 00:10:57,969 idea of spreading into the content 291 00:10:57,970 --> 00:11:00,909 just via broadcast 292 00:11:00,910 --> 00:11:03,609 from satellites is really nice. 293 00:11:03,610 --> 00:11:06,249 But I don't think that using 294 00:11:06,250 --> 00:11:08,769 a secret protocol and closer's 295 00:11:08,770 --> 00:11:11,049 software is a good idea for that. 296 00:11:11,050 --> 00:11:13,509 You should really use open 297 00:11:13,510 --> 00:11:16,599 standards or open information, 298 00:11:16,600 --> 00:11:18,399 whatever you want to call it. 299 00:11:18,400 --> 00:11:21,189 Also, some amateur radio operators 300 00:11:21,190 --> 00:11:24,219 had started playing without net 301 00:11:24,220 --> 00:11:26,829 or it's really fun because, 302 00:11:26,830 --> 00:11:29,409 well, most real 303 00:11:29,410 --> 00:11:30,729 one point two gigahertz. 304 00:11:30,730 --> 00:11:32,679 It's one point to the nearest amateur 305 00:11:32,680 --> 00:11:33,279 band. 306 00:11:33,280 --> 00:11:36,009 So most one point two gigahertz 307 00:11:36,010 --> 00:11:37,689 equipment for amateur radio. 308 00:11:37,690 --> 00:11:38,829 It's expensive. 309 00:11:38,830 --> 00:11:40,929 It gets some money 310 00:11:40,930 --> 00:11:42,609 to get into that. 311 00:11:42,610 --> 00:11:44,829 But this is some very cheap 312 00:11:44,830 --> 00:11:47,049 way to start playing with microwaves. 313 00:11:47,050 --> 00:11:49,329 So amateur radio operators 314 00:11:49,330 --> 00:11:51,099 started with that. 315 00:11:51,100 --> 00:11:53,199 And as I said before, 316 00:11:53,200 --> 00:11:55,479 well, a close source and secret 317 00:11:55,480 --> 00:11:58,359 protocols are also very bad from 318 00:11:58,360 --> 00:12:00,519 a for amateur radio. 319 00:12:00,520 --> 00:12:02,739 So what do they have before 320 00:12:02,740 --> 00:12:04,809 starting what the 321 00:12:04,810 --> 00:12:06,969 receiver is like a black box 322 00:12:06,970 --> 00:12:09,219 where you have the binary running 323 00:12:09,220 --> 00:12:11,409 and the RF goes in and then 324 00:12:11,410 --> 00:12:13,689 magically you have files coming 325 00:12:13,690 --> 00:12:16,339 out of the other and so 326 00:12:16,340 --> 00:12:18,789 on. And, you know, this is figures 327 00:12:18,790 --> 00:12:20,889 from out on that, that the bitrate of 328 00:12:20,890 --> 00:12:23,229 the signal is two kilobits per second. 329 00:12:23,230 --> 00:12:25,029 So that's not so much. 330 00:12:25,030 --> 00:12:27,249 But if you are running for 331 00:12:27,250 --> 00:12:29,079 the whole day and you're expected to 332 00:12:29,080 --> 00:12:31,359 leave it running, like for many 333 00:12:31,360 --> 00:12:33,669 days, you get 20 megabytes 334 00:12:33,670 --> 00:12:34,670 of content per day. 335 00:12:35,980 --> 00:12:38,829 I also had access to these 336 00:12:38,830 --> 00:12:40,569 clotheshorse software. 337 00:12:40,570 --> 00:12:42,639 It's free to download, but 338 00:12:42,640 --> 00:12:44,349 it's clotheshorse. 339 00:12:44,350 --> 00:12:46,539 And I mostly looked at 340 00:12:46,540 --> 00:12:48,669 the AMD 64 341 00:12:48,670 --> 00:12:51,129 bit version of the software. 342 00:12:51,130 --> 00:12:53,259 It's an older version because now they've 343 00:12:53,260 --> 00:12:55,539 dropped support for 344 00:12:55,540 --> 00:12:57,789 Intel machines and now 345 00:12:57,790 --> 00:13:00,909 everything is only for ALM. 346 00:13:00,910 --> 00:13:03,219 So the clotheshorse receiver has 347 00:13:03,220 --> 00:13:06,219 two parts. One is called SDR 100. 348 00:13:06,220 --> 00:13:08,409 That's the SDR receiver, 349 00:13:08,410 --> 00:13:10,629 which talks to the Artlessly 350 00:13:10,630 --> 00:13:12,849 are dongle and gets 351 00:13:12,850 --> 00:13:15,009 the bits from the RF 352 00:13:15,010 --> 00:13:17,379 and then those bits get sent 353 00:13:17,380 --> 00:13:19,689 into something called own. 354 00:13:19,690 --> 00:13:22,389 That's a Sandeman which is running and 355 00:13:22,390 --> 00:13:24,939 that's all the work of reconstructing 356 00:13:24,940 --> 00:13:27,489 the files and so on. 357 00:13:27,490 --> 00:13:29,799 And I also had access to some 358 00:13:29,800 --> 00:13:32,079 AQ recordings by my 359 00:13:32,080 --> 00:13:34,659 fellow amateur operators, Cochairmen 360 00:13:34,660 --> 00:13:36,519 Kilo for Kilo, Delta, Romeo. 361 00:13:36,520 --> 00:13:39,019 So in fact I don't have myself 362 00:13:39,020 --> 00:13:40,659 a receiver for that. 363 00:13:40,660 --> 00:13:42,819 I could set up one very easily, but 364 00:13:42,820 --> 00:13:45,489 in these days it's much, much easier 365 00:13:45,490 --> 00:13:48,249 to go to some amateur operator 366 00:13:48,250 --> 00:13:50,589 who already has a disagreement on 367 00:13:50,590 --> 00:13:50,739 it. 368 00:13:50,740 --> 00:13:53,499 Well, can you give me an IQ recording? 369 00:13:53,500 --> 00:13:55,149 I want to play with it. 370 00:13:55,150 --> 00:13:57,399 And in a couple of days 371 00:13:57,400 --> 00:13:59,499 I had my IQ recordings and I could 372 00:13:59,500 --> 00:14:00,500 start working. 373 00:14:02,100 --> 00:14:03,330 So let's see how 374 00:14:04,620 --> 00:14:06,600 we get from RF to bits, 375 00:14:08,250 --> 00:14:10,349 when I look at an RF signal, I 376 00:14:10,350 --> 00:14:12,509 do it with my favorite as a star 377 00:14:12,510 --> 00:14:14,379 receiver. This is Leandra's. 378 00:14:14,380 --> 00:14:16,499 It's not very popular, but most 379 00:14:16,500 --> 00:14:18,599 of you have heard of is the R 380 00:14:18,600 --> 00:14:20,369 sharp or whatever. 381 00:14:20,370 --> 00:14:22,499 So this is the waterfall on 382 00:14:22,500 --> 00:14:23,639 the left hand side. 383 00:14:23,640 --> 00:14:25,259 I don't know if I have. 384 00:14:25,260 --> 00:14:26,969 Yes, the mouse cursor. 385 00:14:26,970 --> 00:14:29,819 This is the alternate signal. 386 00:14:29,820 --> 00:14:32,129 This is these are stronger 387 00:14:32,130 --> 00:14:34,829 signals which are I'm not sure 388 00:14:34,830 --> 00:14:37,259 either GPS or 389 00:14:38,310 --> 00:14:40,619 air traffic somewhat. 390 00:14:40,620 --> 00:14:43,409 And here you can see some pocket data. 391 00:14:43,410 --> 00:14:45,479 So this is the first you see 392 00:14:45,480 --> 00:14:48,329 just the waterfall 393 00:14:48,330 --> 00:14:50,159 and in the waterfall, the only thing we 394 00:14:50,160 --> 00:14:52,769 can see, apart from signal strength, 395 00:14:52,770 --> 00:14:55,199 which is OK in this recording, 396 00:14:55,200 --> 00:14:57,779 is that it's four point eight kilohertz 397 00:14:57,780 --> 00:15:00,029 wide and there's no 398 00:15:00,030 --> 00:15:01,829 FITA, there's no pattern in the 399 00:15:01,830 --> 00:15:04,109 waterfall. So it just looks like a hump 400 00:15:04,110 --> 00:15:06,029 in the noise floor is something which is 401 00:15:06,030 --> 00:15:08,519 four point eight kilohertz wide and just 402 00:15:08,520 --> 00:15:10,469 above the noise floor. 403 00:15:10,470 --> 00:15:12,809 So I like this quote from Phalcon 404 00:15:12,810 --> 00:15:13,919 very much. 405 00:15:13,920 --> 00:15:16,679 It says that if you use a good enough 406 00:15:16,680 --> 00:15:18,989 modulation or protocol, then 407 00:15:18,990 --> 00:15:20,699 there are really no features in the 408 00:15:20,700 --> 00:15:23,819 spectrum or in the waterfall. 409 00:15:23,820 --> 00:15:25,469 So what do we do? 410 00:15:25,470 --> 00:15:28,109 Well, we have some experience 411 00:15:28,110 --> 00:15:29,879 and satellite communications. 412 00:15:29,880 --> 00:15:32,259 Most of the time they use 413 00:15:32,260 --> 00:15:34,469 modulation, single channel 414 00:15:34,470 --> 00:15:35,939 know of of D.M. 415 00:15:35,940 --> 00:15:38,039 That's for terrestrial. 416 00:15:38,040 --> 00:15:39,239 But we don't know 417 00:15:40,890 --> 00:15:42,929 if it's Biskupski. 418 00:15:42,930 --> 00:15:45,149 It could also be KWAM. 419 00:15:45,150 --> 00:15:47,879 But we suspect that either PBSC 420 00:15:47,880 --> 00:15:50,009 or Kupinski are the 421 00:15:50,010 --> 00:15:52,019 most likely candidates. 422 00:15:52,020 --> 00:15:54,449 So we use Ganu Radio for the signal 423 00:15:54,450 --> 00:15:56,519 processing. Anyone here familiar 424 00:15:56,520 --> 00:15:57,570 with radio? 