1 00:00:45,000 --> 00:00:45,733 2 00:00:45,733 --> 00:00:47,600 There have been two major ideas... 3 00:00:47,600 --> 00:00:47,866 ... within the computing community within the last decade. 4 00:00:47,866 --> 00:00:50,600 ... within the computing community within the last decade. 5 00:00:50,766 --> 00:00:52,700 The first was time sharing, which ... 6 00:00:52,700 --> 00:00:52,966 The first was time sharing, which ... 7 00:00:52,966 --> 00:00:56,400 ... was being advocated very hard in the beginning of the decade. 8 00:00:56,400 --> 00:00:57,000 The key thing that was being observed was that ... 9 00:00:57,000 --> 00:00:58,900 The key thing that was being observed was that... 10 00:00:58,900 --> 00:01:00,000 ... there was a lot better way to interact with a computer. 11 00:01:00,000 --> 00:01:01,200 ... there was a lot better way to interact with a computer. 12 00:01:02,600 --> 00:01:03,000 The second major idea within the last decade ... 13 00:01:03,000 --> 00:01:05,000 The second major idea within the last decade ... 14 00:01:05,000 --> 00:01:06,000 ... has been the notion that ... 15 00:01:06,000 --> 00:01:06,500 ... has been the notion that ... 16 00:01:06,500 --> 00:01:08,533 ... computer networks were not only needed ... 17 00:01:08,533 --> 00:01:10,400 ... but were valuable... 18 00:01:10,400 --> 00:01:13,133 .. and they were gradually coming into production. 19 00:01:13,133 --> 00:01:13,700 20 00:01:13,700 --> 00:01:15,500 The computing technology ... 21 00:01:15,500 --> 00:01:18,066 ... has been moving in a way that nothing else ... 22 00:01:18,066 --> 00:01:19,400 ... hat people have ever known ... 23 00:01:19,400 --> 00:01:20,333 ... has moved. 24 00:01:20,333 --> 00:01:20,833 25 00:01:20,833 --> 00:01:23,900 Here is a field that gets a thousand times as good in twenty years . 26 00:01:23,900 --> 00:01:24,833 27 00:01:24,833 --> 00:01:28,100 The communication field hasn't been able to keep pace, ... 28 00:01:28,100 --> 00:01:30,866 ... but the melding of computers and communication ... 29 00:01:30,866 --> 00:01:33,800 ... and the switch to digital communication technology, ... 30 00:01:33,800 --> 00:01:35,900 ... aided and abetted by satellites ... 31 00:01:36,100 --> 00:01:39,100 ... is doing something pretty good for communication. 32 00:01:39,566 --> 00:01:42,633 The problem is much like small civilizations or ... 33 00:01:42,633 --> 00:01:44,566 ... or small cities trying to develop separately... 34 00:01:44,566 --> 00:01:45,566 ... or small cities trying to develop separately... 35 00:01:45,566 --> 00:01:48,833 ... and not having any of sharing ... 36 00:01:48,833 --> 00:01:51,033 ... what they learned with other groups. 37 00:01:51,033 --> 00:01:54,100 And if that continues to happen you don't have a civilization. 38 00:01:54,100 --> 00:01:56,466 So that the necessity was ... 39 00:01:56,466 --> 00:01:59,500 ... to provide a mechanism so that what was learned in one place ... 40 00:01:59,500 --> 00:02:01,866 ... ... could be transferred effectively and directly ... 41 00:02:01,866 --> 00:02:04,700 ... to other place, without redoing it all... 42 00:02:04,700 --> 00:02:06,966 ... and learning it all over again at each place. 43 00:02:06,966 --> 00:02:09,633 The problem was not in the computers, we found out, 44 00:02:09,633 --> 00:02:11,366 (computers could talk to one another, ... 45 00:02:11,366 --> 00:02:13,233 ... even though they were very different) 46 00:02:13,666 --> 00:02:15,400 ... and time sharing systems ... 47 00:02:15,400 --> 00:02:17,300 ... that existed within each computer ... 48 00:02:17,300 --> 00:02:19,500 ... were perfectly capable of achieving that. 49 00:02:19,500 --> 00:02:22,300 But the real problem that we found was that 50 00:02:22,300 --> 00:02:24,033 ... the communications was inadequate. 51 00:02:24,200 --> 00:02:25,733 And ideally what we wanted to do was ... 52 00:02:25,733 --> 00:02:27,733 ... was to use a common carrying authoring. 53 00:02:27,733 --> 00:02:31,333 Unfortunately, there was no wide-band, switched ... 54 00:02:31,333 --> 00:02:33,633 ... common carrying authoring at the time. 55 00:02:33,966 --> 00:02:39,000 Our problem at BBN was to build new kind of digital communication system, ... 56 00:02:39,000 --> 00:02:42,733 ... employing wide-band leased lines and message switching. 57 00:02:42,733 --> 00:02:46,333 Message switching is were each path is not established in advance, ... 58 00:02:46,333 --> 00:02:48,700 ... and each message carries an address. 59 00:02:49,266 --> 00:02:51,633 We wanted this new communication system ... 60 00:02:51,633 --> 00:02:53,666 ... to support all the various ... 61 00:02:53,666 --> 00:02:56,033 ... kinds of resource sharing which were ... 62 00:02:56,033 --> 00:02:57,166 ... so important. 63 00:02:57,166 --> 00:03:00,900 At the same time we wanted to build a data communication system... 64 00:03:00,900 --> 00:03:02,966 ... which would stand in its own right ... 65 00:03:02,966 --> 00:03:05,700 ... as a better, more economical ... 66 00:03:05,700 --> 00:03:07,300 ... higher performance ... 67 00:03:07,300 --> 00:03:10,100 ... and faster, and more reliable ... 68 00:03:10,100 --> 00:03:12,233 ... digital communication system. 69 00:03:12,433 --> 00:03:14,966 The underlying concept of that network ... 70 00:03:14,966 --> 00:03:17,900 ... will have to embody the efficient utilization of ... 71 00:03:17,900 --> 00:03:19,833 ... communication resources, ... 72 00:03:19,833 --> 00:03:22,733 ... as well as to provide a system that was both ... 