WEBVTT 1 00:00:00.120 --> 00:00:03.720 You know, we use this digital world every single day, 2 00:00:04.280 --> 00:00:08.800 streaming movies, sending instant messages, even just finding our way 3 00:00:08.839 --> 00:00:10.160 with a quick search on our phones. 4 00:00:10.279 --> 00:00:12.320 Yeah, it's completely woven into our lives. 5 00:00:12.519 --> 00:00:15.720 But how often do we actually stop and think about 6 00:00:15.119 --> 00:00:19.399 the intricate architecture underneath it, all the stuff you don't see. 7 00:00:19.519 --> 00:00:21.199 Not often enough, probably. 8 00:00:21.160 --> 00:00:24.079 So today we're doing a deep dive into exactly the 9 00:00:24.760 --> 00:00:28.879 fascinating foundations of computer networks. They're really the backbone of 10 00:00:28.879 --> 00:00:30.440 our digital lives, right, and. 11 00:00:30.320 --> 00:00:33.560 Our mission here really is to unpack the core ideas 12 00:00:33.960 --> 00:00:36.640 that power everything you do online. We want you to 13 00:00:36.640 --> 00:00:39.320 get not just what makes networks tick, but why they 14 00:00:39.320 --> 00:00:40.840 were built that way and some of. 15 00:00:40.799 --> 00:00:44.159 The ingenious solutions engineers came up with for well some 16 00:00:44.200 --> 00:00:45.359 pretty surprising problems. 17 00:00:45.439 --> 00:00:48.520 Exactly. We're digging into the key insights, basically giving you 18 00:00:48.560 --> 00:00:51.039 a shortcut to being genuinely well informed on this. 19 00:00:51.320 --> 00:00:53.640 And our guide for this is a really foundational book, 20 00:00:53.759 --> 00:00:57.600 Computer Networks Global Edition by Andrew Tannenbaum and Nick Fiemester. 21 00:00:57.960 --> 00:01:01.000 So get ready for some surprise facts, maybe a few 22 00:01:01.439 --> 00:01:05.120 aha moments along the way. Okay, let's start by looking 23 00:01:05.159 --> 00:01:08.280 back a bit. This idea of computers being networked. That 24 00:01:08.359 --> 00:01:10.840 wasn't always the case, was it. Early computers were just 25 00:01:10.879 --> 00:01:13.159 these huge isolated machines. 26 00:01:13.239 --> 00:01:16.640 Oh. Absolutely. For the first couple of decades, computer systems 27 00:01:16.640 --> 00:01:20.040 were like highly centralized, usually stuck in a single room 28 00:01:20.560 --> 00:01:23.799 up and behind glass windows. You know, visitors would just 29 00:01:23.840 --> 00:01:27.239 sort of gawk at the great electronic wonder inside. The 30 00:01:27.359 --> 00:01:32.000 idea that you could have vastly more powerful computers smaller 31 00:01:32.040 --> 00:01:34.920 than a postage stamp pure science fiction back then. 32 00:01:35.200 --> 00:01:37.760 That changed pretty fast, though. What was the main driver? 33 00:01:38.000 --> 00:01:41.280 What pushed us from those isolated giants towards you know, 34 00:01:41.359 --> 00:01:44.280 global conversations, especially the personal media stuff. 35 00:01:44.359 --> 00:01:47.200 Well, it really boiled down to person to person communication 36 00:01:47.519 --> 00:01:49.439 That became the big thing for the twenty first century, 37 00:01:49.519 --> 00:01:51.760 kind of like the telephone was for the nineteenth Ah, 38 00:01:51.840 --> 00:01:55.239 think about email. It just exploded, right and pretty quickly 39 00:01:55.280 --> 00:01:57.480 it started pulling in audio and video too. Then you 40 00:01:57.519 --> 00:02:00.760 had instant messaging. Its roots actually go way back to 41 00:02:00.760 --> 00:02:03.359 a nineteen seventies Unix program. 42 00:02:03.000 --> 00:02:05.680 Called talk Wow the seventies. 43 00:02:05.239 --> 00:02:09.120 Yeah, allowing real time text chat. And then came things 44 00:02:09.199 --> 00:02:12.520 like Twitter, you know, multi person services for short messages, 45 00:02:12.680 --> 00:02:16.439 even video to your friends or well, the entire world. 46 00:02:16.240 --> 00:02:19.280 And with everyone adopting this, totally new uses popped up, 47 00:02:19.280 --> 00:02:21.560 didn't they. Mobile phones becoming central to. 48 00:02:22.479 --> 00:02:26.479 Commerce, exactly, using texts to pay for stuff snacks from 49 00:02:26.520 --> 00:02:30.159 a vending machine, movie tickets and the charge just shows 50 00:02:30.240 --> 00:02:31.199 up on your phone. 