425 00:15:59,680 --> 00:16:01,089 Well, that's good. 426 00:16:01,090 --> 00:16:03,249 Yeah, so if 427 00:16:03,250 --> 00:16:05,409 you're into DSB 428 00:16:05,410 --> 00:16:07,569 or signal processing, it's really easy to 429 00:16:07,570 --> 00:16:09,699 get started with a radio. 430 00:16:09,700 --> 00:16:11,949 If not, you 431 00:16:11,950 --> 00:16:13,839 need to learn a bit, but you can start 432 00:16:13,840 --> 00:16:15,969 playing without knowing too much 433 00:16:15,970 --> 00:16:17,229 of the signal processing. 434 00:16:17,230 --> 00:16:18,939 There are lots of examples. 435 00:16:18,940 --> 00:16:21,399 Some things you can do and are 436 00:16:21,400 --> 00:16:22,309 things you can run. 437 00:16:22,310 --> 00:16:23,310 Sorry. 438 00:16:24,710 --> 00:16:27,439 So here I open my recording 439 00:16:27,440 --> 00:16:28,339 with a radio. 440 00:16:28,340 --> 00:16:30,439 The first step is just to filter out 441 00:16:30,440 --> 00:16:32,539 the signal so the filters signal 442 00:16:32,540 --> 00:16:35,299 is in red and 443 00:16:35,300 --> 00:16:37,459 not so much. To talk about this, I 444 00:16:37,460 --> 00:16:39,739 also shifted to baseband, so 445 00:16:39,740 --> 00:16:42,379 I sent her my signal 446 00:16:42,380 --> 00:16:44,719 about the frequency of zero 447 00:16:44,720 --> 00:16:46,939 hertz. Here you can see it, zero 448 00:16:46,940 --> 00:16:47,839 kilohertz. 449 00:16:47,840 --> 00:16:50,029 So that I use the 450 00:16:50,030 --> 00:16:52,249 usual procedure in signal processing. 451 00:16:52,250 --> 00:16:55,099 You center your signal and you 452 00:16:55,100 --> 00:16:56,100 filter. 453 00:16:57,370 --> 00:17:00,219 Now we want to find the upscaled. 454 00:17:00,220 --> 00:17:02,289 So there are some standard way to 455 00:17:02,290 --> 00:17:04,469 do this is you take your seat. 456 00:17:04,470 --> 00:17:05,858 This is my filter signal. 457 00:17:05,859 --> 00:17:08,289 It comes in here and I multiply 458 00:17:08,290 --> 00:17:09,759 it by itself. 459 00:17:09,760 --> 00:17:11,828 So that's like raising the signal to the 460 00:17:11,829 --> 00:17:13,449 power to. 461 00:17:13,450 --> 00:17:15,549 And what happens 462 00:17:15,550 --> 00:17:17,618 when I look at the spectrum, I 463 00:17:17,619 --> 00:17:18,939 get a DC spike. 464 00:17:18,940 --> 00:17:21,489 So there's a strong frequency component 465 00:17:21,490 --> 00:17:22,989 next to zero on. 466 00:17:22,990 --> 00:17:25,899 This means that the signal is PBSC. 467 00:17:25,900 --> 00:17:28,779 If the signal would be Kupinski, 468 00:17:28,780 --> 00:17:30,879 I wouldn't get a spike 469 00:17:30,880 --> 00:17:32,979 for the power to I would 470 00:17:32,980 --> 00:17:35,019 have to go to compete in the fourth 471 00:17:35,020 --> 00:17:37,449 power. So that's multiplying the signal 472 00:17:37,450 --> 00:17:39,669 with itself four times in 473 00:17:39,670 --> 00:17:41,729 order to get a decent 474 00:17:41,730 --> 00:17:42,909 spike in the middle. 475 00:17:42,910 --> 00:17:45,339 And you can sort of go 476 00:17:45,340 --> 00:17:47,469 on, do it for higher 477 00:17:47,470 --> 00:17:49,959 order. GSK, for instance, ETPs 478 00:17:49,960 --> 00:17:51,339 K, it's also popular. 479 00:17:52,360 --> 00:17:54,489 So there's a fairly simple mathematical 480 00:17:54,490 --> 00:17:56,559 reason why this works, but I don't want 481 00:17:56,560 --> 00:17:58,089 to go too much into it. 482 00:17:58,090 --> 00:18:00,459 If anyone is if anyone's interested, 483 00:18:00,460 --> 00:18:01,949 please ask me later. 484 00:18:04,250 --> 00:18:06,679 We also want to know the board's rate. 485 00:18:06,680 --> 00:18:09,739 So how many symbol's per second? 486 00:18:09,740 --> 00:18:11,749 There's also standard way to do this 487 00:18:11,750 --> 00:18:13,809 analysis. It's called cycler, 488 00:18:13,810 --> 00:18:15,679 stationary analysis. 489 00:18:15,680 --> 00:18:17,689 And this involves the following 490 00:18:17,690 --> 00:18:19,789 computation, which is much better 491 00:18:19,790 --> 00:18:21,889 described in these 492 00:18:21,890 --> 00:18:23,659 GANU radio flow graph. 493 00:18:23,660 --> 00:18:25,879 You take your signal and you multiply 494 00:18:25,880 --> 00:18:28,219 it with the conjugate of the delayed 495 00:18:28,220 --> 00:18:29,299 signal. 496 00:18:29,300 --> 00:18:31,519 It's not critical how many samples 497 00:18:31,520 --> 00:18:32,899 you're delaying your signal. 498 00:18:32,900 --> 00:18:34,459 One sample works, OK? 499 00:18:34,460 --> 00:18:37,249 In this case, perhaps ten samples 500 00:18:37,250 --> 00:18:39,349 would work as well, or even better, 501 00:18:39,350 --> 00:18:41,899 but you must delayed a little bit 502 00:18:41,900 --> 00:18:44,179 and then you look at the frequency 503 00:18:44,180 --> 00:18:46,339 components so you 504 00:18:46,340 --> 00:18:48,319 get a DC frequency component. 505 00:18:48,320 --> 00:18:50,419 But that's not what we're interested 506 00:18:50,420 --> 00:18:52,609 into. We're interested into 507 00:18:52,610 --> 00:18:55,159 the next strong frequency component, 508 00:18:55,160 --> 00:18:57,409 and that happens at four point 509 00:18:57,410 --> 00:18:59,029 two kilohertz. 510 00:18:59,030 --> 00:19:01,729 So that tells me that's the symbol rate. 511 00:19:01,730 --> 00:19:03,889 The symbol rate is four point 512 00:19:03,890 --> 00:19:04,890 two kilowatt. 513 00:19:05,750 --> 00:19:06,750 All right. 514 00:19:09,100 --> 00:19:11,469 Now that I know that I have a big 515 00:19:11,470 --> 00:19:13,539 beeps signal and it's 516 00:19:13,540 --> 00:19:15,789 four point to kill a bot, 517 00:19:15,790 --> 00:19:18,939 I can do the beeps modulation. 518 00:19:18,940 --> 00:19:20,409 This is fairly standard. 519 00:19:20,410 --> 00:19:22,749 There's a really nice radio 520 00:19:22,750 --> 00:19:24,969 tutorial about how to do this. 521 00:19:24,970 --> 00:19:27,579 If anyone, starting with a radio 522 00:19:27,580 --> 00:19:29,889 after doing two or three very 523 00:19:29,890 --> 00:19:32,109 basic things, I recommend 524 00:19:32,110 --> 00:19:34,989 you should go and do these tutorial. 525 00:19:34,990 --> 00:19:37,179 So again, really stand 526 00:19:37,180 --> 00:19:39,009 for beeps. Shady modulation. 527 00:19:39,010 --> 00:19:41,589 You have an AGC 528 00:19:41,590 --> 00:19:44,949 just to maintain a constant signal level 529 00:19:44,950 --> 00:19:48,009 which makes signal processing work best. 530 00:19:48,010 --> 00:19:50,289 Then you have a clock recovery 531 00:19:50,290 --> 00:19:52,659 with the polar face clock 532 00:19:52,660 --> 00:19:54,609 recovery algorithm. 533 00:19:54,610 --> 00:19:56,859 And this is where our four point 534 00:19:56,860 --> 00:19:59,439 two kilobits enters the picture 535 00:19:59,440 --> 00:20:01,659 here. Samples per symbol works 536 00:20:01,660 --> 00:20:02,619 out to. 537 00:20:02,620 --> 00:20:04,929 Well, I have so many samples 538 00:20:04,930 --> 00:20:06,849 per second in my recording. 539 00:20:06,850 --> 00:20:08,079 I don't recall how many. 540 00:20:08,080 --> 00:20:10,839 I have four point to kill samples per 541 00:20:10,840 --> 00:20:13,029 kilo. Symbol's per second in 542 00:20:13,030 --> 00:20:13,969 my signal. 543 00:20:13,970 --> 00:20:16,090 So that's the quotient. 544 00:20:17,660 --> 00:20:20,329 And then this recovers 545 00:20:20,330 --> 00:20:23,119 the clock and gets like 546 00:20:23,120 --> 00:20:25,339 the symbols and then 547 00:20:25,340 --> 00:20:27,679 Acosta's loop for frame 548 00:20:27,680 --> 00:20:30,349 frequency tuning, so the frequency 549 00:20:30,350 --> 00:20:32,599 is slightly out and 550 00:20:32,600 --> 00:20:33,859 it's always going to be 551 00:20:35,060 --> 00:20:37,369 this constellation diagram wouldn't 552 00:20:37,370 --> 00:20:39,469 be steel. It would rotate like 553 00:20:39,470 --> 00:20:40,909 this. And you don't want that. 554 00:20:40,910 --> 00:20:43,189 So close the loop takes care 555 00:20:43,190 --> 00:20:44,309 of that. 556 00:20:44,310 --> 00:20:46,849 And this is the Constellation diagram 557 00:20:46,850 --> 00:20:48,919 for any one which is not used to 558 00:20:48,920 --> 00:20:50,269 these diagrams. 