73 00:03:22,733 --> 00:03:24,800 ... reliable, error free... 74 00:03:24,800 --> 00:03:27,133 ... and provide the high-band needed for ... 75 00:03:27,133 --> 00:03:29,033 ... interactive use of those resources. 76 00:03:29,033 --> 00:03:30,566 This means that ... 77 00:03:30,566 --> 00:03:34,500 ... users sitting at a terminal would be able to hit a key and see a response ... 78 00:03:34,500 --> 00:03:37,166 ... virtually instantly, almost as if that computer, ... 79 00:03:37,166 --> 00:03:40,300 wherever it were, looked like it was in the same room. 80 00:03:40,933 --> 00:03:41,900 A simple way to interconnect ... 81 00:03:41,900 --> 00:03:42,700 A simple way to interconnect ... 82 00:03:42,700 --> 00:03:44,900 ... computers into such a network... 83 00:03:44,900 --> 00:03:48,300 ... is to place wide-band leased circuits ... 84 00:03:48,300 --> 00:03:51,633 ... (in the case of the APRAnet 50kb/sec circuits) ... 85 00:03:51,633 --> 00:03:53,500 ... between each of the computers, ... 86 00:03:54,600 --> 00:03:57,366 ... and then to interconnect ... 87 00:03:57,366 --> 00:03:58,666 ... each of the computers ... 88 00:03:58,666 --> 00:04:01,466 ... to each other to form a fully connected network. 89 00:04:01,466 --> 00:04:03,600 As more sites come onto the network, ... 90 00:04:04,200 --> 00:04:07,466 ... the requirement is than to connect that site to every other one, which ... 91 00:04:07,466 --> 00:04:10,666 ... that the extension of the network is just not a graceful thing. 92 00:04:11,100 --> 00:04:14,566 And so this naturally leads to the concept of a ... 93 00:04:14,566 --> 00:04:16,900 ... store and forward technique in order to ... 94 00:04:16,900 --> 00:04:19,833 ... cut down on the expenses of building such a network. 95 00:04:21,866 --> 00:04:22,766 And so let's ... 96 00:04:23,200 --> 00:04:24,766 ... erase these lines ... 97 00:04:25,833 --> 00:04:26,833 ... over here ... 98 00:04:27,633 --> 00:04:29,800 ... to leave ourselves for the moment ... 99 00:04:30,166 --> 00:04:31,466 ... with a loop network ... 100 00:04:31,933 --> 00:04:33,033 ... which can be extended. 101 00:04:33,700 --> 00:04:35,133 And in this type of network, ... 102 00:04:35,133 --> 00:04:36,900 ... this computer for example ... 103 00:04:36,900 --> 00:04:38,466 ... would talk to this computer ... 104 00:04:38,466 --> 00:04:41,566 ... not by sending it a massage directly, since there is no circuit, ... 105 00:04:41,566 --> 00:04:44,133 ... but by sending a message first to this computer ... 106 00:04:44,133 --> 00:04:47,166 ... which would then store it and forward it toward this computer, ... 107 00:04:47,166 --> 00:04:48,933 ... thereby acting as a relay. 108 00:04:49,166 --> 00:04:50,533 In order to ... 109 00:04:51,666 --> 00:04:54,100 ... have a reliable network of this sort, ... 110 00:04:54,100 --> 00:04:57,566 each of these individual computers must be sufficiently reliable ... 111 00:04:57,566 --> 00:05:00,433 to maintain the kinds of communication that's needed. 112 00:05:00,433 --> 00:05:01,833 But unfortunately ... 113 00:05:01,833 --> 00:05:04,800 ... most computer installations are just not reliable enough... 114 00:05:04,800 --> 00:05:06,333 ... and this leads to the notion ... 115 00:05:06,333 --> 00:05:09,066 ... of a small, mini processor ... 116 00:05:09,066 --> 00:05:11,433 ... to take on the functions of the computers ... 117 00:05:11,433 --> 00:05:13,833 ... and to allow a single design ... 118 00:05:13,833 --> 00:05:16,933 ...to then be propagated among all the installations. 119 00:05:16,933 --> 00:05:18,300 So that ... 120 00:05:18,300 --> 00:05:21,966 So we would put a little mini processor at each computer, ... 121 00:05:23,400 --> 00:05:24,266 ... like that, ... 122 00:05:27,700 --> 00:05:30,766 ... disconnect the 50kb circuits ... 123 00:05:30,766 --> 00:05:32,900 ... linking the computers themselves, ... 124 00:05:36,166 --> 00:05:38,566 ... and then interconnect these little mini processors, ... 125 00:05:39,500 --> 00:05:40,366 ... or IMPs ... 126 00:05:42,100 --> 00:05:43,733 ... with the wide-band circuits, ... 127 00:05:44,100 --> 00:05:47,133 ... and then connect the computers to the IMPs, ... 128 00:05:47,133 --> 00:05:48,466 ... in this fashion. 129 00:05:50,266 --> 00:05:51,966 Now such a network would then operate ... 130 00:05:52,700 --> 00:05:55,166 ... essentially in the same fashion as the previous one ... 131 00:05:55,166 --> 00:05:57,166 ... with this computer talking to this computer ... 132 00:05:57,166 --> 00:05:59,533 ... by first sending a message to its IMP ... 133 00:05:59,533 --> 00:06:00,766 ... having this IMP . 134 00:06:00,766 --> 00:06:02,166 ... relaying it to this IMP... 135 00:06:02,166 --> 00:06:03,966 ... and this IMP relaying it to this IMP, ... 136 00:06:03,966 --> 00:06:05,500 ... and then this IMP delivering it ... 137 00:06:05,500 --> 00:06:07,366 ... to the final destination. 138 00:06:07,666 --> 00:06:11,433 Now this kind of network can be made to be extremely reliable ... 139 00:06:11,433 --> 00:06:15,700 since an effective control can be placed upon the design of each of these IMPs ... 140 00:06:15,700 --> 00:06:18,766 ... since there is no large political problem in ... 141 00:06:18,766 --> 00:06:20,766 ... getting a large number of sites ... 142 00:06:20,766 --> 00:06:24,333 ... to cooperate in the design and building of the communications part of the system. 143 00:06:24,333 --> 00:06:28,133 We operate a network control centre here at BBN in Cambridge... 