1 00:00:00,340 --> 00:00:02,260 Hello, and welcome to the transport layer. 2 00:00:02,260 --> 00:00:06,0 We are working our way up from the bottom of our four layer architecture up through 3 00:00:06,0 --> 00:00:09,290 the top. So we're kind of at the half way point. 4 00:00:09,290 --> 00:00:12,306 We have covered the link layer, like ethernet, and then we covered the 5 00:00:12,306 --> 00:00:16,120 internetwork layer which is kind of like the postcard layer. 6 00:00:16,120 --> 00:00:19,890 So lets, lets review the magic of IP, right. 7 00:00:19,890 --> 00:00:24,375 The Magic IP is this postcard layer that bounces these packets with From addresses 8 00:00:24,375 --> 00:00:28,600 and To addresses, through you know, 15 or so hops, getting them their, getting 9 00:00:28,600 --> 00:00:33,150 packets to their destination network from one network to another network as well as 10 00:00:33,150 --> 00:00:37,893 it can. And then when it gets to the destination 11 00:00:37,893 --> 00:00:40,379 network it finds the final computer on that network. 12 00:00:43,20 --> 00:00:48,780 The magic of this is there is no interim long term storage inside the network. 13 00:00:48,780 --> 00:00:52,530 All the long term storage is outside, and now we're going to talk about that. 14 00:00:52,530 --> 00:00:55,430 So one of the things that makes IP so fast is that it is not demanded to be 15 00:00:55,430 --> 00:00:58,280 perfect, it is not demanded to, to deliver data in order, and it's not 16 00:00:58,280 --> 00:01:04,190 demanded to, it, it's, there's no requirement that it doesn't lose data. 17 00:01:04,190 --> 00:01:07,390 Its fast and barely ever loses data and when it does there is a layer to recover 18 00:01:07,390 --> 00:01:10,295 and that's we are going to talk about next. 19 00:01:10,295 --> 00:01:15,805 So the internet protocol is this multi-hop packets can take multiple 20 00:01:15,805 --> 00:01:22,710 paths, they can make use of all kinds of crazy links. 21 00:01:22,710 --> 00:01:25,840 But now, we're going to move up one, right? 22 00:01:25,840 --> 00:01:32,8 We're going to move up to the TCP layer. So the TCP layer is both simple and 23 00:01:32,8 --> 00:01:35,291 complex. The purpose of the TCP layer is to 24 00:01:35,291 --> 00:01:39,260 compensate for the possible errors in the IP layer as well as make best use of 25 00:01:39,260 --> 00:01:45,550 available resources. So if the network, if the overall network 26 00:01:45,550 --> 00:01:50,200 from the overall network from here to here is extremely fast we want to send 27 00:01:50,200 --> 00:01:55,386 data really fast right. If on the other hand the overall network 28 00:01:55,386 --> 00:01:58,610 from here to here is really slow we want to send data slowly and be efficient 29 00:01:58,610 --> 00:02:04,112 cause remember part of the goal of the. TCP IP networking is to share effectively 30 00:02:04,112 --> 00:02:08,620 and so we need to be aware of whether our network is fast or slow. 31 00:02:08,620 --> 00:02:11,150 And those are the kind of problems we solve with a TSP layer. 32 00:02:11,150 --> 00:02:16,64 How fast is the underlying network, how reliable is it and if something goes 33 00:02:16,64 --> 00:02:23,706 wrong what do we do to deal with that. So the key idea in TCP/IP is that when we 34 00:02:23,706 --> 00:02:29,380 send some data, we break it into packets and then we send each one. 35 00:02:29,380 --> 00:02:34,645 And then we keep them until they get an acknowledgement from the other side and 36 00:02:34,645 --> 00:02:41,394 then and only then do we throw them away. And at some point, if a packet gets lost 37 00:02:41,394 --> 00:02:45,682 It can be sent again, and again, and again until it finally is acknowledged in 38 00:02:45,682 --> 00:02:51,410 the destination system. And so that's basically what TCP does, is 39 00:02:51,410 --> 00:02:58,120 it figures out which packets have or have not made it across the Internet layer. 40 00:02:58,120 --> 00:03:01,237 So here's an example scenario. So we got a message that's broken up into 41 00:03:01,237 --> 00:03:04,564 five packets. We've got sort of the first hundred 42 00:03:04,564 --> 00:03:09,422 characters, the, the, you know, first characters, the second, and third packet. 