559 00:20:50,270 --> 00:20:52,699 Each dot is one of the symbols 560 00:20:52,700 --> 00:20:55,189 we're receiving and they cluster 561 00:20:55,190 --> 00:20:57,559 around this point, which is, let's say, 562 00:20:57,560 --> 00:20:59,749 the big one and 563 00:20:59,750 --> 00:21:01,609 the point on the left, which is the big 564 00:21:01,610 --> 00:21:03,739 zero. So if you have 565 00:21:03,740 --> 00:21:05,629 two clouds which are really well 566 00:21:05,630 --> 00:21:07,699 separated, then your signal is 567 00:21:07,700 --> 00:21:10,039 very good. You draw Lenny here 568 00:21:10,040 --> 00:21:12,169 in the middle and everything which is 569 00:21:12,170 --> 00:21:14,479 on the left is rather than as a zero, 570 00:21:14,480 --> 00:21:16,220 everything on the right is the one. 571 00:21:17,840 --> 00:21:20,149 So the problem is when these 572 00:21:20,150 --> 00:21:22,309 two clouds start to merge and you 573 00:21:22,310 --> 00:21:24,419 have some Beta's 574 00:21:24,420 --> 00:21:26,539 OK, so in this case, 575 00:21:26,540 --> 00:21:27,890 it's more or less good. 576 00:21:28,970 --> 00:21:29,899 We can proceed. 577 00:21:29,900 --> 00:21:32,329 Doing things are 578 00:21:32,330 --> 00:21:34,229 what happens with the coating. 579 00:21:34,230 --> 00:21:36,949 Well, I have four point to clear board, 580 00:21:36,950 --> 00:21:39,169 but the bitrate is only about half. 581 00:21:39,170 --> 00:21:40,979 It's only two kilobits per second. 582 00:21:40,980 --> 00:21:43,249 That's a number taken from 583 00:21:43,250 --> 00:21:45,109 out on that Web page. 584 00:21:45,110 --> 00:21:47,629 So we suspect that 585 00:21:47,630 --> 00:21:49,759 Faik is in use and our equal 586 00:21:49,760 --> 00:21:51,709 one half a code. 587 00:21:51,710 --> 00:21:52,909 What does that mean? 588 00:21:52,910 --> 00:21:55,129 That means that half of 589 00:21:55,130 --> 00:21:58,069 the bits are spent for error 590 00:21:58,070 --> 00:22:00,769 decoding. If you get some errors, 591 00:22:00,770 --> 00:22:01,849 there's no problem. 592 00:22:01,850 --> 00:22:05,119 You're spending half of your bitstream 593 00:22:05,120 --> 00:22:07,220 in using a code which is able to 594 00:22:08,450 --> 00:22:11,269 resolve those beta errors. 595 00:22:11,270 --> 00:22:13,429 So once again, we have no idea 596 00:22:13,430 --> 00:22:15,589 about which forward error 597 00:22:15,590 --> 00:22:17,629 correction code is in use. 598 00:22:17,630 --> 00:22:19,819 We try first whatever is most 599 00:22:19,820 --> 00:22:21,739 popular and this is really, really 600 00:22:21,740 --> 00:22:22,139 popular. 601 00:22:22,140 --> 00:22:25,129 Artists use any were the 602 00:22:25,130 --> 00:22:27,289 DS Convolutional Code. 603 00:22:27,290 --> 00:22:29,689 It's a convolutional code. 604 00:22:29,690 --> 00:22:32,509 And anywhere armatures sitilides 605 00:22:32,510 --> 00:22:34,579 long range space probes out 606 00:22:34,580 --> 00:22:35,580 on that as well. 607 00:22:36,790 --> 00:22:38,979 But there are some parameters, so this 608 00:22:38,980 --> 00:22:41,409 code has some variations 609 00:22:41,410 --> 00:22:43,839 to find the proper parameters. 610 00:22:43,840 --> 00:22:46,089 There's a very nice tool from Balin 611 00:22:46,090 --> 00:22:48,339 Siva, it's called or Tofik. 612 00:22:49,690 --> 00:22:52,329 You just block this block into your 613 00:22:52,330 --> 00:22:54,909 radio flow graph and he tries 614 00:22:54,910 --> 00:22:57,009 to use different parameters for 615 00:22:57,010 --> 00:22:59,409 the Viterbi decoder for 616 00:22:59,410 --> 00:23:01,719 Fassie, and 617 00:23:01,720 --> 00:23:03,939 it tries on different parameters until 618 00:23:03,940 --> 00:23:06,339 it finds the combination, which 619 00:23:06,340 --> 00:23:08,409 makes the bit error rate go 620 00:23:08,410 --> 00:23:10,509 very, very low somewhere when the 621 00:23:10,510 --> 00:23:12,159 bit error rate goes low. 622 00:23:12,160 --> 00:23:14,349 Then, you know, you have the correct 623 00:23:14,350 --> 00:23:15,579 parameters. 624 00:23:15,580 --> 00:23:16,750 So using that, 625 00:23:17,920 --> 00:23:20,769 I found that the CCDs 626 00:23:20,770 --> 00:23:23,499 code used is the standard one, 627 00:23:23,500 --> 00:23:26,139 albeit with the two polynomial swopped. 628 00:23:26,140 --> 00:23:28,609 That's a technical detail. 629 00:23:28,610 --> 00:23:30,609 There are two polynomials which are used 630 00:23:30,610 --> 00:23:32,739 to obtain the code. 631 00:23:32,740 --> 00:23:34,989 So those are called usually Poly A 632 00:23:34,990 --> 00:23:36,219 and B. 633 00:23:36,220 --> 00:23:38,319 First you go with Polya, then you go with 634 00:23:38,320 --> 00:23:40,299 Jollibee out and it does the opposite 635 00:23:40,300 --> 00:23:42,369 thing. You first use Polybius and 636 00:23:42,370 --> 00:23:44,409 then you use poly. 637 00:23:44,410 --> 00:23:46,989 So you have to know that otherwise 638 00:23:46,990 --> 00:23:48,729 you are not going to get anything from 639 00:23:48,730 --> 00:23:49,979 the stream. 640 00:23:49,980 --> 00:23:52,149 OK, so knowing that we 641 00:23:52,150 --> 00:23:54,549 can use the stock radio 642 00:23:54,550 --> 00:23:56,649 Viterbi Decoder and it 643 00:23:56,650 --> 00:23:59,019 goes like this first 644 00:23:59,020 --> 00:24:01,209 we have to swap every pair 645 00:24:01,210 --> 00:24:04,029 of symbols 646 00:24:04,030 --> 00:24:06,639 because of the Polya and B thing 647 00:24:06,640 --> 00:24:08,059 which are swapped. 648 00:24:08,060 --> 00:24:10,359 Then this is the stock radio 649 00:24:10,360 --> 00:24:12,549 decoder there 650 00:24:12,550 --> 00:24:14,739 which spits bytes. 651 00:24:14,740 --> 00:24:17,049 It's each of them has 652 00:24:17,050 --> 00:24:19,959 eight valid bits or symbols. 653 00:24:19,960 --> 00:24:21,249 So this just 654 00:24:22,450 --> 00:24:25,479 takes each byte and outputs 655 00:24:25,480 --> 00:24:27,789 eight different bits 656 00:24:27,790 --> 00:24:29,739 because we want to look at the B stream 657 00:24:29,740 --> 00:24:31,569 to see if there's an important. 658 00:24:32,900 --> 00:24:35,029 And I have to Viterbi 659 00:24:35,030 --> 00:24:37,159 decoders running in parallel, 660 00:24:37,160 --> 00:24:39,829 one is running on the stream 661 00:24:39,830 --> 00:24:41,599 and the other one is running on the 662 00:24:41,600 --> 00:24:42,789 delayed stream. 663 00:24:42,790 --> 00:24:43,819 Why? 664 00:24:43,820 --> 00:24:45,879 Because when you do, 665 00:24:45,880 --> 00:24:48,829 Viterbi, decoding the symbols 666 00:24:48,830 --> 00:24:50,629 are going to be in. 667 00:24:50,630 --> 00:24:53,119 So you want to process them in person 668 00:24:53,120 --> 00:24:54,949 and you are looking at a symbol and you 669 00:24:54,950 --> 00:24:57,229 don't know whether its fur is a symbol 670 00:24:57,230 --> 00:24:59,839 coming just before it or just after 671 00:24:59,840 --> 00:25:00,769 it. 672 00:25:00,770 --> 00:25:02,989 So you just 673 00:25:02,990 --> 00:25:05,539 try something in one branch 674 00:25:05,540 --> 00:25:07,639 of processing and you try the 675 00:25:07,640 --> 00:25:09,439 other thing. You're the branch of 676 00:25:09,440 --> 00:25:11,569 processing and one of them is going 677 00:25:11,570 --> 00:25:13,549 to succeed and the other one would fail 678 00:25:13,550 --> 00:25:15,829 completely. But you don't care, you know, 679 00:25:15,830 --> 00:25:17,479 at some point in the chain. 680 00:25:17,480 --> 00:25:19,669 OK, so I have 681 00:25:19,670 --> 00:25:21,439 to decoders. 682 00:25:21,440 --> 00:25:23,119 I'll put in the bitstream. 683 00:25:23,120 --> 00:25:25,669 And while it just looked garbage, 684 00:25:25,670 --> 00:25:27,739 it's completely random. 685 00:25:27,740 --> 00:25:29,929 So that that means we need 686 00:25:29,930 --> 00:25:32,179 a descrambler that was 687 00:25:32,180 --> 00:25:34,669 expected because if not, 688 00:25:34,670 --> 00:25:37,279 patterns from the data can 689 00:25:37,280 --> 00:25:39,649 can be seen in the waterfall. 690 00:25:39,650 --> 00:25:41,139 And we had no pattern. 691 00:25:41,140 --> 00:25:43,579 So there's some scrambling 692 00:25:43,580 --> 00:25:45,649 which takes our data, which is 693 00:25:45,650 --> 00:25:47,539 full of patterns into something which 694 00:25:47,540 --> 00:25:48,529 looks random. 695 00:25:48,530 --> 00:25:50,329 And that's much better for radio 696 00:25:50,330 --> 00:25:51,229 transmission. 697 00:25:51,230 --> 00:25:53,630 But we have to reverse the process. 