144 00:06:28,133 --> 00:06:31,833 ... and each IMP every half second sends us a message ... 145 00:06:31,833 --> 00:06:33,466 ... telling us how it feels and ... 146 00:06:33,466 --> 00:06:36,800 and how each of its lines are, and how each of its hosts are, ... 147 00:06:36,800 --> 00:06:38,900 ... and what kind of loading it's got. 148 00:06:39,233 --> 00:06:42,166 And we man that centre 24 hours a day ... 149 00:06:42,166 --> 00:06:45,100 ... and we use those stats-reports ... 150 00:06:45,100 --> 00:06:46,533 ... from the IMPs ... 151 00:06:46,533 --> 00:06:49,966 ... to generate statistics about network performance ... 152 00:06:49,966 --> 00:06:52,033 ... and also to alert the operators ... 153 00:06:52,033 --> 00:06:55,666 ... to any immediate needs for maintenance, either in circuits of in IMPs. 154 00:06:55,866 --> 00:06:58,166 Here is an instance of the ARPAnet ... 155 00:06:58,166 --> 00:07:00,133 ... as it was recently configured ... 156 00:07:00,133 --> 00:07:01,500 ... so you can see there are some ... 157 00:07:01,500 --> 00:07:03,566 ... 25, 30 sites in it. 158 00:07:04,566 --> 00:07:08,800 The transmission of a message, say from a node over here, ... 159 00:07:08,800 --> 00:07:11,266 ... to a node over here might go as follows: 160 00:07:11,266 --> 00:07:11,700 The computer at this point would send a message to its local IMP ... 161 00:07:11,700 --> 00:07:15,600 The computer at this point would send a message to its local IMP ... 162 00:07:15,600 --> 00:07:18,200 which would break it down into thousand bits packets. 163 00:07:18,800 --> 00:07:21,500 The packets would then be transmitted from IMP to IMP ... 164 00:07:21,500 --> 00:07:24,166 ... along a route selected by the IMP themselves. 165 00:07:24,800 --> 00:07:26,100 At the destination IMP, ... 166 00:07:26,466 --> 00:07:29,133 the package would be reassembled in the proper order ... 167 00:07:29,133 --> 00:07:30,466 ... and delivered to the computer. 168 00:07:30,966 --> 00:07:32,533 And then a message would go back ... 169 00:07:32,533 --> 00:07:34,300 ... along perhaps a different route, ... 170 00:07:34,300 --> 00:07:37,166 ... to indicate that the original message was received. 171 00:07:37,466 --> 00:07:40,566 The whole transmission cycle typically takes no more than ... 172 00:07:40,566 --> 00:07:41,500 ... a few tenth of a second. 173 00:07:42,500 --> 00:07:46,333 The system is completely independent of the ups and downs of ... 174 00:07:46,333 --> 00:07:48,300 ... small numbers of lines; for example ... 175 00:07:48,300 --> 00:07:50,400 ... if this circuit over here broke ... 176 00:07:50,400 --> 00:07:51,900 ... in the midst of the transmission, ... 177 00:07:51,900 --> 00:07:53,733 ... the message has gotten that far ... 178 00:07:53,733 --> 00:07:55,800 ... it might then backtrack to back here and ... 179 00:07:55,800 --> 00:07:58,733 ... possibly takes some other route until it gets to the destination. 180 00:07:58,733 --> 00:08:00,066 There is error-checking ... 181 00:08:00,066 --> 00:08:01,700 ... in between each IMPs, so ... 182 00:08:01,700 --> 00:08:05,166 ... that in the transmission from this node to this node ... 183 00:08:06,133 --> 00:08:08,133 ... the message that was send would actually be ... 184 00:08:08,133 --> 00:08:10,900 ... error-checked and an acknowledgment would be send back ... 185 00:08:10,900 --> 00:08:13,000 ... and than from here to here ... 186 00:08:13,000 --> 00:08:15,000 it would be error-checked and if it got ... 187 00:08:15,000 --> 00:08:16,366 ... correctly accepted ... 188 00:08:16,366 --> 00:08:18,766 ... an acknowledgement would be go back an so ... 189 00:08:18,766 --> 00:08:20,933 ... forth among each set of IMPs along the path. 190 00:08:21,566 --> 00:08:25,733 A very unusual feature of this system which I thing is quite new ... 191 00:08:25,733 --> 00:08:28,366 ... is that it's possible for us to debug ... 192 00:08:28,366 --> 00:08:30,666 ... any of the running IMPs in the field. 193 00:08:30,666 --> 00:08:34,200 We can actually examine core, test and restart them ... 194 00:08:34,200 --> 00:08:37,933 ... and even reload the program from our Cambridge location. 195 00:08:38,933 --> 00:08:42,000 In order to add a new resource into the network ... 196 00:08:44,900 --> 00:08:45,800 ... such as this ... 197 00:08:46,333 --> 00:08:47,500 ... it would get its own IMP ... 198 00:08:48,000 --> 00:08:50,666 ... it would get its own IMP and would require one or two or ... 199 00:08:50,666 --> 00:08:53,300 ... perhaps three connections to other IMPs in the network. 200 00:08:54,100 --> 00:08:57,200 That would be all that would be required the effect that connection. 201 00:08:57,200 --> 00:08:59,233 So we could in principle let's say ... 202 00:08:59,233 --> 00:09:01,433 ... get rid of that connection and the network would work, ... 203 00:09:01,433 --> 00:09:03,633 ... or we could put in that connection over there ... 204 00:09:03,633 --> 00:09:05,333 ... to add a little more reliability. 205 00:09:05,833 --> 00:09:08,066 As soon as the network began working ... 206 00:09:08,100 --> 00:09:11,266 ... we really had a nation-wide resource complex. 207 00:09:12,066 --> 00:09:16,133 And this resource complex was very attractive to many users ... 208 00:09:16,133 --> 00:09:19,133 ... who had no resource of their own to really contribute. 209 00:09:19,133 --> 00:09:19,733 210 00:09:19,733 --> 00:09:23,533 These people wanted direct terminal access to the net, ... 211 00:09:23,533 --> 00:09:27,100 ... even though they had no host to get that access through. 