51 00:02:30.960 --> 00:02:35.080 Bill or NFC nearfield communication, just tapping your phone to 52 00:02:35.120 --> 00:02:35.840 pay right. 53 00:02:35.800 --> 00:02:38.319 Acting like an RFID smart card. And it was good 54 00:02:38.319 --> 00:02:41.400 for both sides. Stores saved on credit card fees. 55 00:02:41.439 --> 00:02:46.680 And customers got convenience but also power right like price 56 00:02:46.800 --> 00:02:48.280 checking competitors right there in the. 57 00:02:48.280 --> 00:02:50.599 Store precisely, or seeing where else you could buy that 58 00:02:50.680 --> 00:02:54.080 item nearby, maybe cheaper. That kind of instant information access 59 00:02:54.199 --> 00:02:55.199 changed things. 60 00:02:55.039 --> 00:02:57.840 And the uses for mobile and wireless just keep expanding, 61 00:02:57.879 --> 00:03:01.240 sometimes in ways nobody really predicted. Yeah, like sensor networks 62 00:03:01.240 --> 00:03:02.439 in cars, Oh yeah. 63 00:03:02.199 --> 00:03:06.199 That's a fascinating area. Cars gathering data on location, speed, vibration, 64 00:03:06.360 --> 00:03:08.680 how much fuel you're using, uploading it all. 65 00:03:08.840 --> 00:03:10.280 So what do they do with that data? 66 00:03:10.360 --> 00:03:13.400 Well, it can help map out potholes across a whole city, 67 00:03:13.800 --> 00:03:17.319 plan roots to avoid traffic jams based on real time conditions, 68 00:03:17.840 --> 00:03:20.479 or even get this, tell you if you're a gas 69 00:03:20.479 --> 00:03:22.919 guzzler compared to other people driving the same road. 70 00:03:23.159 --> 00:03:26.919 Huh, that's quite specific. It's a whole new level of 71 00:03:27.000 --> 00:03:30.240 data insight. Really, it really is so okay, with all 72 00:03:30.280 --> 00:03:35.159 this complexity, this massive growth, how on earth do engineers 73 00:03:35.199 --> 00:03:39.120 even start designing and managing these huge networks. You can't 74 00:03:39.159 --> 00:03:40.919 just lump it all together, surely. 75 00:03:40.639 --> 00:03:43.639 No, you definitely can't. And the elegant solution they came 76 00:03:43.680 --> 00:03:46.680 up with is called layering. It's fundamental layering, Yeah, to 77 00:03:46.719 --> 00:03:49.759 make the design manageable. Most networks are organized as a 78 00:03:49.800 --> 00:03:52.520 stack of layers or levels. Each one is built on 79 00:03:52.560 --> 00:03:55.319 top of the one below it, and the key idea, 80 00:03:55.439 --> 00:03:58.520 the genius of it, is that each layer offers specific 81 00:03:58.560 --> 00:04:01.080 services to the layer above it, while hiding all the 82 00:04:01.120 --> 00:04:03.840 messy details of how it actually provides those services. 83 00:04:04.199 --> 00:04:07.000 So each layer doesn't need to know the inner workings 84 00:04:07.000 --> 00:04:08.159 of the others exactly. 85 00:04:08.159 --> 00:04:10.520 It's like each layer is a kind of virtual machine 86 00:04:10.560 --> 00:04:12.439 just serving the layer directly above it. 87 00:04:12.719 --> 00:04:17.079 So it's sort of like sending a package overseas. Is 88 00:04:17.079 --> 00:04:17.920 that a fair analogy? 89 00:04:18.079 --> 00:04:21.199 That's a great analogy. Yeah, think about it. One layer, 90 00:04:21.439 --> 00:04:25.439 maybe the application prepares the contents. Another layer, like transport, 91 00:04:25.759 --> 00:04:29.360 packages it up securely, puts the address on. Then maybe 92 00:04:29.360 --> 00:04:33.160 a network layer handles the customs forms, the cross border stuff, gotcha. 93 00:04:33.560 --> 00:04:36.399 And finally, the physical layer is what actually moves the 94 00:04:36.439 --> 00:04:40.040 box across the ocean or through the air. Each step 95 00:04:40.079 --> 00:04:42.959 has its own rules, its own job, and it doesn't 96 00:04:43.000 --> 00:04:45.240 need to know the nitty gritty of the other steps. 97 00:04:45.480 --> 00:04:50.800 And below that lowest layer, that's the actual wires. 98 00:04:50.720 --> 00:04:54.279 Or airwaves precisely, that's the physical medium where the actual 99 00:04:54.319 --> 00:04:57.279 communication happens, the real transmission of signals. 100 00:04:57.360 --> 00:04:59.560 So how does this connect to the internet we use? 101 00:05:00.000 --> 00:05:04.240 This layering concept leads us straight to the TCPIP reference model. 