698 00:25:56,480 --> 00:25:58,999 OK, so now what do you do? 699 00:25:59,000 --> 00:26:01,209 Again, you try the most popular, 700 00:26:01,210 --> 00:26:03,259 the scramblers, you know, but in this 701 00:26:03,260 --> 00:26:06,169 case, none of which I know 702 00:26:06,170 --> 00:26:07,459 did work. 703 00:26:07,460 --> 00:26:09,619 What can I do? Oh, let's recall, 704 00:26:09,620 --> 00:26:11,879 I have been working 705 00:26:11,880 --> 00:26:14,749 a binary for the receiver 706 00:26:14,750 --> 00:26:17,119 so I can look at the assembler code. 707 00:26:17,120 --> 00:26:19,219 And this is the 708 00:26:19,220 --> 00:26:20,779 AMD 64 bit. 709 00:26:20,780 --> 00:26:23,209 I don't know if it 710 00:26:23,210 --> 00:26:25,489 can be read fairly well. 711 00:26:26,660 --> 00:26:29,089 So this is the assembler code 712 00:26:29,090 --> 00:26:31,789 for the descrambler 713 00:26:31,790 --> 00:26:34,609 in the clotheshorse receiver. 714 00:26:34,610 --> 00:26:36,529 Well, you just take your assembler code 715 00:26:36,530 --> 00:26:38,659 and you start in writing a C code 716 00:26:38,660 --> 00:26:40,819 translation and of 717 00:26:40,820 --> 00:26:43,099 course you simplify your code until it's 718 00:26:43,100 --> 00:26:44,509 human readable. 719 00:26:44,510 --> 00:26:46,579 And then there's 720 00:26:46,580 --> 00:26:48,949 a really nice descrambler 721 00:26:48,950 --> 00:26:49,999 function. 722 00:26:50,000 --> 00:26:52,099 Anyone who seem to descrambler could 723 00:26:52,100 --> 00:26:54,739 go on and understand what's going on. 724 00:26:54,740 --> 00:26:56,869 So this is nothing 725 00:26:56,870 --> 00:26:58,069 I had seen before. 726 00:26:58,070 --> 00:27:00,409 But actually it turns out 727 00:27:00,410 --> 00:27:03,229 it's really popular in the geostationary 728 00:27:03,230 --> 00:27:04,879 satellite industry. 729 00:27:04,880 --> 00:27:07,009 It's called ISIS 730 00:27:07,010 --> 00:27:08,449 three zero eight. 731 00:27:08,450 --> 00:27:11,299 It's from standard from Intelsat. 732 00:27:12,690 --> 00:27:14,969 And actually, this thunder, it's 733 00:27:14,970 --> 00:27:17,189 not open, they only 734 00:27:17,190 --> 00:27:19,769 give it to our partners 735 00:27:19,770 --> 00:27:22,289 or business partners or whatever. 736 00:27:22,290 --> 00:27:24,779 This is from a data sheet of 737 00:27:24,780 --> 00:27:25,889 whatever component. 738 00:27:25,890 --> 00:27:27,989 I find this randomly on the 739 00:27:27,990 --> 00:27:30,089 Internet. I have a very nice 740 00:27:30,090 --> 00:27:31,019 description. 741 00:27:31,020 --> 00:27:34,079 So if anyone anyone's into the scramblers 742 00:27:34,080 --> 00:27:36,089 this piece in the middle with a shift 743 00:27:36,090 --> 00:27:38,309 register, oh, it just 744 00:27:38,310 --> 00:27:40,889 looks like a multiplicative descrambler 745 00:27:40,890 --> 00:27:42,689 and that's what it is. 746 00:27:42,690 --> 00:27:45,509 But the new thing in this 747 00:27:45,510 --> 00:27:48,479 scrambler is that these counter. 748 00:27:48,480 --> 00:27:50,549 So this is just like a multiplicative 749 00:27:50,550 --> 00:27:52,670 descrambler, but with a counter. 750 00:27:54,150 --> 00:27:56,549 And once we have 751 00:27:56,550 --> 00:27:59,249 the code and how it works, 752 00:27:59,250 --> 00:28:01,859 we go implement it into a radio 753 00:28:01,860 --> 00:28:02,860 block. 754 00:28:03,980 --> 00:28:05,959 Plug it into the processing chain in 755 00:28:05,960 --> 00:28:08,929 there and now we look at the data 756 00:28:08,930 --> 00:28:11,149 and you can start seeing some 757 00:28:11,150 --> 00:28:13,129 patterns, there are some. 758 00:28:13,130 --> 00:28:15,379 So this must be read from 759 00:28:15,380 --> 00:28:17,839 left to right, from top to bottom. 760 00:28:17,840 --> 00:28:19,309 So you have some 761 00:28:20,630 --> 00:28:22,549 like some pieces of bits which are 762 00:28:22,550 --> 00:28:24,229 constantly you can see there. 763 00:28:24,230 --> 00:28:25,279 Right. 764 00:28:25,280 --> 00:28:27,289 So this looks fairly good. 765 00:28:28,360 --> 00:28:30,759 OK, the next step is 766 00:28:30,760 --> 00:28:33,129 framing, we have to know where 767 00:28:33,130 --> 00:28:35,499 frame start and where they end 768 00:28:35,500 --> 00:28:36,500 and so on. 769 00:28:37,300 --> 00:28:39,489 So once again, you 770 00:28:39,490 --> 00:28:41,829 could look at the bitstream and 771 00:28:41,830 --> 00:28:43,959 start looking to see if 772 00:28:43,960 --> 00:28:46,389 you figure out any patterns which 773 00:28:46,390 --> 00:28:48,579 might indicate start of frame or 774 00:28:48,580 --> 00:28:49,819 something else. 775 00:28:49,820 --> 00:28:51,969 But that's in this case, it's much 776 00:28:51,970 --> 00:28:53,019 easier. 777 00:28:53,020 --> 00:28:55,089 So I was looking at the as 778 00:28:55,090 --> 00:28:57,249 the 100 binary, I can read 779 00:28:57,250 --> 00:28:59,289 the names of the functions in the 780 00:28:59,290 --> 00:29:01,569 assembler code and several 781 00:29:01,570 --> 00:29:03,939 of those have the name etched DLC 782 00:29:03,940 --> 00:29:06,249 in them. So DLC is 783 00:29:06,250 --> 00:29:07,539 a framing protocol. 784 00:29:07,540 --> 00:29:09,729 It's using amateur radio, Puckett's 785 00:29:09,730 --> 00:29:12,129 radio and many, 786 00:29:12,130 --> 00:29:14,289 many other things. 787 00:29:14,290 --> 00:29:16,809 So, well, we suspect 788 00:29:16,810 --> 00:29:19,749 a DLC framing and 789 00:29:19,750 --> 00:29:22,359 there are tools in radio 790 00:29:22,360 --> 00:29:23,609 to frame it. 791 00:29:23,610 --> 00:29:25,949 Still see there's a stock 792 00:29:25,950 --> 00:29:28,179 defamer, but I prefer to use my 793 00:29:28,180 --> 00:29:31,209 own jerkies pocket 794 00:29:31,210 --> 00:29:33,319 to the frame 795 00:29:33,320 --> 00:29:34,239 HPLC. 796 00:29:34,240 --> 00:29:35,680 So this is what we do. 797 00:29:36,790 --> 00:29:38,889 This is the output of this of the 798 00:29:38,890 --> 00:29:40,029 descrambler. 799 00:29:40,030 --> 00:29:42,279 We feed it into the DLC, 800 00:29:42,280 --> 00:29:44,439 the framer. And the nice thing about my 801 00:29:44,440 --> 00:29:46,869 own Defamer is that you have this check 802 00:29:46,870 --> 00:29:48,389 if six Parmeter. 803 00:29:48,390 --> 00:29:51,729 So HPLC uses said 804 00:29:51,730 --> 00:29:53,799 CCRC 16 bit 805 00:29:53,800 --> 00:29:56,229 to check the integrity of the frames. 806 00:29:56,230 --> 00:29:58,149 Many times when you're doing these sort 807 00:29:58,150 --> 00:30:00,519 of things, you want to skip the check 808 00:30:00,520 --> 00:30:02,739 and get frames even if they have some 809 00:30:02,740 --> 00:30:05,079 errors. So if you put it to false, 810 00:30:05,080 --> 00:30:07,209 it's handy, it spits, frames, 811 00:30:07,210 --> 00:30:09,279 whatever this reality check might 812 00:30:09,280 --> 00:30:10,509 be. 813 00:30:10,510 --> 00:30:12,999 And slight thing here. 814 00:30:13,000 --> 00:30:16,739 When Ussery when you're receiving PBSC, 815 00:30:16,740 --> 00:30:18,879 there is a phase ambiguity of 816 00:30:18,880 --> 00:30:20,499 one hundred and eighty degrees. 817 00:30:20,500 --> 00:30:23,349 You don't know if you're looking at 818 00:30:23,350 --> 00:30:24,669 a one or a zero. 819 00:30:24,670 --> 00:30:26,739 So either you're looking at the good 820 00:30:26,740 --> 00:30:29,199 bitstream or the inverted 821 00:30:29,200 --> 00:30:30,399 bitstream. 822 00:30:30,400 --> 00:30:32,589 So you have to take and 823 00:30:32,590 --> 00:30:34,689 again, you assume it's the 824 00:30:34,690 --> 00:30:36,699 good bitstream on one branch of the 825 00:30:36,700 --> 00:30:39,129 chain. You assume it's the inverted 826 00:30:39,130 --> 00:30:41,229 bitstream. So you inverted once more in 827 00:30:41,230 --> 00:30:43,419 the other branch and only one branch 828 00:30:43,420 --> 00:30:45,429 is going to succeed at one time, but you 829 00:30:45,430 --> 00:30:46,599 don't care. 830 00:30:46,600 --> 00:30:48,729 So any one of 831 00:30:48,730 --> 00:30:50,949 these two will output frames at 832 00:30:50,950 --> 00:30:54,009 the end. And of course, recall that 833 00:30:54,010 --> 00:30:56,199 we had to Viterbi decoder 834 00:30:56,200 --> 00:30:57,249 branches. 835 00:30:57,250 --> 00:30:59,469 So each of those has 836 00:30:59,470 --> 00:31:00,969 to Defamer. 837 00:31:00,970 --> 00:31:02,709 So we are running for different means in 838 00:31:02,710 --> 00:31:03,710 parallel. 839 00:31:05,140 --> 00:31:07,629 And with this, 840 00:31:07,630 --> 00:31:09,130 I start, we get frames. 841 00:31:10,470 --> 00:31:12,539 So now what do we do with the frames? 