43 00:03:09,422 --> 00:03:16,330 Now, what TCP does, is it speculatively sends a few packets. 44 00:03:16,330 --> 00:03:19,782 Says okay, let's get ahead. Cause if you sent one and waited Then you 45 00:03:19,782 --> 00:03:22,914 might not make best use of the network, so you kind of guess and dump a few 46 00:03:22,914 --> 00:03:27,6 packets out there. And so you send, let's say we send three 47 00:03:27,6 --> 00:03:30,790 packets, cue them up for sending. And as fast as we can send them out the 48 00:03:30,790 --> 00:03:35,90 back of our network, we start spooling them out the back of their network. 49 00:03:35,90 --> 00:03:38,290 And somehow, two of them make it across the internet, to the destination, but 50 00:03:38,290 --> 00:03:41,540 somehow that poor second packet, just like every time I'm doing a lecture, the 51 00:03:41,540 --> 00:03:46,257 poor second packet never gets it. You don't want to be the second packet in 52 00:03:46,257 --> 00:03:49,90 one of my lectures, 'cause you're going to get it. 53 00:03:49,90 --> 00:03:54,74 But somehow, the second packet has gotten vaporized. 54 00:03:54,74 --> 00:03:58,791 Now, time passes. And[COUGH], the receiver gets this s-, 55 00:03:58,791 --> 00:04:04,520 s-, this sensation that maybe, maybe, maybe something's missing. 56 00:04:04,520 --> 00:04:08,178 And so it sends and acknowledgement back, and it says, I'm going to send a note 57 00:04:08,178 --> 00:04:13,463 back that says, look, I'm ready for 200. I do have 100, and I'm, I'm ready for 58 00:04:13,463 --> 00:04:16,385 200. Matter of fact I might even through away 59 00:04:16,385 --> 00:04:20,877 300 cause it's been so long. So I want you to start me over at 200. 60 00:04:20,877 --> 00:04:23,680 [COUGH] But sender all of a sudden knows that 100 has been sent. 61 00:04:23,680 --> 00:04:28,450 And so it knows that is can throw this one away now right it's been 62 00:04:28,450 --> 00:04:33,268 acknowledged. So there's an acknowledgement stream 63 00:04:33,268 --> 00:04:37,684 going back and forth, so now the sender sends 200 and 300 and then they make it a 64 00:04:37,684 --> 00:04:45,150 cost, the receiver says I have got 200. Thus the 300 and then they send 400 65 00:04:45,150 --> 00:04:50,170 speculatively. 300 makes it, 400 makes it and receiver 66 00:04:50,170 --> 00:04:54,70 says. I got 400, which means now it can cross 67 00:04:54,70 --> 00:04:58,670 off 300 and 400, right? Check those off and throw them away, and 68 00:04:58,670 --> 00:05:03,80 then send 500 cross, and at some point check those off. 69 00:05:03,80 --> 00:05:07,250 It's got 500, and now the sender can sort of like empty, check everything off. 70 00:05:07,250 --> 00:05:10,140 Everything's been sent, and it's been acknowledged, and we know. 71 00:05:10,140 --> 00:05:14,361 And so that's kind of an oversimplified view of the kind of bookkeeping It's 72 00:05:14,361 --> 00:05:18,380 going on, on both sides of a TCP connection. 73 00:05:18,380 --> 00:05:24,906 And so this, right here is in a sense the genius of the internet. 74 00:05:24,906 --> 00:05:30,210 The storage requirements in the middle, the routers, this is the IP basically the 75 00:05:30,210 --> 00:05:37,470 network of networks, they, we didn't design these to require a lot of storage. 76 00:05:37,470 --> 00:05:40,980 We wanted them to be fast, we wanted them to be agile, we wanted to be dynamic, we 77 00:05:40,980 --> 00:05:45,20 wanted to be clever, but we also gave the right to fail. 78 00:05:45,20 --> 00:05:48,684 So we didn't demand storage. We didn't say, hey, hold on to packets, 79 00:05:48,684 --> 00:05:53,132 you know, all over the place. Just, store up piles of packets, piles of 80 00:05:53,132 --> 00:05:57,90 packets and pile up in these, we don't ask that. 81 00:05:57,90 --> 00:06:00,716 No router has to keep the packets. As a matter of fact, routers are supposed 82 00:06:00,716 --> 00:06:04,728 to throw packets away to communicate back and forth between the system that maybe 83 00:06:04,728 --> 00:06:10,300 things aren't working so well and, don't use me to get to California. 