102 00:05:04.319 --> 00:05:07.040 That's the actual architecture running the global network you're using 103 00:05:07.079 --> 00:05:10.959 right now. Okay, TCPIP, I've heard that its ancestor really 104 00:05:11.040 --> 00:05:14.079 was the Arpinent, a research network funded by the US 105 00:05:14.079 --> 00:05:16.959 Department of Defense, and it was designed right from the 106 00:05:17.000 --> 00:05:21.000 start to seamlessly connect multiple different networks together. 107 00:05:20.879 --> 00:05:23.839 Like connecting networks was the goal. Now here's something that 108 00:05:23.920 --> 00:05:29.439 surprised me from the book. Why TCPIP became the standard. 109 00:05:29.560 --> 00:05:32.399 It wasn't necessarily because it was the best theoretical model. 110 00:05:32.839 --> 00:05:34.240 There was another one OSI. 111 00:05:34.600 --> 00:05:37.000 That's a fascinating piece of history, isn't it. Yeah, the 112 00:05:37.000 --> 00:05:40.800 OSI model was a major contender, very thoroughly designed. But 113 00:05:40.839 --> 00:05:43.279 the book points out that one of the first implementations 114 00:05:43.279 --> 00:05:46.439 of TCPIP was part of Berkeley Unity X that was 115 00:05:46.519 --> 00:05:49.759 quite good and free, not to mention free, Yeah, exactly. 116 00:05:50.199 --> 00:05:54.319 That practical reality was huge. People started using it, improving it, building. 117 00:05:54.000 --> 00:05:55.839 On it, so it wasn't just theory, it was availability 118 00:05:55.879 --> 00:05:57.319 and usability pretty much. 119 00:05:57.439 --> 00:06:01.399 Early OSI implementations, the book notes, were off, huge, unwieldy, 120 00:06:01.439 --> 00:06:05.319 and slow, so people started associating OSI with poor quality. 121 00:06:05.680 --> 00:06:10.800 Fairly or unfairly. TCPIP just worked and its spread, so that. 122 00:06:10.800 --> 00:06:15.040 Layered approach, specifically tcpip's implementation. That's why my phone can 123 00:06:15.079 --> 00:06:17.240 talk to a server halfway across the world without breaking 124 00:06:17.240 --> 00:06:17.759 a sweat. 125 00:06:17.920 --> 00:06:21.439 That's the magic. Yes, your apps don't need to know 126 00:06:21.639 --> 00:06:24.560 about the fiber optic cables or the satellite links or 127 00:06:24.560 --> 00:06:28.120 the Wi Fi signals involved. The layers handle all that 128 00:06:28.199 --> 00:06:29.360 complexity underneath. 129 00:06:29.480 --> 00:06:31.759 Okay, so we've got this layered system. But let's go 130 00:06:31.879 --> 00:06:34.879 right down to the bottom that physical layer. How does 131 00:06:34.959 --> 00:06:37.839 the actual data the ones and next, how do they 132 00:06:38.000 --> 00:06:39.839 literally get from one device to another? 133 00:06:40.040 --> 00:06:42.800 Right? Good question. Down at the physical layer, you're dealing 134 00:06:42.800 --> 00:06:46.120 with the actual transmission media, and there's a real diversity there, 135 00:06:46.160 --> 00:06:49.600 each with its own characteristics and challenges. 136 00:06:49.839 --> 00:06:52.040 Like the wires the cables we actually see. 137 00:06:51.879 --> 00:06:56.079 Sometimes exactly for wired connections, you've got twisted pair cabling. 138 00:06:56.399 --> 00:06:59.040 Think of your standard Ethernet cable plugging into your computer 139 00:06:59.160 --> 00:07:02.560 or router. It's super common, widely used because it's cheap 140 00:07:02.600 --> 00:07:05.480 and the performance is adequate for shorter runs within a 141 00:07:05.480 --> 00:07:08.759 building or home hundreds of megabits per second easily. 142 00:07:09.000 --> 00:07:13.000 Then there's the cable that brings TV or Internet into 143 00:07:13.040 --> 00:07:13.399 the house. 144 00:07:13.560 --> 00:07:17.360 Yeah, coaxial coaxial pable. Yeah, it's got that solid core 145 00:07:17.439 --> 00:07:21.800 wire insulation than a braided shield. That structure gives it 146 00:07:21.879 --> 00:07:25.639 a good combination of high bandwidth and excellent noise immunity. 147 00:07:25.720 --> 00:07:27.600 It resists interference. 148 00:07:27.040 --> 00:07:29.439 Well, so it's tougher than twisted pair generally. 149 00:07:29.600 --> 00:07:31.920 Yes, it used to be used for long distance phone lines, 150 00:07:31.959 --> 00:07:35.319 but fiber optics mostly took over that role. Still, it's 151 00:07:35.439 --> 00:07:38.439 very common for cable TV and bringing Internet into homes 152 00:07:38.560 --> 00:07:40.120 metropolitan area networks. 