842 00:31:13,770 --> 00:31:16,229 Well, the Knicks just look at the 843 00:31:16,230 --> 00:31:18,389 hex dumps for the frames 844 00:31:18,390 --> 00:31:20,579 very carefully and something 845 00:31:20,580 --> 00:31:21,959 much clever you can do. 846 00:31:21,960 --> 00:31:23,969 You have the own binary. 847 00:31:23,970 --> 00:31:26,249 It gets frames from as they are 848 00:31:26,250 --> 00:31:28,629 100 with a unique socket, 849 00:31:28,630 --> 00:31:30,779 so you can feed frames into 850 00:31:30,780 --> 00:31:32,999 the unique socket and see what happens. 851 00:31:33,000 --> 00:31:35,879 So you can figure out that way. 852 00:31:35,880 --> 00:31:38,099 And well, the protocols 853 00:31:38,100 --> 00:31:40,289 that it used in Tonette, 854 00:31:40,290 --> 00:31:42,569 they are custom or they look that 855 00:31:42,570 --> 00:31:44,669 I haven't seen anything like those. 856 00:31:44,670 --> 00:31:46,799 So I have named them as I 857 00:31:46,800 --> 00:31:47,800 liked. 858 00:31:49,110 --> 00:31:51,689 And this is a typical frame, can anyone 859 00:31:51,690 --> 00:31:53,939 see any instructor in here, 860 00:31:53,940 --> 00:31:54,940 please? 861 00:31:56,720 --> 00:31:57,720 No. 862 00:31:58,590 --> 00:32:00,689 OK, this five 863 00:32:00,690 --> 00:32:03,599 thing at the very beginning is 864 00:32:03,600 --> 00:32:05,849 so really caught my attention, this 865 00:32:05,850 --> 00:32:08,009 looks like a broadcast Ethernet 866 00:32:08,010 --> 00:32:09,089 address. 867 00:32:09,090 --> 00:32:11,639 This is a broken Ethernet address. 868 00:32:11,640 --> 00:32:13,409 It makes very much sense. 869 00:32:13,410 --> 00:32:15,959 This is a broadcast network. 870 00:32:15,960 --> 00:32:18,209 So, in fact, this is an 871 00:32:18,210 --> 00:32:19,499 Ethernet frame. 872 00:32:19,500 --> 00:32:21,569 You have the broadcast destination, 873 00:32:21,570 --> 00:32:23,669 the source mark of whatever 874 00:32:23,670 --> 00:32:25,969 PC is running on outand 875 00:32:25,970 --> 00:32:28,529 out on its ground station. 876 00:32:28,530 --> 00:32:31,499 And the effort I discuss them 877 00:32:31,500 --> 00:32:34,019 all the frames or the largest 878 00:32:34,020 --> 00:32:36,209 frames, which are most of the frames 879 00:32:36,210 --> 00:32:38,849 are two hundred and seventy six bytes. 880 00:32:38,850 --> 00:32:41,129 And I think that's chosen because 881 00:32:41,130 --> 00:32:43,289 it takes approximately one second to 882 00:32:43,290 --> 00:32:45,599 transmit such a frame of the year 883 00:32:45,600 --> 00:32:48,899 with the modulation in use by outlet. 884 00:32:48,900 --> 00:32:51,059 So that's really handy if you're setting 885 00:32:51,060 --> 00:32:53,219 up your antenna within one second, 886 00:32:53,220 --> 00:32:55,289 you know, if you're getting friends or 887 00:32:55,290 --> 00:32:56,290 not. 888 00:32:58,010 --> 00:33:00,199 So let's look first at the level 889 00:33:00,200 --> 00:33:02,089 three protocol. 890 00:33:02,090 --> 00:33:03,949 It's called the Alternate Protocol, or 891 00:33:03,950 --> 00:33:05,269 OPIS. 892 00:33:05,270 --> 00:33:07,879 And since this is a broadcast 893 00:33:07,880 --> 00:33:10,009 network, it's only role 894 00:33:10,010 --> 00:33:12,019 is the kind of fragmentation. 895 00:33:12,020 --> 00:33:14,389 So, in fact, it's a very convoluted 896 00:33:14,390 --> 00:33:16,459 way of handling fragmentation. 897 00:33:16,460 --> 00:33:18,709 But we do have one byte 898 00:33:18,710 --> 00:33:20,929 which indicates whether this 899 00:33:20,930 --> 00:33:23,509 is the last fragment or not. 900 00:33:23,510 --> 00:33:25,639 And in case you have 901 00:33:25,640 --> 00:33:28,339 a fragmented packet, 902 00:33:28,340 --> 00:33:31,279 you have one byte which indicates 903 00:33:31,280 --> 00:33:33,769 the number of fragments 904 00:33:33,770 --> 00:33:36,499 in the packet or the 905 00:33:36,500 --> 00:33:38,579 number of the last fragment. 906 00:33:38,580 --> 00:33:40,699 And you also have a bite which 907 00:33:40,700 --> 00:33:43,009 goes contin up from zero to the 908 00:33:43,010 --> 00:33:45,169 total number of fragments in 909 00:33:45,170 --> 00:33:47,569 your pocket. So this only handles 910 00:33:47,570 --> 00:33:49,099 fragmentation. 911 00:33:49,100 --> 00:33:51,859 You have also a packet size 912 00:33:51,860 --> 00:33:53,209 in there, 16 bits, 913 00:33:54,350 --> 00:33:55,999 the layer four protocol. 914 00:33:56,000 --> 00:33:58,189 It's called LDP and it's really 915 00:33:58,190 --> 00:34:00,349 similar to UDP. 916 00:34:00,350 --> 00:34:02,509 So you have some concept of 917 00:34:02,510 --> 00:34:04,639 ports. They're used to identify 918 00:34:04,640 --> 00:34:07,339 different services within the network. 919 00:34:07,340 --> 00:34:09,109 I'm not sure. And it's really hard to 920 00:34:09,110 --> 00:34:11,419 tell with the data we have. 921 00:34:11,420 --> 00:34:13,488 If there is the concept of source 922 00:34:13,489 --> 00:34:15,829 port and destination port, 923 00:34:15,830 --> 00:34:16,979 I really don't care. 924 00:34:16,980 --> 00:34:19,549 So I have two ports and 925 00:34:19,550 --> 00:34:21,859 the power of two ports identifies 926 00:34:21,860 --> 00:34:24,439 the service and I call those the A field 927 00:34:24,440 --> 00:34:25,729 and the B field. 928 00:34:27,170 --> 00:34:29,329 And you have packet size 929 00:34:29,330 --> 00:34:31,849 on a checksum. So very much like UDP. 930 00:34:33,659 --> 00:34:35,999 Apart from transferring files, 931 00:34:36,000 --> 00:34:38,069 something out of the dust 932 00:34:38,070 --> 00:34:40,439 is to broadcast time packets, 933 00:34:40,440 --> 00:34:42,238 and these are used to sell the receiver 934 00:34:42,239 --> 00:34:44,428 clock because it's intended 935 00:34:44,429 --> 00:34:46,379 to run without Internet access. 936 00:34:46,380 --> 00:34:48,269 So you cannot use A.P. 937 00:34:48,270 --> 00:34:50,698 and these are very 938 00:34:50,699 --> 00:34:52,859 easy. They have here 939 00:34:52,860 --> 00:34:54,928 the ASCII code for in this 940 00:34:54,929 --> 00:34:56,249 case, Odissi two. 941 00:34:56,250 --> 00:34:58,859 That means alternate datacasting to 942 00:34:58,860 --> 00:35:01,019 it's the ID for 943 00:35:01,020 --> 00:35:02,819 the ground station in the America 944 00:35:02,820 --> 00:35:05,099 satellite. This comes from Scott's 945 00:35:05,100 --> 00:35:07,919 recording in the East Coast. 946 00:35:07,920 --> 00:35:09,989 You have something which I don't know 947 00:35:09,990 --> 00:35:12,089 what it is, but you have the Unix 948 00:35:12,090 --> 00:35:13,439 timestamp. 949 00:35:13,440 --> 00:35:15,629 So any time the receiver gets 950 00:35:15,630 --> 00:35:17,699 a packet such as this, it reads the 951 00:35:17,700 --> 00:35:20,279 Unix timestamp and updates updates 952 00:35:20,280 --> 00:35:21,280 it clock. 953 00:35:22,310 --> 00:35:24,499 And there is some padding in the 954 00:35:24,500 --> 00:35:26,569 end because pockets 955 00:35:26,570 --> 00:35:29,059 have to be at least 46 956 00:35:29,060 --> 00:35:31,669 bites. This is some limitation 957 00:35:31,670 --> 00:35:34,280 of the Ethernet fiscal protocol. 958 00:35:37,040 --> 00:35:39,709 Let's go to how files are broadcast, 959 00:35:39,710 --> 00:35:41,899 so one file, one file, it's 960 00:35:41,900 --> 00:35:45,079 broadcast at its time, 961 00:35:45,080 --> 00:35:46,369 this need not be so. 962 00:35:46,370 --> 00:35:48,529 The protocol allows for simultaneous 963 00:35:48,530 --> 00:35:50,059 broadcast. 964 00:35:50,060 --> 00:35:53,299 Every file gets split into 242 965 00:35:53,300 --> 00:35:55,279 byte blocks. So you have many, many 966 00:35:55,280 --> 00:35:56,280 blocks. 967 00:35:56,900 --> 00:35:59,119 There's LDC codes used 968 00:35:59,120 --> 00:36:02,269 to recover the file, even if you 969 00:36:02,270 --> 00:36:04,339 failed to receive some of 970 00:36:04,340 --> 00:36:06,289 the blocks. So if you have holes, you can 971 00:36:06,290 --> 00:36:09,469 feel them using the LDP code. 972 00:36:09,470 --> 00:36:11,929 And there are three types of packets. 