212 00:09:27,800 --> 00:09:29,833 So we designed a new kind of IMP, ... 213 00:09:29,833 --> 00:09:33,233 ... we called it a Terminal Interface Message Processor, ... 214 00:09:33,233 --> 00:09:35,333 ... we called it a TIP, for short. 215 00:09:35,833 --> 00:09:40,033 And this machine really includes a very tiny mini-host. 216 00:09:40,533 --> 00:09:43,800 With this TIP, many different kinds of terminals ... 217 00:09:43,800 --> 00:09:46,433 ... can be directly connected to the TIP... 218 00:09:46,433 --> 00:09:50,200 ... or can be dialed in through low-speed lines to the TIP, ... 219 00:09:50,200 --> 00:09:54,500 ... and can provide acces to the nation wide resource pool ... 220 00:09:54,500 --> 00:09:57,033 ... to users at various kinds of terminals. 221 00:09:57,033 --> 00:09:59,633 Let me illustrate such a device down here ... 222 00:09:59,633 --> 00:10:02,366 ... which can than be connected into the network ... 223 00:10:02,366 --> 00:10:04,066 ... (nothing more than an IMP), ... 224 00:10:04,066 --> 00:10:06,666 ... and than terminal devices, some large number of them, ... 225 00:10:06,666 --> 00:10:09,566 ... can directly connect to that device. 226 00:10:10,333 --> 00:10:13,866 The most immediate cost-benefit of the network ... 227 00:10:13,866 --> 00:10:18,966 ... for the users: the hardware and resource sharing being able to... 228 00:10:19,666 --> 00:10:22,266 ... getting at large computers ... 229 00:10:23,033 --> 00:10:26,300 ... time sharing computers and other specialized computer capabilities... 230 00:10:26,300 --> 00:10:27,833 ... throughout the country that ... 231 00:10:27,833 --> 00:10:29,166 ... meets his needs. 232 00:10:29,166 --> 00:10:31,566 This eliminates the needs for him to have a ... 233 00:10:31,566 --> 00:10:35,000 ... medium scaled machine that does everything in a mediocre way. 234 00:10:35,766 --> 00:10:38,100 Specialized hardware facilities ... 235 00:10:38,100 --> 00:10:41,000 ... tend to be expensive, but very efficient. 236 00:10:42,266 --> 00:10:46,266 If there isn't any way to distribute their use, ... 237 00:10:46,266 --> 00:10:50,233 ... to make it available all over the country or all over the world, ... 238 00:10:50,233 --> 00:10:53,000 ... it may be economically ... 239 00:10:53,000 --> 00:10:55,300 ... impractical to provide them; 240 00:10:55,300 --> 00:10:59,566 ... because there isn't a large enough need for them in any one place. 241 00:11:00,166 --> 00:11:02,300 On the other hand, if they can be distributed, ... 242 00:11:02,300 --> 00:11:05,300 .... than specialized hardware facilities ... 243 00:11:05,300 --> 00:11:07,166 ... can be very effective... 244 00:11:07,166 --> 00:11:09,766 ... and can do a lot of things that we couldn't otherwise do. 245 00:11:09,766 --> 00:11:13,233 In the case of the large superfiles, ... 246 00:11:13,233 --> 00:11:17,833 ... the ten to the eleventh bit weather files which we're putting on the ELLIAC for example, ... 247 00:11:17,833 --> 00:11:23,100 ... these would never be available without being on the network; 248 00:11:23,100 --> 00:11:26,500 ... because it wouldn't be worthwhile for one person to have it all himself. 249 00:11:26,500 --> 00:11:30,200 The thing that makes the computer network communication network special ... 250 00:11:30,200 --> 00:11:32,400 ... is that it puts the workers, ... 251 00:11:32,400 --> 00:11:35,300 ... the team members that are geographically distributed, ... 252 00:11:35,300 --> 00:11:37,533 ... in touch not only with one another, ... 253 00:11:37,533 --> 00:11:40,966 ... but with the information base with which they work all the time. 254 00:11:40,966 --> 00:11:43,566 So that when they get to ... 255 00:11:43,566 --> 00:11:45,300 ... developing plans, ... 256 00:11:45,300 --> 00:11:48,133 ... the blue-prints as it were, ... 257 00:11:48,133 --> 00:11:51,666 ... don't have to be copied and send all around the country. 258 00:11:51,666 --> 00:11:53,766 The blue-prints come out of the database ... 259 00:11:53,766 --> 00:11:55,700 ... and appear on everybody's scopes ... 260 00:11:56,233 --> 00:11:59,033 ... and the correlation of coordination of the activity ... 261 00:11:59,100 --> 00:12:02,500 ... is essentially right there in the computer network itself. 262 00:12:02,733 --> 00:12:05,333 And this is obviously going to make a tremendous difference ... 263 00:12:05,333 --> 00:12:08,500 ... in how we plan, organize, and execute ... 264 00:12:08,500 --> 00:12:11,266 ... almost everything of any intellectual consequence. 265 00:12:11,266 --> 00:12:15,700 The main thing about this data-file sharing is that ... 266 00:12:15,700 --> 00:12:19,533 ... you don't want to update and maintain it at different places simultaneously. 267 00:12:19,533 --> 00:12:24,533 This becomes even very error prone if not very costly. 268 00:12:25,266 --> 00:12:28,033 Well it must be amusing to anyone outside the computer field ... 269 00:12:28,033 --> 00:12:29,600 ... to watch people inside it ... 270 00:12:29,600 --> 00:12:33,866 ... rewriting and recreating the same old programs over and over and over. 271 00:12:33,866 --> 00:12:38,566 And programs are recreated partly because it's fun to program ... 272 00:12:38,566 --> 00:12:40,800 ... but mainly because it's very difficult ... 273 00:12:40,800 --> 00:12:42,833 ... to get hold of them and use them ... 274 00:12:42,833 --> 00:12:46,833 ... if they have been programmed a few hundred or a few thousand miles away ... 275 00:12:46,833 --> 00:12:48,900 ... for use on different equipment ... 