153 00:07:40.560 --> 00:07:44.600 And then the really high speed stuff relies on fiber optics, right, 154 00:07:44.639 --> 00:07:46.560 sending light down glass threads. 155 00:07:46.240 --> 00:07:49.480 That's the one using light pulses through these ultra thin 156 00:07:49.680 --> 00:07:52.839 fibers of glass. It works because of a principle called 157 00:07:52.959 --> 00:07:57.160 total internal reflection, which basically traps the light inside the fiber, 158 00:07:57.240 --> 00:08:00.839 letting it travel long distances. And the potential bandwidth is 159 00:08:01.000 --> 00:08:04.560 just mind blowing. The book mentions figures in excess of 160 00:08:04.600 --> 00:08:07.720 fifty thousand gbps. That's fifty terabits per second. 161 00:08:07.759 --> 00:08:08.639 Fifty terabits. 162 00:08:08.720 --> 00:08:12.160 Yeah, and we're apparently nowhere near reaching these limits. The 163 00:08:12.199 --> 00:08:15.560 current practical limits maybe around one hundred gbps per fiber, 164 00:08:15.800 --> 00:08:18.199 are more about how fast we can convert electricity to 165 00:08:18.279 --> 00:08:19.160 light and back again. 166 00:08:19.480 --> 00:08:22.680 Incredible. Okay, so that's wires and light. What about wireless 167 00:08:23.639 --> 00:08:24.600 through the air? Right? 168 00:08:24.680 --> 00:08:28.439 Wireless connections, You've got radio waves and different frequency bands 169 00:08:28.439 --> 00:08:33.519 behave differently lower frequencies vlf, LFMF. They pass through buildings easily, 170 00:08:33.879 --> 00:08:36.320 which is why your radio works indoors. 171 00:08:35.919 --> 00:08:37.559 But they don't carry much data, right. 172 00:08:37.440 --> 00:08:42.159 Exactly, low bandwidth for data transmission. Higher frequencies carry more 173 00:08:42.240 --> 00:08:44.720 data but might struggle with obstacles. 174 00:08:44.799 --> 00:08:47.279 And the main one we all use daily is Wi 175 00:08:47.279 --> 00:08:49.120 Fi eighth two point eleven. 176 00:08:49.240 --> 00:08:53.240 That's the dominant standard for wireless local networks. Yes, and 177 00:08:53.279 --> 00:08:57.440 it uses a system called csmcka carrier sense multiple access 178 00:08:57.440 --> 00:09:00.879 with collision avoidance, trying to prevent device is transmitting at 179 00:09:00.879 --> 00:09:02.360 the same time and causing. 180 00:09:02.000 --> 00:09:03.320 Collisions trying to avoid them. 181 00:09:03.440 --> 00:09:04.679 Does it always work well? 182 00:09:04.720 --> 00:09:07.679 That leads to a really interesting kind of counterintuitive issue, 183 00:09:07.679 --> 00:09:09.080 the hidden terminal problem. 184 00:09:09.159 --> 00:09:11.000 In terminal, okay, what's that match you. 185 00:09:11.039 --> 00:09:15.480 You have three laptops A, B, and C. Laptop A 186 00:09:15.639 --> 00:09:17.840 and laptop C are too far apart. They can't hear 187 00:09:17.840 --> 00:09:19.039 each other's radio signals. 188 00:09:19.039 --> 00:09:21.559 Okay at a range, but laptop B is in the 189 00:09:21.600 --> 00:09:24.840 metal within range of both A and C. Now suppose 190 00:09:24.919 --> 00:09:28.759 A wants to send data to B. It listens, here's nothing, 191 00:09:28.840 --> 00:09:32.120 and starts transmitting. Makes sense, But C also wants to 192 00:09:32.120 --> 00:09:35.799 send to B. C listens, doesn't hear A because A 193 00:09:35.960 --> 00:09:38.320 is hidden from it, thinks the coast is clear, and 194 00:09:38.399 --> 00:09:39.799 also starts transmitting to B. 195 00:09:40.200 --> 00:09:42.240 Uh. Oh, so B gets signals from both A and 196 00:09:42.279 --> 00:09:43.000 C at the same. 197 00:09:42.799 --> 00:09:46.480 Time exactly, A collision happens at B and probably neither 198 00:09:46.519 --> 00:09:49.960 transmission gets through. Queenly, A and C didn't know about 199 00:09:49.960 --> 00:09:52.240 each other, but they interfered with each other at the receiver. 200 00:09:52.480 --> 00:09:54.960 Huh. That is tricky. So how do they fix that? 201 00:09:55.120 --> 00:09:56.919 You can't just hope it doesn't happen, right. 