973 00:36:11,930 --> 00:36:14,059 First comes the file announcement, 974 00:36:14,060 --> 00:36:16,279 which tells you all the details about 975 00:36:16,280 --> 00:36:17,280 the file. 976 00:36:18,110 --> 00:36:20,239 Then you have a bunch of file 977 00:36:20,240 --> 00:36:21,889 blocks, which are the contents of the 978 00:36:21,890 --> 00:36:22,999 file. 979 00:36:23,000 --> 00:36:24,499 Lastly, you have the 980 00:36:25,550 --> 00:36:27,709 fact box, which are the party 981 00:36:27,710 --> 00:36:31,009 tech bytes for the LDP codes, 982 00:36:31,010 --> 00:36:33,139 and these are used to 983 00:36:33,140 --> 00:36:35,869 fill in blocks in case you 984 00:36:35,870 --> 00:36:38,599 missed some of the file blocks. 985 00:36:38,600 --> 00:36:41,059 So these five books and 986 00:36:41,060 --> 00:36:42,679 books are sent interleaved. 987 00:36:44,240 --> 00:36:46,339 The file enhancement packets 988 00:36:46,340 --> 00:36:49,219 contain an XML in ASCII 989 00:36:49,220 --> 00:36:51,349 with information about the file 990 00:36:51,350 --> 00:36:53,659 and they are signed with an X five 991 00:36:53,660 --> 00:36:56,119 nine certificate Y, probably 992 00:36:56,120 --> 00:36:57,589 to prevent spoofing. 993 00:36:57,590 --> 00:36:59,959 So the outliner receiver has 994 00:36:59,960 --> 00:37:02,079 the certificate for out on 995 00:37:02,080 --> 00:37:04,549 it and it checks 996 00:37:04,550 --> 00:37:06,409 that the file announcement is properly 997 00:37:06,410 --> 00:37:08,509 signed. And if not, I 998 00:37:08,510 --> 00:37:09,709 don't know what happens. 999 00:37:09,710 --> 00:37:11,660 Probably it discards the file. 1000 00:37:13,160 --> 00:37:15,529 So the information you want 1001 00:37:15,530 --> 00:37:17,839 in here, the idea, the 1002 00:37:17,840 --> 00:37:19,909 size, the name of the file, 1003 00:37:19,910 --> 00:37:22,189 Hosch to check and some 1004 00:37:22,190 --> 00:37:24,679 information about the LDP code 1005 00:37:24,680 --> 00:37:26,779 in use for the file and the block size 1006 00:37:26,780 --> 00:37:28,369 is always two hundred and forty two 1007 00:37:29,480 --> 00:37:31,669 so far, but it's very 1008 00:37:31,670 --> 00:37:34,129 easy. You have in blue the file ID 1009 00:37:34,130 --> 00:37:35,659 in green the block number. 1010 00:37:35,660 --> 00:37:37,849 It goes counting from zero to 1011 00:37:37,850 --> 00:37:39,929 whatever number and then the 1012 00:37:39,930 --> 00:37:41,749 block contents. Two hundred and forty two 1013 00:37:41,750 --> 00:37:42,750 bytes. 1014 00:37:43,900 --> 00:37:45,880 So that's how podcasts are 1015 00:37:46,930 --> 00:37:49,029 transmitted, and once we 1016 00:37:49,030 --> 00:37:51,279 have all of these information together, 1017 00:37:51,280 --> 00:37:53,499 we can just go and code 1018 00:37:53,500 --> 00:37:56,019 a Python implementation that 1019 00:37:56,020 --> 00:37:58,389 gets the frames and recovers 1020 00:37:58,390 --> 00:37:59,439 files. 1021 00:37:59,440 --> 00:38:01,659 So we haven't been able to reverse 1022 00:38:01,660 --> 00:38:04,039 engineering LDP codes yet. 1023 00:38:04,040 --> 00:38:06,249 That means you need to get all 1024 00:38:06,250 --> 00:38:07,449 the blocks for the file. 1025 00:38:07,450 --> 00:38:09,579 If you miss a single block, you have a 1026 00:38:09,580 --> 00:38:11,649 hole in your data and you can get 1027 00:38:11,650 --> 00:38:14,409 the file, but otherwise you're fine. 1028 00:38:14,410 --> 00:38:16,509 So this is the Python implementation. 1029 00:38:16,510 --> 00:38:18,819 It prints out a lot of 1030 00:38:18,820 --> 00:38:21,459 debugging, informational curiosities. 1031 00:38:21,460 --> 00:38:23,619 You may want to know the mark 1032 00:38:23,620 --> 00:38:26,619 of the ground station or whatever, 1033 00:38:26,620 --> 00:38:28,719 but the nice thing is that 1034 00:38:28,720 --> 00:38:30,289 here where is it? 1035 00:38:30,290 --> 00:38:32,799 There's a new file announcement for 1036 00:38:32,800 --> 00:38:35,739 Amazon.com Wikipedia page, 1037 00:38:35,740 --> 00:38:37,839 and that's around 1038 00:38:37,840 --> 00:38:39,069 eight 03. 1039 00:38:39,070 --> 00:38:41,229 And 20 minutes later, that's what it 1040 00:38:41,230 --> 00:38:43,419 takes to transmit 1041 00:38:43,420 --> 00:38:45,249 that Wikipedia page. 1042 00:38:45,250 --> 00:38:47,559 You get all your file blocks and 1043 00:38:47,560 --> 00:38:49,819 these managed to recover the file 1044 00:38:49,820 --> 00:38:51,820 and save it onto your hard drive. 1045 00:38:54,110 --> 00:38:56,179 So what do we have now? 1046 00:38:56,180 --> 00:38:58,129 Well, there's lots of documentation on 1047 00:38:58,130 --> 00:39:00,559 everything I've talked about here 1048 00:39:00,560 --> 00:39:03,109 and a little bit more 1049 00:39:03,110 --> 00:39:04,789 in my blog. 1050 00:39:04,790 --> 00:39:06,859 There's the a radio receiver. 1051 00:39:06,860 --> 00:39:09,049 You can put frames 1052 00:39:09,050 --> 00:39:11,149 in real time with UDP or you 1053 00:39:11,150 --> 00:39:13,069 can do a case recording. 1054 00:39:13,070 --> 00:39:15,169 And then you have the Python code, 1055 00:39:15,170 --> 00:39:17,329 which takes either real time 1056 00:39:17,330 --> 00:39:19,969 frames from the UDP socket 1057 00:39:19,970 --> 00:39:23,059 or the Keys file recording and 1058 00:39:23,060 --> 00:39:26,179 gets the files from from those. 1059 00:39:26,180 --> 00:39:28,819 So what about all the decoding? 1060 00:39:28,820 --> 00:39:30,919 We've done some progress, not so 1061 00:39:30,920 --> 00:39:31,920 much. 1062 00:39:32,610 --> 00:39:34,999 There's some track of the progress 1063 00:39:35,000 --> 00:39:37,069 on GitHub if anyone's interested 1064 00:39:37,070 --> 00:39:39,229 to come talk with me. 1065 00:39:39,230 --> 00:39:41,479 Currently, I don't have much motivation 1066 00:39:41,480 --> 00:39:43,879 to go into the binary code on reverse 1067 00:39:43,880 --> 00:39:46,129 engineering, but we might 1068 00:39:46,130 --> 00:39:47,300 be able to do something. 1069 00:39:48,680 --> 00:39:50,929 So some miscellaneous fun 1070 00:39:50,930 --> 00:39:51,930 stuff. 1071 00:39:52,750 --> 00:39:55,569 Well, the satellite modem, 1072 00:39:55,570 --> 00:39:57,969 so the X five or nine certificates 1073 00:39:59,410 --> 00:40:01,749 they use are seeing 1074 00:40:01,750 --> 00:40:04,329 these addresses, Odissi to dot alternate 1075 00:40:04,330 --> 00:40:06,759 is for you say Odissi 1076 00:40:06,760 --> 00:40:08,739 three is for Europe. 1077 00:40:08,740 --> 00:40:10,359 So what's in there? 1078 00:40:10,360 --> 00:40:12,609 You can think that machine, it thinks 1079 00:40:12,610 --> 00:40:15,039 you can go by HTP. 1080 00:40:15,040 --> 00:40:16,089 Why not? 1081 00:40:16,090 --> 00:40:18,309 And now the TDP port, 1082 00:40:18,310 --> 00:40:19,299 it's blocked. 1083 00:40:19,300 --> 00:40:20,679 So they've noticed. 1084 00:40:20,680 --> 00:40:23,109 But before I swear, 1085 00:40:23,110 --> 00:40:25,239 I got the login page of 1086 00:40:25,240 --> 00:40:27,399 the satellite modem they use 1087 00:40:27,400 --> 00:40:29,619 on the ground station to upload 1088 00:40:29,620 --> 00:40:31,629 the data to satellite. 1089 00:40:31,630 --> 00:40:33,849 So huge security flow. 1090 00:40:33,850 --> 00:40:36,009 Of course, I didn't get past that. 1091 00:40:36,010 --> 00:40:38,289 It was open probably for week. 1092 00:40:38,290 --> 00:40:40,869 Well, surely for weeks, perhaps 1093 00:40:40,870 --> 00:40:42,449 even for more than a month. 1094 00:40:43,580 --> 00:40:45,649 Huge security for what 1095 00:40:45,650 --> 00:40:48,559 I'm interested is what's 1096 00:40:48,560 --> 00:40:51,079 the model of the modem, it's an M7 1097 00:40:51,080 --> 00:40:53,269 modem and now you go 1098 00:40:53,270 --> 00:40:55,459 Google it and there's 1099 00:40:55,460 --> 00:40:58,099 lots of documentation and some pictures 1100 00:40:58,100 --> 00:40:59,419 for your modem fans. 1101 00:41:01,440 --> 00:41:03,779 This is the data sheet from the moment 1102 00:41:03,780 --> 00:41:05,969 I've marked in yellow what 1103 00:41:05,970 --> 00:41:07,559 is used in AlterNet. 1104 00:41:07,560 --> 00:41:09,419 Let's see if I can zoom in here. 1105 00:41:09,420 --> 00:41:12,799 Yes, so. 1106 00:41:12,800 --> 00:41:14,899 Yeah, probably not a very good 1107 00:41:14,900 --> 00:41:16,280 idea to do it like this. 1108 00:41:17,720 --> 00:41:20,479 So visas for the FCC, 1109 00:41:20,480 --> 00:41:22,909 which they used to be modulation 1110 00:41:22,910 --> 00:41:25,099 and so on, as you can see, only the 1111 00:41:25,100 --> 00:41:27,679 most basic are 1112 00:41:27,680 --> 00:41:30,499 options of the modem are in use. 1113 00:41:30,500 --> 00:41:33,289 They could have used Khyam 1114 00:41:33,290 --> 00:41:35,689 and there are very nice LDC 1115 00:41:35,690 --> 00:41:38,329 codes, probably much harder 1116 00:41:38,330 --> 00:41:39,579 to reverse engineer. 