276 00:12:48,900 --> 00:12:50,100 ... from that which you have. 277 00:12:50,266 --> 00:12:53,100 Now, the network idea makes it ... 278 00:12:53,100 --> 00:12:56,100 ... practical to use those programs ... 279 00:12:56,166 --> 00:12:59,100 ... on the machines for which they were written and on which ... 280 00:12:59,100 --> 00:13:01,366 ... they have been debugged and tested and checked out. 281 00:13:02,300 --> 00:13:04,200 The effect of the network for a user ... 282 00:13:04,200 --> 00:13:06,500 ... is that now he can ... 283 00:13:06,500 --> 00:13:09,500 ... start to attack a new problem that he may come up with ... 284 00:13:09,500 --> 00:13:10,100 285 00:13:10,100 --> 00:13:13,800 ... with any resource that may be available within the net ... 286 00:13:13,800 --> 00:13:16,800 ... rather than the very limited set of resources that is available ... 287 00:13:16,800 --> 00:13:18,800 ... through his local computation centre. 288 00:13:19,433 --> 00:13:22,300 This has a tremendously valuable effect ... 289 00:13:22,300 --> 00:13:24,400 ... because as he starts to look at a problem ... 290 00:13:24,400 --> 00:13:26,800 ... he can use any language he can use any resource; 291 00:13:26,800 --> 00:13:30,100 ... the reliability is hard because he can choose between several machines ... 292 00:13:30,100 --> 00:13:32,300 ... for a particular run. 293 00:13:33,133 --> 00:13:36,500 And his personal situation is ... 294 00:13:36,500 --> 00:13:38,900 ... greatly magnifying what he can do. 295 00:13:39,466 --> 00:13:42,100 Because he doesn't have to depend on a very limited ... 296 00:13:42,100 --> 00:13:44,000 ...set of resources to do the job. 297 00:13:44,000 --> 00:13:47,500 The ARPA-network provides a new research opportunity ... 298 00:13:47,500 --> 00:13:51,900 ... for experimenting with computer to computer communication. 299 00:13:52,900 --> 00:13:54,500 To make computers talk to each other ... 300 00:13:54,500 --> 00:13:57,100 ... a number of functions normally performed ... 301 00:13:57,100 --> 00:13:58,400 ... by human users ... 302 00:13:58,400 --> 00:14:00,800 ... must be moved into computer programs. 303 00:14:01,066 --> 00:14:03,800 To gain some hands-on, practical experience ... 304 00:14:03,800 --> 00:14:05,700 ... with computer talking to each other ... 305 00:14:05,700 --> 00:14:08,500 ... we extended an air-traffic control simulator ... 306 00:14:08,500 --> 00:14:10,200 ... which we had previously developed. 307 00:14:10,633 --> 00:14:13,200 Using this as a model, we partitioned ... 308 00:14:13,200 --> 00:14:14,800 ... the simulation ... 309 00:14:14,800 --> 00:14:18,000 ... into a number of air-spaces. 310 00:14:18,000 --> 00:14:20,700 Using the ARPA-net as a communication medium, ... 311 00:14:20,700 --> 00:14:22,400 ... we were able then to couple ... 312 00:14:22,400 --> 00:14:25,300 ... a number of the independent simulations together ... 313 00:14:25,300 --> 00:14:27,000 ... making them talk to each other ... 314 00:14:27,000 --> 00:14:28,600 ... for purposes of ... 315 00:14:28,600 --> 00:14:31,200 ... handling air crafts off and for ... 316 00:14:31,200 --> 00:14:33,500 ... changing flight plans and the like. 317 00:14:34,166 --> 00:14:36,500 We were able to dynamically reconfigure ... 318 00:14:36,500 --> 00:14:38,800 ... one of these multi-computer simulations ... 319 00:14:38,800 --> 00:14:42,800 ... by ordering the simulation program for one particular area ... 320 00:14:42,800 --> 00:14:45,000 ... to move into another computer ... 321 00:14:45,000 --> 00:14:47,600 ... in the middle of a simulation run. 322 00:14:47,600 --> 00:14:49,900 The net effect was an ability ... 323 00:14:49,900 --> 00:14:53,200 ... to conduct a large simulation exercise ... 324 00:14:53,200 --> 00:14:55,700 ... involving several autonomous programs ... 325 00:14:55,700 --> 00:14:57,800 ... in several independent computer. 326 00:14:57,800 --> 00:15:01,000 We believe that this is an interesting example of ... 327 00:15:01,000 --> 00:15:04,600 ... future network capabilities in failure tolerance ... 328 00:15:04,600 --> 00:15:07,100 ... and in automatic load leveling. 329 00:15:07,100 --> 00:15:08,300 One of the ... 330 00:15:08,300 --> 00:15:12,200 ... things that is coming out as people begin using networks more and more ... 331 00:15:12,200 --> 00:15:15,500 ... is that it must have as the node of the network .... 332 00:15:15,500 --> 00:15:17,400 ... computers which are ... 333 00:15:17,400 --> 00:15:20,600 ... of a mulitplex class or essentially ... 334 00:15:20,600 --> 00:15:23,200 ... a major computing utility ... 335 00:15:23,200 --> 00:15:24,300 ... which is capable ... 336 00:15:25,500 --> 00:15:27,800 ... of manipulating information, because ... 337 00:15:27,800 --> 00:15:29,900 ... the network itself doesn't ... 338 00:15:29,900 --> 00:15:32,600 ... hold information; it doesn't keep information. 339 00:15:32,600 --> 00:15:35,500 – an important design criterium in fact. 340 00:15:35,500 --> 00:15:37,100 Research now going on ... 341 00:15:37,100 --> 00:15:39,900 ... will someday permit a network user to log in ... 342 00:15:39,900 --> 00:15:42,700 ... and not really care where his computation takes place. 343 00:15:43,533 --> 00:15:45,400 With a distributed operating system ... 344 00:15:45,400 --> 00:15:48,100 ... the user will log in into a network of computers, ... 345 00:15:48,100 --> 00:15:51,500 ... the system deciding which computer can best perform his job. 346 00:15:52,200 --> 00:15:54,700 This development will make possible much more efficient ... 