202 00:09:57.320 --> 00:10:01.159 Wi Fi has an optional mechanism called RTSC tests request 203 00:10:01.320 --> 00:10:05.559 to send clear descent RTSCTS. Okay, so laptop A, before 204 00:10:05.600 --> 00:10:08.120 sending its big chunk of data to B, first sends 205 00:10:08.120 --> 00:10:11.039 a very short RTS frame to B requesting to send 206 00:10:11.120 --> 00:10:13.399 got it. If B here's that and is ready, it 207 00:10:13.480 --> 00:10:17.279 broadcasts back a short CTS frame clear descent. Now here's 208 00:10:17.320 --> 00:10:21.000 the clever part. Laptop C can hear B. So when 209 00:10:21.039 --> 00:10:23.759 B sends a CTS, oh see, here's the clear descend 210 00:10:23.799 --> 00:10:25.840 meant for A, and knows someone is about to transmit 211 00:10:25.840 --> 00:10:28.039 to B, even if it couldn't hear the original request 212 00:10:28.039 --> 00:10:31.200 from A. Precisely, the CTS acts like a do not 213 00:10:31.279 --> 00:10:36.279 disturb sign for anyone within d's range. C virtually senses 214 00:10:36.759 --> 00:10:39.720 the channel is busy and holds off its transmission, avoiding 215 00:10:39.720 --> 00:10:40.320 the collision. 216 00:10:40.840 --> 00:10:44.080 That's really clever, using the central node B to signal 217 00:10:44.080 --> 00:10:47.679 everyone else. It's amazing that different techniques needed just for 218 00:10:47.720 --> 00:10:51.919 the physical layer, isn't it wires, light, radio waves avoiding 219 00:10:52.000 --> 00:10:52.799 hidden collisions. 220 00:10:52.840 --> 00:10:55.519 It's a whole world of specialized engineering just to get 221 00:10:55.559 --> 00:10:58.080 those bits reliably from one point to another. 222 00:10:58.159 --> 00:11:01.159 Okay, so we've got bits moving physically. But we mentioned 223 00:11:01.200 --> 00:11:03.919 packets earlier. How does a packet of data send from 224 00:11:03.919 --> 00:11:07.120 my phone navigate the entire Internet, potentially crossing dozens of 225 00:11:07.159 --> 00:11:11.279 networks to find one specific server somewhere else. That seems 226 00:11:11.279 --> 00:11:12.440 like a huge challenge. 227 00:11:12.519 --> 00:11:14.960 It is, and that's where the Internet Protocol or IP, 228 00:11:15.200 --> 00:11:16.200 really comes into its own. 229 00:11:16.240 --> 00:11:17.519 Yeah, p addresses, right. 230 00:11:17.360 --> 00:11:19.840 IP addresses are part of it. Yes, IP provides the 231 00:11:19.840 --> 00:11:23.960 fundamental addressing and routing mechanism. It offers a best effort 232 00:11:24.000 --> 00:11:26.639 and that's a key term, best effort way to transport 233 00:11:26.679 --> 00:11:28.440 packets from source to destination. 234 00:11:28.679 --> 00:11:32.679 Best effort meaning it tries, but no promises pretty much. Yeah. 235 00:11:32.679 --> 00:11:36.639 It doesn't guarantee delivery or order or protect against errors. 236 00:11:37.000 --> 00:11:39.240 Its main job is to get the packet towards the 237 00:11:39.279 --> 00:11:44.039 destination hop by hop across different networks, without needing to 238 00:11:44.080 --> 00:11:47.879 know the specifics of those networks. It's the universal translator 239 00:11:48.000 --> 00:11:48.440 in a way. 240 00:11:48.600 --> 00:11:52.679 We mostly use IPv four addresses, those familiar numbers. 241 00:11:52.399 --> 00:11:55.519 Correct the thirty two bit addresses, but as you probably know, 242 00:11:55.559 --> 00:11:57.360 we've basically run out of them. It's like running out 243 00:11:57.360 --> 00:11:58.440 of phone numbers for the planet. 244 00:11:58.600 --> 00:12:01.279 Yeah, I've heard about that. So this polution is IPv six. 245 00:12:01.399 --> 00:12:04.879 IPv six, Yes, with its massive one hundred and twenty 246 00:12:04.879 --> 00:12:08.679 eight bit addresses, the address space is practically infinite for 247 00:12:08.720 --> 00:12:12.240 the foreseeable future. It's actually been an official Internet standard 248 00:12:12.320 --> 00:12:13.480 since nineteen ninety eight. 249 00:12:13.600 --> 00:12:16.320 Nineteen ninety eight, but we're still mostly on IPv four. 250 00:12:16.440 --> 00:12:19.480 Yeah, adoption has been slower than expected. The book says 251 00:12:19.519 --> 00:12:22.360 itv six is deployed and used in only about twenty 252 00:12:22.399 --> 00:12:25.360 five percent of the Internet. Even now. Inertia is a 253 00:12:25.399 --> 00:12:26.639 powerful thing, and. 254 00:12:26.559 --> 00:12:30.200 That scarcity of the old IPv four addresses, it's actually 255 00:12:30.240 --> 00:12:32.039 created like a market for them. Oh. 256 00:12:32.080 --> 00:12:35.799 Absolutely, get this. The book mentions each IPv four address 257 00:12:35.919 --> 00:12:38.360 is now worth as much as nineteen dollars. 