1117 00:41:39,580 --> 00:41:41,989 It are a much better 1118 00:41:41,990 --> 00:41:44,119 performance that they could have used 1119 00:41:44,120 --> 00:41:46,699 instead of the really simple Cissie, 1120 00:41:46,700 --> 00:41:49,409 the 60s or 1121 00:41:49,410 --> 00:41:50,410 Viterbi code. 1122 00:41:52,870 --> 00:41:54,099 What else? 1123 00:41:54,100 --> 00:41:56,169 Well, you have the eyepiece 1124 00:41:56,170 --> 00:41:58,269 for the ground stations 1125 00:41:58,270 --> 00:42:00,669 so you can look them, use 1126 00:42:00,670 --> 00:42:02,739 any old line to maybe you want 1127 00:42:02,740 --> 00:42:04,809 to know where the ground stations are. 1128 00:42:04,810 --> 00:42:07,239 So the Americas is in Toronto. 1129 00:42:07,240 --> 00:42:09,189 The Europe. Africa is in Amsterdam. 1130 00:42:09,190 --> 00:42:10,749 I don't know if there's anyone from 1131 00:42:10,750 --> 00:42:12,459 Amsterdam in this room. 1132 00:42:12,460 --> 00:42:14,989 It would be. Yes, yes. 1133 00:42:14,990 --> 00:42:17,259 So it would be very, very nice. 1134 00:42:17,260 --> 00:42:19,389 If you can listen on the uplink signal, I 1135 00:42:19,390 --> 00:42:21,369 don't know the frequency, but maybe you 1136 00:42:21,370 --> 00:42:23,439 can search or maybe you can manage to 1137 00:42:23,440 --> 00:42:25,509 do it. So listen on the uplink 1138 00:42:25,510 --> 00:42:27,699 signal and see if you can 1139 00:42:27,700 --> 00:42:29,469 just receive the uplink signal. 1140 00:42:29,470 --> 00:42:30,549 The format is the same. 1141 00:42:30,550 --> 00:42:32,979 The satellite is like a bent 1142 00:42:32,980 --> 00:42:34,209 pipe transponder. 1143 00:42:34,210 --> 00:42:36,069 Whatever goes up goes down. 1144 00:42:36,070 --> 00:42:38,409 So if you're in Amsterdam or anywhere 1145 00:42:38,410 --> 00:42:40,509 near all the ground stations, see if 1146 00:42:40,510 --> 00:42:42,969 you can receive the uplink. 1147 00:42:42,970 --> 00:42:45,189 And another thing you can do 1148 00:42:45,190 --> 00:42:47,469 if you can receive the uplink is go 1149 00:42:47,470 --> 00:42:49,809 on, do some radio direction, 1150 00:42:49,810 --> 00:42:53,139 finding maybe you can find the 1151 00:42:53,140 --> 00:42:55,059 ground station. Probably it has very, 1152 00:42:55,060 --> 00:42:57,339 very large satellite dishes. 1153 00:42:59,030 --> 00:43:01,549 So the actually Asia-Pacific, 1154 00:43:01,550 --> 00:43:03,679 it's in some small village in 1155 00:43:03,680 --> 00:43:04,680 New Zealand. 1156 00:43:06,200 --> 00:43:08,359 What else? Finally, the data 1157 00:43:08,360 --> 00:43:11,059 throughput, so 20 megabytes 1158 00:43:11,060 --> 00:43:13,339 of content per day, that's what 1159 00:43:13,340 --> 00:43:14,299 it promises. 1160 00:43:14,300 --> 00:43:15,499 Is this true? 1161 00:43:15,500 --> 00:43:17,229 Well, let's run the numbers. 1162 00:43:18,680 --> 00:43:21,199 So we have files which are syncing 1163 00:43:21,200 --> 00:43:23,389 242 blocks, 1164 00:43:23,390 --> 00:43:25,519 bytes per block, 1165 00:43:25,520 --> 00:43:27,619 and they feed inside 1166 00:43:27,620 --> 00:43:29,479 Ethernet frames, which are two hundred 1167 00:43:29,480 --> 00:43:31,159 and seventy two bytes each. 1168 00:43:31,160 --> 00:43:33,559 So that's a 12 percent overhead 1169 00:43:33,560 --> 00:43:35,629 for headers Arsène 1170 00:43:35,630 --> 00:43:38,059 Ethernet headers, opis 1171 00:43:38,060 --> 00:43:39,320 on LDP headers. 1172 00:43:40,490 --> 00:43:42,739 Then most of the files 1173 00:43:42,740 --> 00:43:45,139 use LDP codes as well. 1174 00:43:45,140 --> 00:43:47,299 The rate is around zero point 1175 00:43:47,300 --> 00:43:49,609 eighty three. So this is a 20 1176 00:43:49,610 --> 00:43:51,739 percent overhead on top of the 1177 00:43:51,740 --> 00:43:54,139 12 percent overhead that works 1178 00:43:54,140 --> 00:43:56,329 out to a total overhead of 30 1179 00:43:56,330 --> 00:43:57,829 percent. 1180 00:43:57,830 --> 00:44:00,199 And the bitrate is only two point 1181 00:44:00,200 --> 00:44:02,629 one kilobits per second. 1182 00:44:02,630 --> 00:44:04,879 And that's at most you would have 1183 00:44:04,880 --> 00:44:07,399 to factor in the bit stuffing 1184 00:44:07,400 --> 00:44:09,679 done by SDLC. 1185 00:44:09,680 --> 00:44:12,589 So when you work out how many 1186 00:44:12,590 --> 00:44:14,929 information per day you have, it's 1187 00:44:14,930 --> 00:44:17,029 only 15 megabytes 1188 00:44:17,030 --> 00:44:19,219 per day with the 1189 00:44:19,220 --> 00:44:21,679 invention that one megabyte 1190 00:44:21,680 --> 00:44:23,869 is 124 kilobytes 1191 00:44:23,870 --> 00:44:25,939 and so on. If you was if you 1192 00:44:25,940 --> 00:44:28,369 use the 1000 convention, you would get, 1193 00:44:28,370 --> 00:44:29,780 I don't know, 16 megabit, 1194 00:44:31,100 --> 00:44:33,199 but nothing near the 20 1195 00:44:33,200 --> 00:44:35,809 megabit we were promised. 1196 00:44:35,810 --> 00:44:37,130 So thank you very much. 1197 00:44:47,600 --> 00:44:48,600 Thank you, Danielle, 1198 00:44:49,820 --> 00:44:50,989 do we have questions 1199 00:44:52,250 --> 00:44:54,739 for questions, please use the microphones 1200 00:44:54,740 --> 00:44:56,899 provided over here, here and here and 1201 00:44:56,900 --> 00:44:59,269 here, one to four. 1202 00:44:59,270 --> 00:45:00,499 And we also have a signal. 1203 00:45:00,500 --> 00:45:02,149 Angel, do we have questions from the 1204 00:45:02,150 --> 00:45:03,150 Internet? 1205 00:45:05,860 --> 00:45:08,049 Yes, we have, um, 1206 00:45:08,050 --> 00:45:10,269 actually the scrambler 1207 00:45:10,270 --> 00:45:12,399 is that basically encryption 1208 00:45:12,400 --> 00:45:14,499 or is it for, you 1209 00:45:14,500 --> 00:45:16,359 know, another RF purpose? 1210 00:45:19,140 --> 00:45:21,089 You mean like this whole reverse 1211 00:45:21,090 --> 00:45:23,789 engineering work, 1212 00:45:23,790 --> 00:45:26,249 so it's it's most 1213 00:45:26,250 --> 00:45:28,649 of the techniques are useful for 1214 00:45:28,650 --> 00:45:30,759 for many other stuff in 1215 00:45:30,760 --> 00:45:33,689 in RF, the RF processing, 1216 00:45:33,690 --> 00:45:35,969 most of it, it's thunder in 1217 00:45:35,970 --> 00:45:39,029 armor to satellites or there are 1218 00:45:39,030 --> 00:45:41,699 commercial satellite communication. 1219 00:45:41,700 --> 00:45:44,010 So I'm not sure if that was the question. 1220 00:45:45,480 --> 00:45:47,699 The question was you build a 1221 00:45:47,700 --> 00:45:48,959 descrambler. 1222 00:45:48,960 --> 00:45:51,509 So why is the signal's scrambled at all? 1223 00:45:51,510 --> 00:45:53,669 Why is the signal scrambled at all, 1224 00:45:53,670 --> 00:45:56,759 sir? It happens like this. 1225 00:45:56,760 --> 00:45:59,189 If you have data most of the time 1226 00:45:59,190 --> 00:46:01,319 the data is going to have some pattern 1227 00:46:01,320 --> 00:46:03,419 in it. For instance, in here, every 1228 00:46:03,420 --> 00:46:05,639 time we have the Ethernet header, we 1229 00:46:05,640 --> 00:46:08,139 have this whole bunch of 1230 00:46:08,140 --> 00:46:10,799 F data, which is the 1231 00:46:10,800 --> 00:46:12,539 broadcast mark. 1232 00:46:12,540 --> 00:46:14,729 So any any time you have 1233 00:46:14,730 --> 00:46:16,799 patterns in your bitstream, 1234 00:46:16,800 --> 00:46:18,869 if you modulate with any sort of 1235 00:46:18,870 --> 00:46:20,939 modulation, beat beeps 1236 00:46:20,940 --> 00:46:23,129 or anything else, that 1237 00:46:23,130 --> 00:46:25,349 means some pattern in the 1238 00:46:25,350 --> 00:46:28,379 frequency in the frequency domain. 1239 00:46:28,380 --> 00:46:30,269 So that's not very good from the 1240 00:46:30,270 --> 00:46:32,399 performance point of view of 1241 00:46:32,400 --> 00:46:34,509 RF engineering. 