347 00:15:54,700 --> 00:15:56,500 ... and reliable use of computers. 348 00:15:56,800 --> 00:16:00,500 Now the interesting thing is that as time goes on we find that ... 349 00:16:01,000 --> 00:16:02,700 ... the powerful nodes ... 350 00:16:02,700 --> 00:16:04,300 ... also need the network; 351 00:16:05,400 --> 00:16:08,900 ... because as one begins to take on larger and larger problems, ... 352 00:16:09,900 --> 00:16:14,200 ... one of these systems that one becomes more and more dependent on, ... 353 00:16:14,700 --> 00:16:15,800 ... which one must ... 354 00:16:16,500 --> 00:16:19,000 ... be certain of being up and running. 355 00:16:19,500 --> 00:16:20,900 And in order to do that, ... 356 00:16:20,900 --> 00:16:25,200 ... one has to have tremendous reliability agains all kinds of catastrophes. 357 00:16:25,200 --> 00:16:27,800 The user of distributed operating systems ... 358 00:16:27,800 --> 00:16:30,200 ... will know a new kind of reliability. 359 00:16:30,200 --> 00:16:32,800 For example, his files could be backed up ... 360 00:16:32,800 --> 00:16:34,300 ... on more than one site, ... 361 00:16:34,300 --> 00:16:36,500 ... so that if one of the sites would go down, ... 362 00:16:36,500 --> 00:16:39,500 ... he could obtain his files from another site in the network. 363 00:16:41,800 --> 00:16:43,400 I'm involved in the ... 364 00:16:44,000 --> 00:16:47,800 ... ARPA program to develop a speech understanding system. 365 00:16:48,200 --> 00:16:49,600 And I think that that program ... 366 00:16:50,300 --> 00:16:51,800 ... illustrates some ... 367 00:16:51,800 --> 00:16:53,700 ... potentials of the network. 368 00:16:55,100 --> 00:16:57,400 This is a program that involves ... 369 00:16:58,100 --> 00:17:00,400 ... at present time eight laboratories, ... 370 00:17:01,000 --> 00:17:04,600 ... some labs will develop some sub- systems and others other sub-systems. 371 00:17:05,100 --> 00:17:06,600 Maybe two or three will develop ... 372 00:17:06,600 --> 00:17:09,000 ... entire speech understanding systems, ... 373 00:17:09,000 --> 00:17:11,100 ... but even then it's unlikely that ... 374 00:17:11,100 --> 00:17:13,300 ... one of them will have the best of everything. 375 00:17:14,000 --> 00:17:16,400 There's also a need for a database ... 376 00:17:16,400 --> 00:17:19,700 ... of controlled and well- understood speech samples ... 377 00:17:19,700 --> 00:17:22,400 ... for use in development and in testing ... 378 00:17:22,400 --> 00:17:25,000 ... of the various sub-systems and overall-systems. 379 00:17:25,000 --> 00:17:27,800 And finally there's the requirement that ... 380 00:17:28,800 --> 00:17:30,800 ... the people involved in all these projects ... 381 00:17:30,800 --> 00:17:32,500 ... communicate with each other. 382 00:17:32,700 --> 00:17:37,000 The network information centre which is located in Standford Research Institute in California ... 383 00:17:37,800 --> 00:17:40,600 ... sees the network as this multi-layered experiment ... 384 00:17:40,600 --> 00:17:43,700 ... in resource sharing where the resource is available for people, ... 385 00:17:43,700 --> 00:17:45,600 ... computers, data. 386 00:17:45,600 --> 00:17:47,900 At one level you've got the technology that's ... 387 00:17:48,600 --> 00:17:51,800 ... informating the network; the circuits that send the bits from place to place. 388 00:17:51,800 --> 00:17:54,700 At another legel you've got the protocols and procedures ... 389 00:17:54,700 --> 00:17:56,300 ... that allow data to be shared, ... 390 00:17:56,300 --> 00:17:56,700 391 00:17:56,700 --> 00:17:58,500 ... computers to talk to each other and so forth. 392 00:17:59,100 --> 00:18:00,600 And at other levels you've got ... 393 00:18:00,600 --> 00:18:03,000 ... the facilities to allow our peoples to get together ... 394 00:18:03,000 --> 00:18:05,900 ... who are geographically dispersed and allow them to work together, ... 395 00:18:05,900 --> 00:18:08,200 ... find out information, resources and so forth ... 396 00:18:08,200 --> 00:18:10,600 ... that they need to do their job and bring them together. 397 00:18:11,300 --> 00:18:13,400 So we see our job is twofold: 398 00:18:13,400 --> 00:18:17,200 one, to provide information about resources, about people, ... 399 00:18:17,200 --> 00:18:19,500 ... about the network, that people need; 400 00:18:19,500 --> 00:18:22,500 ... to bring these things together to do their job. 401 00:18:22,500 --> 00:18:25,900 And once they've gotten together these things particular groups of people, ... 402 00:18:25,900 --> 00:18:30,200 ... we like to provide services that will aide their collaboration and their working together. 403 00:18:31,000 --> 00:18:34,500 The succes of the network will depend by and large ... 404 00:18:34,500 --> 00:18:35,200 405 00:18:35,200 --> 00:18:36,700 how the user views them. 406 00:18:36,700 --> 00:18:38,700 You have to give the user ... 407 00:18:38,700 --> 00:18:41,800 ... something more than he gets at present time, ... 408 00:18:41,800 --> 00:18:45,200 ... otherwise he's not going to be particularly interested in the network. 409 00:18:45,200 --> 00:18:48,200 And there are lots of political problems associated with that. 410 00:18:48,200 --> 00:18:50,400 He presently has a computer center. 411 00:18:50,400 --> 00:18:52,700 That computer center by and large ... 412 00:18:52,700 --> 00:18:55,700 provides some kind of service for him. 413 00:18:55,700 --> 00:18:58,500 The people who run the computer center ... 