258 00:12:38.000 --> 00:12:40.519 Nineteen dollars for a single IP address yep. 259 00:12:41.039 --> 00:12:43.440 And it even cites a case in twenty nineteen where 260 00:12:43.440 --> 00:12:47.000 a man was convicted for illegally stockpiling seven hundred and 261 00:12:47.120 --> 00:12:50.960 fifty thousand IP addresses, worth about fourteen million dollars at 262 00:12:51.000 --> 00:12:53.000 the time, and selling them on the black market. 263 00:12:53.480 --> 00:12:57.679 Wow, that really shows the tangible value of these abstract 264 00:12:57.759 --> 00:12:58.799 numbers that run the Internet. 265 00:12:58.879 --> 00:12:59.720 It definitely does. 266 00:13:00.120 --> 00:13:03.679 But back to best effort. If IP doesn't guarantee delivery. 267 00:13:03.879 --> 00:13:06.559 How do we handle things where reliability is crucial, like 268 00:13:06.840 --> 00:13:09.360 checking my bank balance or a video call where I 269 00:13:09.440 --> 00:13:12.480 need things in order. Lost packets would be bad, right. 270 00:13:12.600 --> 00:13:15.759 You need something on top of IP to provide that reliability. 271 00:13:15.799 --> 00:13:18.919 And that's what the Transmission control Protocol or TCP comes in. 272 00:13:19.000 --> 00:13:20.480 TCP, okay, works with IP. 273 00:13:20.440 --> 00:13:24.720 Exactly, TCPIP they work as a pair. TCP is connection oriented. 274 00:13:24.960 --> 00:13:27.120 Think of it like making a phone call house. You 275 00:13:27.159 --> 00:13:29.799 first establish a connection like dialing and the other person 276 00:13:29.879 --> 00:13:32.919 picking up. Then you exchange your data over that established connection. 277 00:13:33.120 --> 00:13:36.159 Then you explicitly release the connection hanging up. 278 00:13:36.200 --> 00:13:38.080 Okay, So there's a setup and tear down. 279 00:13:38.399 --> 00:13:42.159 Yes, And during that connection, TCP works hard to create 280 00:13:42.200 --> 00:13:46.440 a reliable ordered byte stream. If IP drops a packet, 281 00:13:46.519 --> 00:13:50.240 PCP notices and retransmits it. If packets arrive out of order, 282 00:13:50.320 --> 00:13:52.960 TCP puts them back in the right sequence before handing 283 00:13:53.000 --> 00:13:53.960 them up to the application. 284 00:13:54.200 --> 00:13:56.639 So it fixes the best effort problems of IP. 285 00:13:56.960 --> 00:13:59.759 It does. It also handles flow control, making sure a 286 00:13:59.759 --> 00:14:03.600 fact sender doesn't overwhelm a slow receiver, and congestion control 287 00:14:03.799 --> 00:14:06.480 trying to prevent the network itself from getting overloaded. 288 00:14:06.759 --> 00:14:10.159 It's quite sophisticated, but not everything uses TCP. Right, There's 289 00:14:10.200 --> 00:14:12.279 another one, UDP, correct. 290 00:14:12.039 --> 00:14:15.559 User Data Ground Protocol or UDP. It's the lightweight alternative. 291 00:14:15.679 --> 00:14:18.639 It's connectionless, no setup or tear down needed, so it 292 00:14:18.720 --> 00:14:22.279 just sends pretty much. UDP does almost nothing beyond sending 293 00:14:22.320 --> 00:14:25.399 packets between applications. As the book puts it, it relies 294 00:14:25.440 --> 00:14:29.440 on ip's best effort delivery and doesn't add reliability, ordering 295 00:14:29.639 --> 00:14:30.840 or flow control itself. 296 00:14:30.879 --> 00:14:33.440 Why would you want that? Less reliable sounds. 297 00:14:33.200 --> 00:14:37.120 Bad because it's faster, no connection set up overhead, no 298 00:14:37.279 --> 00:14:41.720 waiting for acknowledgments or retransmissions. For things like say online 299 00:14:41.759 --> 00:14:45.799 gaming or streaming live video or audio, speed is often 300 00:14:45.840 --> 00:14:47.759 more critical than perfect reliability. 301 00:14:48.000 --> 00:14:51.240 Ah okay, If a single frame of video drops in 302 00:14:51.279 --> 00:14:54.519 a stream, it's maybe a slight glitch, but the stream continues. 