1242 00:46:34,510 --> 00:46:36,899 You get less 1243 00:46:36,900 --> 00:46:39,029 performance with respect to 1244 00:46:39,030 --> 00:46:41,099 Signal-to-noise performance, 1245 00:46:41,100 --> 00:46:43,289 also interference to other 1246 00:46:43,290 --> 00:46:45,359 channels. So when you want to 1247 00:46:45,360 --> 00:46:47,789 transmit some digital data on RF, 1248 00:46:47,790 --> 00:46:50,069 you want that your data looks completely 1249 00:46:50,070 --> 00:46:52,229 random. And what you do is you use 1250 00:46:52,230 --> 00:46:54,989 some simpler scramble function, 1251 00:46:54,990 --> 00:46:57,329 which takes your data on outputs, 1252 00:46:57,330 --> 00:47:00,359 something which is easily recoverable 1253 00:47:00,360 --> 00:47:02,819 but looks random enough 1254 00:47:02,820 --> 00:47:05,159 so that you don't get any patterns 1255 00:47:05,160 --> 00:47:07,019 in your RF spectrum. 1256 00:47:09,750 --> 00:47:10,750 Thank you. 1257 00:47:11,160 --> 00:47:12,769 Question for Michael. 1258 00:47:12,770 --> 00:47:15,239 One high, I understand 1259 00:47:15,240 --> 00:47:16,919 the overhead for the forward error 1260 00:47:16,920 --> 00:47:19,289 correction, but why do they use 1261 00:47:19,290 --> 00:47:21,419 14 bytes for the Internet header 1262 00:47:21,420 --> 00:47:22,420 and the idea? 1263 00:47:23,820 --> 00:47:26,399 Well, the thing is that if you go back 1264 00:47:26,400 --> 00:47:28,499 to the documentation about the 1265 00:47:28,500 --> 00:47:29,550 satellite modem 1266 00:47:30,570 --> 00:47:33,179 or were presented 1267 00:47:33,180 --> 00:47:34,399 probably here, 1268 00:47:35,760 --> 00:47:37,979 I hear so on the top, 1269 00:47:37,980 --> 00:47:40,319 it supports like 1270 00:47:40,320 --> 00:47:43,289 several different input 1271 00:47:43,290 --> 00:47:43,869 formats. 1272 00:47:43,870 --> 00:47:46,049 So you're going to have a PC, your PC 1273 00:47:46,050 --> 00:47:48,389 like machine, which is running 1274 00:47:48,390 --> 00:47:50,549 all the broadcast 1275 00:47:50,550 --> 00:47:53,189 software in the alternate station, 1276 00:47:53,190 --> 00:47:55,229 and then you plug that into the satellite 1277 00:47:55,230 --> 00:47:57,899 modem. And of course, 1278 00:47:57,900 --> 00:47:59,159 they have to talk to each other. 1279 00:47:59,160 --> 00:48:01,620 So they have to know a common protocol 1280 00:48:02,790 --> 00:48:04,949 that should be simple enough for 1281 00:48:04,950 --> 00:48:07,139 them to implement an Ethernet is 1282 00:48:07,140 --> 00:48:08,159 OK. 1283 00:48:08,160 --> 00:48:10,949 It's very easy to wiretap 1284 00:48:10,950 --> 00:48:12,299 via Ethernet. 1285 00:48:12,300 --> 00:48:14,549 Why do the Ethernet headers 1286 00:48:14,550 --> 00:48:16,199 go over the air? 1287 00:48:16,200 --> 00:48:18,329 So this is intended in the commercial 1288 00:48:18,330 --> 00:48:21,209 application, as far as I understand, 1289 00:48:21,210 --> 00:48:24,239 for you to have like an Ethernet, 1290 00:48:24,240 --> 00:48:26,369 a real Ethernet network, but 1291 00:48:26,370 --> 00:48:28,949 over satellite where you can have 1292 00:48:28,950 --> 00:48:31,349 many different stations on, it acts 1293 00:48:31,350 --> 00:48:33,839 just like another bridge. 1294 00:48:33,840 --> 00:48:34,840 So that's why 1295 00:48:36,090 --> 00:48:37,949 you microphone number two, please. 1296 00:48:37,950 --> 00:48:40,169 Hey, did you see any 1297 00:48:40,170 --> 00:48:42,439 packages with this nation, which is 1298 00:48:42,440 --> 00:48:44,249 the broadcast destination, like unicast 1299 00:48:44,250 --> 00:48:46,589 packets or. 1300 00:48:46,590 --> 00:48:48,569 I think I did. 1301 00:48:48,570 --> 00:48:50,879 So I can I can go back 1302 00:48:50,880 --> 00:48:53,099 to my recordings 1303 00:48:53,100 --> 00:48:55,379 by Scott and do have done dumps 1304 00:48:55,380 --> 00:48:57,839 of these messages. And we're sort of 1305 00:48:57,840 --> 00:49:00,149 not in my head, but yes, in my computer. 1306 00:49:00,150 --> 00:49:02,219 So in my computer, I have 1307 00:49:02,220 --> 00:49:04,379 all the files I've used. 1308 00:49:04,380 --> 00:49:06,599 I think I saw something which 1309 00:49:06,600 --> 00:49:09,509 was not broadcast or at least 1310 00:49:09,510 --> 00:49:11,579 which was not like this 1311 00:49:11,580 --> 00:49:13,829 scheme where you have your customers 1312 00:49:13,830 --> 00:49:16,019 type and you feed your 1313 00:49:16,020 --> 00:49:18,299 alternate pocket inside. 1314 00:49:18,300 --> 00:49:20,609 So maybe AARP 1315 00:49:20,610 --> 00:49:22,799 or something else, 1316 00:49:22,800 --> 00:49:24,989 which was probably not intended to be 1317 00:49:24,990 --> 00:49:26,339 broadcast over there. 1318 00:49:26,340 --> 00:49:28,559 So, yeah, I saw some 1319 00:49:28,560 --> 00:49:30,629 of that. I'm not sure from the top 1320 00:49:30,630 --> 00:49:32,159 of my head what it was. 1321 00:49:32,160 --> 00:49:33,540 I can look it afterwards. 1322 00:49:35,130 --> 00:49:37,589 So yes, they happen sometimes. 1323 00:49:37,590 --> 00:49:39,000 Unintended broadcast. 1324 00:49:40,660 --> 00:49:42,579 Thanks. And I think we have a question 1325 00:49:42,580 --> 00:49:44,229 from the Internet. 1326 00:49:44,230 --> 00:49:45,429 Thank you. 1327 00:49:45,430 --> 00:49:47,679 First, I have a comment before four 1328 00:49:47,680 --> 00:49:50,289 in front of the UNIX timestamp. 1329 00:49:50,290 --> 00:49:52,839 Is it possible that this might be a 64 1330 00:49:52,840 --> 00:49:54,039 bit timestamp? 1331 00:49:54,040 --> 00:49:56,139 So that was not a question, just just 1332 00:49:56,140 --> 00:49:57,399 a comment from the Internet. 1333 00:49:57,400 --> 00:49:59,619 The question is, there's no user 1334 00:49:59,620 --> 00:50:00,620 based uplink. 1335 00:50:01,540 --> 00:50:03,369 So so what is broadcasted? 1336 00:50:03,370 --> 00:50:05,380 Who is deciding what is broadcasted? 1337 00:50:06,580 --> 00:50:08,679 OK, so the thing goes more 1338 00:50:08,680 --> 00:50:10,779 or less like this, I'm not really 1339 00:50:10,780 --> 00:50:12,909 sure, but at least this is what they 1340 00:50:12,910 --> 00:50:13,809 say. 1341 00:50:13,810 --> 00:50:15,909 And in fact, there are some people which 1342 00:50:15,910 --> 00:50:18,099 are interesting in monitoring what 1343 00:50:18,100 --> 00:50:20,379 is being broadcast because there's 1344 00:50:20,380 --> 00:50:22,719 not like a public list of what 1345 00:50:22,720 --> 00:50:25,059 goes out and when it goes out. 1346 00:50:25,060 --> 00:50:27,550 So the idea is that most of the content 1347 00:50:28,630 --> 00:50:31,119 regarding bandwidth is Wikipedia 1348 00:50:31,120 --> 00:50:33,339 files, and this is from 1349 00:50:33,340 --> 00:50:37,329 the most popular Wikipedia 1350 00:50:37,330 --> 00:50:38,319 files. 1351 00:50:38,320 --> 00:50:39,909 So this seems somewhat random. 1352 00:50:39,910 --> 00:50:42,609 But Wikipedia files, they also broadcast 1353 00:50:42,610 --> 00:50:45,129 some weather information 1354 00:50:45,130 --> 00:50:47,199 which is most useful for 1355 00:50:47,200 --> 00:50:50,229 sailors. Sailing on the high seas. 1356 00:50:50,230 --> 00:50:52,659 Are some small files for 1357 00:50:52,660 --> 00:50:54,759 amateur radio apress? 1358 00:50:54,760 --> 00:50:56,209 Anyone heard of April? 1359 00:50:56,210 --> 00:50:58,389 So some apress whatever 1360 00:50:58,390 --> 00:51:00,489 goes with the Ultranet tagging 1361 00:51:00,490 --> 00:51:03,309 it, it goes sent back 1362 00:51:03,310 --> 00:51:05,829 and the users can also 1363 00:51:05,830 --> 00:51:08,739 propose content to be broadcast. 1364 00:51:08,740 --> 00:51:11,019 So files at most 10 1365 00:51:11,020 --> 00:51:13,329 kilobytes each and they 1366 00:51:13,330 --> 00:51:15,759 get screened before 1367 00:51:15,760 --> 00:51:18,999 and moderated to prevent 1368 00:51:19,000 --> 00:51:21,759 whatever malicious or 1369 00:51:21,760 --> 00:51:24,389 whatever. So they moderate and 1370 00:51:24,390 --> 00:51:26,589 the process of moderation and so 1371 00:51:26,590 --> 00:51:27,939 on is not open. 1372 00:51:27,940 --> 00:51:30,159 So that's a very interesting question. 1373 00:51:30,160 --> 00:51:31,870 What gets sent and why? 1374 00:51:34,390 --> 00:51:35,769 Thank you. 1375 00:51:35,770 --> 00:51:38,020 Another question from the Internet. 1376 00:51:41,040 --> 00:51:43,559 OK, so in this case, 1377 00:51:43,560 --> 00:51:44,560 thank you, Dana, 1378 00:51:46,020 --> 00:51:47,609 thanks for your time. 1379 00:51:47,610 --> 00:51:49,889 Thank you for your interesting talk and 1380 00:51:49,890 --> 00:51:51,869 I hope we see each other again. 1381 00:51:51,870 --> 00:51:52,949 Thank you very much indeed. 1382 00:51:54,270 --> 00:51:56,849 Please with an applause tonight 1383 00:51:56,850 --> 00:51:57,850 as.