414 00:18:58,500 --> 00:19:02,500 ... are going to continue to want to operate in that mode. 415 00:19:02,500 --> 00:19:05,700 As far as the economic aspects are concerned, ... 416 00:19:05,700 --> 00:19:08,400 ... various computer centers would presumably ... 417 00:19:08,400 --> 00:19:12,400 ... become expert in some area and provide some kind of resource ... 418 00:19:12,400 --> 00:19:15,400 ... that would be useful for all the others and they could concentrate on it; 419 00:19:15,400 --> 00:19:18,000 ... they wouldn't have to cut across the whole board ... 420 00:19:18,000 --> 00:19:21,200 ... and spread out their efforts among many many ... 421 00:19:21,200 --> 00:19:22,600 ... different kinds of programs. 422 00:19:22,600 --> 00:19:26,800 The ARPA network has been criticized as not being typical what's required ... 423 00:19:26,800 --> 00:19:28,700 ... for a real commercial network. 424 00:19:29,700 --> 00:19:33,400 Well let's look at what present day, commercial networks do. 425 00:19:33,400 --> 00:19:36,800 First of all they are only build by very big companies that can afford ... 426 00:19:36,800 --> 00:19:40,300 ... the expense of the design and have leased long lines and so forth. 427 00:19:40,300 --> 00:19:43,400 And secondly they are build around special applications. 428 00:19:43,400 --> 00:19:47,100 This schema of private, purpose-build networks ... 429 00:19:47,100 --> 00:19:49,600 ... is, in my view, a completely false idea ... 430 00:19:49,600 --> 00:19:52,400 ... of what the future of data communication will be. 431 00:19:53,100 --> 00:19:55,700 It seems to me that data communication for a ... 432 00:19:55,700 --> 00:19:58,700 company must evolve in the same way as the company evolves. 433 00:19:58,700 --> 00:20:01,700 Organizations change, they merge, ... 434 00:20:01,700 --> 00:20:04,100 ... they introduce new services; 435 00:20:04,100 --> 00:20:07,000 so in principle every terminal on the network, ... 436 00:20:07,000 --> 00:20:09,800 ... every computer in the company, 437 00:20:09,800 --> 00:20:11,600 ... and computers of different companies, 438 00:20:11,600 --> 00:20:14,200 all have to be able interconnect. 439 00:20:14,200 --> 00:20:16,400 This means that the subscribers of the network ... 440 00:20:16,400 --> 00:20:19,000 ... are a great variety of different terminals ... 441 00:20:19,000 --> 00:20:20,900 ... quite unlike the telephone network. 442 00:20:20,900 --> 00:20:23,700 And to my mind, this implies ... 443 00:20:23,700 --> 00:20:25,100 ... the package-switching principle; 444 00:20:25,100 --> 00:20:27,700 it requires that the terminals ... 445 00:20:27,700 --> 00:20:31,600 ... interconnect at the level of messages and not at particular speed. 446 00:20:32,400 --> 00:20:35,500 Well clearly the public switched network is needed ... 447 00:20:35,500 --> 00:20:38,600 ... and I don't think many people would now disagree with this. 448 00:20:38,600 --> 00:20:41,700 And it must be very versatile and able to connect ... 449 00:20:41,700 --> 00:20:44,400 ... terminals of very widely different types. 450 00:20:44,400 --> 00:20:48,300 In my view, the computer and communication people haven't yet really begun ... 451 00:20:48,300 --> 00:20:51,500 ... a proper dialog; they're not yet speaking the same language. 452 00:20:52,100 --> 00:20:55,700 The thoughts of the communication people are still rooted ... 453 00:20:55,700 --> 00:20:57,800 ...in the technology of the telephone network. 454 00:20:57,800 --> 00:21:00,600 And I think it's necessary, very necessary, ... 455 00:21:00,600 --> 00:21:03,500 ... for computer people in the future ... 456 00:21:03,500 --> 00:21:06,600 ... to learn more about the problems of these networks, ... 457 00:21:06,600 --> 00:21:10,100 ... so they can really join in this dialogue and in this way we can get the ... 458 00:21:10,100 --> 00:21:12,000 ... kind of data network that we need. 459 00:21:12,000 --> 00:21:17,200 Well, it's been hard to share information for years. 460 00:21:17,200 --> 00:21:20,200 The printing press of course, was the ... 461 00:21:20,200 --> 00:21:22,700 ... great step into sharing ... 462 00:21:22,700 --> 00:21:26,800 ... information, but the printing press didn't essentially 463 00:21:26,800 --> 00:21:29,100 ... handle the problem of distributing it. 464 00:21:29,100 --> 00:21:31,300 It handled the problem of copying it. 465 00:21:31,300 --> 00:21:33,600 And we have been needing for a long time ... 466 00:21:33,600 --> 00:21:36,000 ... some better way to distribute information ... 467 00:21:36,000 --> 00:21:37,700 ... than to carry it about. 468 00:21:37,700 --> 00:21:39,600 The print on paper form ... 469 00:21:39,600 --> 00:21:42,900 ... is embarrassing because in order to distribute it ... 470 00:21:42,900 --> 00:21:44,500 ... you've to move the paper around ... 471 00:21:44,500 --> 00:21:47,200 ... and lots of paper get to be bulky and heavy ... 472 00:21:47,200 --> 00:21:49,000 ... and expensive to move about. 473 00:21:49,700 --> 00:21:53,900 There are many million checking accounts in the United States. 474 00:21:53,900 --> 00:21:57,300 These are in some thirteen thousand banks. 475 00:21:57,300 --> 00:22:00,400 And on average there's a cheque written on each ... 476 00:22:00,400 --> 00:22:02,500 ... one of these accounts every business day. 477 00:22:02,500 --> 00:22:05,500 About a hundred million cheques every business day. 478 00:22:05,500 --> 00:22:08,400 And that's a hundred million pieces of paper. 479 00:22:09,000 --> 00:22:12,300 If we get into a mode in which ... 480 00:22:12,300 --> 00:22:15,100 ... everything is handled electronically ... 