303 00:14:54.960 --> 00:14:58.360 Waiting for a retransmission might cause a long freeze, which 304 00:14:58.399 --> 00:14:59.639 is worse exactly. 305 00:14:59.720 --> 00:15:02.960 The application itself might handle some level of error correction 306 00:15:03.200 --> 00:15:06.960 or just tolerate minor losses. So UDP is perfect for 307 00:15:07.000 --> 00:15:08.879 those kinds of real time applications. 308 00:15:08.960 --> 00:15:12.000 So that combination IP for the basic addressing and routing, 309 00:15:12.440 --> 00:15:15.399 then TCP for reliability when you need it and UDP 310 00:15:15.519 --> 00:15:17.879 for speed. When that's the priority that covers a huge 311 00:15:17.960 --> 00:15:18.360 range of. 312 00:15:18.399 --> 00:15:22.039 Uses, it's an incredibly flexible and powerful combination. It really 313 00:15:22.080 --> 00:15:25.639 is the invisible glue holding our incredibly diverse digital world together, 314 00:15:26.000 --> 00:15:29.759 allowing everything from secure banking to watching cat videos. 315 00:15:29.799 --> 00:15:32.960 Okay, so all these layers, all these protocols, they eventually 316 00:15:33.039 --> 00:15:35.799 enable the applications we actually use every day, often without 317 00:15:35.840 --> 00:15:37.519 giving the underlying tech a second thought. 318 00:15:37.840 --> 00:15:42.039 Like the Worldwide Web, absolutely tim berners Lee's invention. Now 319 00:15:42.080 --> 00:15:46.639 this vast worldwide collection of content, and you have organizations 320 00:15:46.720 --> 00:15:50.440 like the W three C, the Worldwide Web Consortium, working 321 00:15:50.440 --> 00:15:54.480 to develop standards and protocols so it all hopefully works together, 322 00:15:54.679 --> 00:15:55.039 and our. 323 00:15:54.960 --> 00:15:56.720 Browsers are doing a lot of work behind the scenes 324 00:15:56.759 --> 00:15:59.919 fetching and displaying stuff. And it's not just static pages anymore. 325 00:16:00.840 --> 00:16:04.159 No, much of the web is highly dynamic, meaning programs 326 00:16:04.200 --> 00:16:07.759 running on servers generate web pages specifically for you based 327 00:16:07.799 --> 00:16:11.240 on your request or your profile. Think online shopping, social 328 00:16:11.279 --> 00:16:11.960 media feeds. 329 00:16:12.120 --> 00:16:14.320 Right, And none of that works without being able to 330 00:16:14.320 --> 00:16:17.080 find the server in the first place, which brings us 331 00:16:17.080 --> 00:16:19.000 to DNS, the domain name. 332 00:16:18.919 --> 00:16:21.679 System, crucial piece of the puzzle. When you type a 333 00:16:21.720 --> 00:16:25.639 website name like www dot example dot com. Your computer 334 00:16:25.720 --> 00:16:28.240 has no idea where that is. It needs the numerical 335 00:16:28.279 --> 00:16:30.600 IP address like a phone book for the Internet kind 336 00:16:30.600 --> 00:16:33.799 of yeah. DNS is described as a hierarchical naming scheme 337 00:16:33.840 --> 00:16:37.320 and a distributed database system. It translates those human friendly 338 00:16:37.399 --> 00:16:41.320 names into the IP addresses that routers actually understand. Without DNS, 339 00:16:41.440 --> 00:16:43.559 browsing would mean memorizing strings of numbers. 340 00:16:43.840 --> 00:16:46.559 No, definitely wouldn't have caught on as well. But this 341 00:16:46.840 --> 00:16:50.720 very openness, this interconnectedness that makes the Internet so powerful, 342 00:16:51.519 --> 00:16:54.440 it also creates big challenges, doesn't it, particularly around security 343 00:16:54.519 --> 00:16:55.960 and privacy. 344 00:16:56.039 --> 00:16:59.960 That's the double edged sword. Yes, the design facilitates community, 345 00:17:00.480 --> 00:17:04.440 but also potential misuse. One major issue highlighted is the 346 00:17:04.480 --> 00:17:08.359 distributed denial of service attack ds DS. 347 00:17:08.400 --> 00:17:09.720 What exactly is happening there? 348 00:17:09.799 --> 00:17:13.359 It's where attackers get many machines on the network, often 349 00:17:13.440 --> 00:17:17.160 compromise computers or devices to send traffic towards a victim 350 00:17:17.160 --> 00:17:20.200 machine in an attempt to exhaust its resources, flood the 351 00:17:20.240 --> 00:17:22.400 target so legitimate users can't get through. 