84 00:06:10,300 --> 00:06:14,128 On the other hand, with hundreds, maybe, you know, hundreds of thousands or 85 00:06:14,128 --> 00:06:19,230 millions of routers but billions of computers on the outside. 86 00:06:19,230 --> 00:06:23,20 So we need a way to make this reliable. So we need to have memory to store the 87 00:06:23,20 --> 00:06:26,980 packets while they're in flight so we can retransmit but we store the packets in 88 00:06:26,980 --> 00:06:32,142 the computers that are outside. And there are billions of these 89 00:06:32,142 --> 00:06:37,85 computers, there are billions of them. Every computer that you can carry around, 90 00:06:37,85 --> 00:06:41,880 every laptop. come here. 91 00:06:41,880 --> 00:06:47,48 Every time we add another computer to the network, we add storage for packets that 92 00:06:47,48 --> 00:06:51,820 are being sent. So when your computer or, or phone is 93 00:06:51,820 --> 00:06:58,460 sending across the network, it is responsible for retaining its own copies. 94 00:06:58,460 --> 00:07:01,782 It does not expect the inner part of the network to do so. 95 00:07:01,782 --> 00:07:08,986 And that is absolutely brilliant, so it's what really makes this work. 96 00:07:08,986 --> 00:07:14,600 And so that's kind of a great oversimplification of what's going on. 97 00:07:14,600 --> 00:07:21,30 And it turns out that there is still a lot of engineering to make that happen. 98 00:07:21,30 --> 00:07:25,125 So I'd like to introduce you to a, to another of the innovators in the, that we 99 00:07:25,125 --> 00:07:29,155 meet in this class his name is Van Jacobson, and I met up with him at Xerox 100 00:07:29,155 --> 00:07:33,575 Park he did the work we're going to talk about while he was at Berkley, and[COUGH] 101 00:07:33,575 --> 00:07:37,800 In the late 1980's there was this prediction that the internet was going to 102 00:07:37,800 --> 00:07:44,790 die. And it seems obvious that it was, it's a 103 00:07:44,790 --> 00:07:48,870 good idea now, but back then there was a bunch of folks that felt like academics 104 00:07:48,870 --> 00:07:54,620 weren't smart enough to make a network. And 105 00:07:54,620 --> 00:07:58,268 And they were like IBM and digital equipment that thought they, the vendors, 106 00:07:58,268 --> 00:08:03,300 should make the network, and we should just pay them to use their network. 107 00:08:03,300 --> 00:08:06,716 And as the NSF net was coming up, and more and more computers were being 108 00:08:06,716 --> 00:08:11,835 connected in the background, the backbone was so slow it started to fail. 109 00:08:11,835 --> 00:08:15,245 And it looked like the predictions of all of the computer vendors that said 110 00:08:15,245 --> 00:08:20,128 academics couldn't build a robust scaleable network were going to be true. 111 00:08:20,128 --> 00:08:24,483 And so Van Jacobson tells the story very differently but the way I saw the story 112 00:08:24,483 --> 00:08:29,450 happening and unfolding in 1987 is Van Jacobson saved us. 113 00:08:29,450 --> 00:08:33,938 the network was crashing, and we all installed Van Jacobson patches and the 114 00:08:33,938 --> 00:08:39,408 network got better. And in my recollection this was the last 115 00:08:39,408 --> 00:08:44,680 time it appeared that the entire internet was going to crash. 116 00:08:44,680 --> 00:08:48,880 So we named it the Van Jacobson protocol. And he doesn't like to call it the Van 117 00:08:48,880 --> 00:08:53,10 Jacobson protocol because he's a shy and unassuming guy. 118 00:08:53,10 --> 00:08:59,85 But I, I think he saved us. And so this interview is him describing 119 00:08:59,85 --> 00:09:02,985 sort of that moment back in the late 80's where he invented the slow start 120 00:09:02,985 --> 00:09:07,210 algorithm that really is part and parcel of every T C P implementation that you 121 00:09:07,210 --> 00:09:11,370 have on every computer that you use as a matter of fact it's being used to float 122 00:09:11,370 --> 00:09:18,680 control on this lecture right this second as we speak. 123 00:09:18,680 --> 00:09:20,663 So here's Van Jacobson.