481 00:22:15,100 --> 00:22:19,200 ... and your only identification is some plastic thing you stick into the machinery, ... 482 00:22:19,200 --> 00:22:21,100 ... then I can imagine that they wanna ... 483 00:22:21,100 --> 00:22:24,100 ... get that settled up with your bank account just right now ... 484 00:22:24,100 --> 00:22:26,200 ... and put it through all the cheques. 485 00:22:26,200 --> 00:22:27,900 And that would require a network. 486 00:22:28,800 --> 00:22:31,400 We have our own network ... 487 00:22:31,400 --> 00:22:36,200 ... which links the federal reserves offices, all 37 of them. 488 00:22:36,200 --> 00:22:39,500 And they in turn are linked with commercial banks and their communities. 489 00:22:39,500 --> 00:22:41,400 That linkages is not yet complete. 490 00:22:42,400 --> 00:22:45,500 The kind of communication required is ... 491 00:22:45,500 --> 00:22:48,700 ... exactly that provided by a computer communications network. 492 00:22:48,700 --> 00:22:51,700 A kind of communication in which you can get in ... 493 00:22:51,700 --> 00:22:55,900 ... for a tenth of a second or for a hundredth of a second or a thousandth of a second, ... 494 00:22:55,900 --> 00:22:58,800 ... do your job and be over with it ... 495 00:22:58,800 --> 00:23:00,800 ... and let somebody else use the facilities. 496 00:23:01,600 --> 00:23:05,500 There is some resistance to changes or people when it happens. 497 00:23:05,500 --> 00:23:09,000 Many people like to get currency, they like the feel of it. 498 00:23:09,600 --> 00:23:12,600 Other people don't like to ... 499 00:23:12,600 --> 00:23:16,200 ... have this mechanized and made a matter of record on a bank account ... 500 00:23:16,200 --> 00:23:19,500 ... because husbands don't want a wife to know what his income is. 501 00:23:20,700 --> 00:23:25,600 There isn't any real need to change things just for the sake of changing. 502 00:23:26,600 --> 00:23:29,600 But I tend to believe that things are gonna be ... 503 00:23:29,600 --> 00:23:32,200 ... considerably better for a lot of people ... 504 00:23:32,200 --> 00:23:36,900 ...when and if we ever get changed over to an essentially electronic base. 505 00:23:36,900 --> 00:23:39,200 It's just fundamental that ... 506 00:23:39,200 --> 00:23:41,500 ... if one wants to deal with information ... 507 00:23:41,500 --> 00:23:45,000 ... you ought to deal with information and not with the paper it's written on. 508 00:23:45,000 --> 00:23:47,300 The network now costs us ... 509 00:23:47,300 --> 00:23:51,400 ... one tenth the cost of mail, for moving paper; 510 00:23:51,400 --> 00:23:55,600 ... and this this cost will continue to go down as labour costs go up. 511 00:23:56,600 --> 00:23:58,300 So that it's quite clear that ... 512 00:23:58,300 --> 00:24:00,600 ... material will be moved ... 513 00:24:00,600 --> 00:24:02,600 ... and handled and stored in ... 514 00:24:02,600 --> 00:24:05,600 ... computer systems rather than in filing cabinets. 515 00:24:05,600 --> 00:24:09,200 Right now, it's possible to buy for about a million dollars ... 516 00:24:09,200 --> 00:24:13,100 ... an information store that will hold the equivalent ... 517 00:24:13,100 --> 00:24:15,200 ... of about a hundred thousand books. 518 00:24:15,200 --> 00:24:19,200 So one can store – or one can buy the store for a book – 519 00:24:19,200 --> 00:24:21,500 ... for about the same amount as he can buy the book. 520 00:24:22,000 --> 00:24:23,000 So that, ... 521 00:24:23,000 --> 00:24:26,000 ... if everyone had a display console ... 522 00:24:26,000 --> 00:24:27,700 ... in his home and in his office, ... 523 00:24:28,900 --> 00:24:31,200 ... he could be reading from ... 524 00:24:31,200 --> 00:24:34,400 ... electronically stored information instead of from a book. 525 00:24:34,400 --> 00:24:37,000 And the difference is, he could have access ... 526 00:24:37,000 --> 00:24:40,100 ... to anything he wanted to read instead of ... 527 00:24:40,100 --> 00:24:42,100 ... just what was within reach. 528 00:24:42,100 --> 00:24:45,000 Well, it turns out to be surprisingly inexpensive, ... 529 00:24:45,000 --> 00:24:48,000 ... if you get wide band-width transmission facilities, ... 530 00:24:48,000 --> 00:24:49,600 ... to send the stuff ... 531 00:24:51,100 --> 00:24:52,700 ... right when it has to be read ... 532 00:24:52,700 --> 00:24:56,700 ... instead of sending it to a local bookstore or local library ... 533 00:24:56,700 --> 00:24:58,100 ... in the hope that it might be read. 534 00:24:58,100 --> 00:25:01,000 The network experiment to be successful ... 535 00:25:01,000 --> 00:25:03,300 ... has got to include more than just the technology ... 536 00:25:03,300 --> 00:25:05,700 ... of getting computer A to talk to computer B. 537 00:25:05,700 --> 00:25:08,600 It's got to include the human institutions that will ... 538 00:25:08,600 --> 00:25:12,700 ... bring together these resources for people to solve real problems of real people. 539 00:25:12,700 --> 00:25:16,200 The processing and distribution technology ... 540 00:25:16,200 --> 00:25:17,900 ... and the storage technology ... 541 00:25:17,900 --> 00:25:20,600 ... are gonna make it possible to get over onto ... 542 00:25:20,600 --> 00:25:23,600 ... a new technological base for intellectual efforts ... 543 00:25:23,600 --> 00:25:24,600 ... before ... 544 00:25:26,100 --> 00:25:29,100 ... our ponderous social processes will let us. 545 00:25:29,100 --> 00:25:33,400 And I think more people ought to get in there and think about the social processes.