352 00:17:22.599 --> 00:17:24.880 And where do these attacking machines come from? Traditionally it 353 00:17:24.920 --> 00:17:26.200 was hacked PCs. 354 00:17:25.839 --> 00:17:29.240 Right, Historically, yes, botnets of infected computers, But the book 355 00:17:29.279 --> 00:17:32.559 points out a major new vector the proliferation of insecure 356 00:17:32.599 --> 00:17:36.279 IoT devices Internet of things devices. 357 00:17:35.720 --> 00:17:38.160 Like smart light bulbs, security. 358 00:17:37.720 --> 00:17:42.160 Cameras, exactly, thermostats, appliances, anything connected to the Internet. And 359 00:17:42.200 --> 00:17:45.039 here's the really startling bit from the source material. Can 360 00:17:45.079 --> 00:17:48.599 a coordinated attack by a million Internet connected smart toasters 361 00:17:48.680 --> 00:17:49.519 take down Google? 362 00:17:49.839 --> 00:17:51.400 Seriously smart toasters? 363 00:17:51.599 --> 00:17:55.400 It sounds absurd, but the underlying point is serious. The 364 00:17:55.400 --> 00:17:58.160 book states, unfortunately that much of the IoT industry in 365 00:17:58.200 --> 00:18:01.640 particular is unconcerned with software security. They just want to 366 00:18:01.720 --> 00:18:04.000 ship cheap devices quickly, so they. 367 00:18:03.920 --> 00:18:07.440 Become easy targets to rope into these massive attacks. 368 00:18:07.039 --> 00:18:10.720 Precisely, and it leaves the network operators, the ISPs, and 369 00:18:10.759 --> 00:18:14.799 others trying to defend against potentially huge floods of traffic 370 00:18:15.039 --> 00:18:19.359 coming from these everyday insecure gadgets. It's a massive headache. 371 00:18:19.400 --> 00:18:22.960 Wow. Okay, beyond attacks, there's the privacy angle too. With 372 00:18:23.119 --> 00:18:23.960 everything connected. 373 00:18:24.160 --> 00:18:27.480 Yeah, data collection is pervasive. As the book says, it 374 00:18:27.519 --> 00:18:30.759 is becoming increasingly easier for various parties to collect data 375 00:18:30.799 --> 00:18:32.240 about how each of us uses the. 376 00:18:32.200 --> 00:18:34.319 Network and who are these various parties. 377 00:18:34.400 --> 00:18:37.559 It's a long list, your internet service provider, your mobile 378 00:18:37.599 --> 00:18:43.599 phone carrier, applications, websites, cloud hosting services, content delivery networks, 379 00:18:43.759 --> 00:18:49.359 device manufacturers, advertisers, and web tracking software vendors. Basically almost 380 00:18:49.359 --> 00:18:52.680 everyone involved in delivering your online experience could be collecting data. 381 00:18:52.920 --> 00:18:55.319 That's a lot of potential watchers. Now, just to be 382 00:18:55.400 --> 00:18:58.160 really clear for our listeners, our aim here is just 383 00:18:58.200 --> 00:19:00.759 to explain the technology and the challenge and just discussed 384 00:19:00.759 --> 00:19:03.559 in the field as presented in our source. We're not 385 00:19:03.599 --> 00:19:05.880 taking a stance on the debates around data collection or 386 00:19:05.920 --> 00:19:07.400 IoT security ourselves. 387 00:19:07.440 --> 00:19:11.039 Absolutely, We're just unpacking the technical realities and the issues 388 00:19:11.119 --> 00:19:14.559 network engineers and researchers grapple with based on the text. 389 00:19:14.640 --> 00:19:17.400 So wrapping this section up, it feels like the digital 390 00:19:17.440 --> 00:19:22.279 age is this constant balancing act, incredible innovation and convenience 391 00:19:22.319 --> 00:19:23.519 on one side. 392 00:19:23.240 --> 00:19:26.720 And this ever present, always evolving need for security and 393 00:19:26.759 --> 00:19:29.880 privacy on the other, and it's all fundamentally shaped by 394 00:19:29.920 --> 00:19:33.319 the network architecture itself. It's a dynamic, ongoing tension. 395 00:19:33.440 --> 00:19:34.920 Yeah. Absolutely, So when you. 396 00:19:34.880 --> 00:19:37.799 Really connect all these dots, the digital experiences you have 397 00:19:37.920 --> 00:19:41.200 every single day, that simple message you send, the complex 398 00:19:41.279 --> 00:19:44.599 movie you stream, they're all built on this incredibly complex, 399 00:19:44.920 --> 00:19:49.640 often invisible, but really clever foundation of network architecture. 400 00:19:49.720 --> 00:19:51.160 It really is invisible, isn't it. 401 00:19:51.440 --> 00:19:55.160