1
00:00:00,459 --> 00:00:20,381
[MUSIC][NOISE] 

2
00:00:20,381 --> 00:00:25,551
. 
I was extraordinarily lucky and happened 

3
00:00:25,551 --> 00:00:29,775
to be at the Xerox Research, Xerox Palo 
Alto Research Center when a, a problem 

4
00:00:29,775 --> 00:00:35,877
evolved that had never before occurred, 
the problem had never occurred. 

5
00:00:35,877 --> 00:00:39,570
And that was the problem of having a 
building full of personal computers. 

6
00:00:40,810 --> 00:00:44,650
And I was the networking guy, so they 
turned to me and said, network these 

7
00:00:44,650 --> 00:00:48,393
puppies. 
And we bad just finished starting the 

8
00:00:48,393 --> 00:00:51,570
internet. 
It was the called the arpanet, which was 

9
00:00:51,570 --> 00:00:55,629
packet switching. 
And it was pretty clear we wanted this 

10
00:00:55,629 --> 00:01:00,150
network to connect to the, it wasn't yet 
called the internet, thing. 

11
00:01:00,150 --> 00:01:03,660
So, it would be packet switch, we were 
pretty sure. 

12
00:01:03,660 --> 00:01:07,820
At the same time, we were building, 
arguably, I don't want to get into that 

13
00:01:07,820 --> 00:01:11,502
argument. 
The first laser printer, which in our 

14
00:01:11,502 --> 00:01:15,670
case, the first one, whose name was Ears. 
That's a whole other story. 

15
00:01:15,670 --> 00:01:19,452
it was a page per second, 500 dots per 
inch, and if you do the math, that's 

16
00:01:19,452 --> 00:01:24,754
about 20 megabits per second. 
So the existing methods of 

17
00:01:24,754 --> 00:01:29,350
interconnection had a lot of problems. 
One is that they were all home run, so 

18
00:01:29,350 --> 00:01:32,760
all these wires, one from every desk, are 
all coming to this one place in the 

19
00:01:32,760 --> 00:01:38,579
building and put a big rat's nest. 
We called it a rat's nest by the way. 

20
00:01:38,579 --> 00:01:43,250
Two is they generally ran at 200 bits per 
second or. 

21
00:01:43,250 --> 00:01:47,282
If you really rate them up, they go to 14 
4 kilobits per second, and that wasn't 

22
00:01:47,282 --> 00:01:52,197
even close to 20 megabits per second. 
And we wanted to be able to keep the 

23
00:01:52,197 --> 00:01:55,461
printer busy by sending documents from 
all these PCs which hadn't been built 

24
00:01:55,461 --> 00:01:58,792
yet. 
 >> Except at park. 

25
00:01:58,792 --> 00:02:01,478
 >> Well, they hadn't been built even at 
Park at that point. 

26
00:02:01,478 --> 00:02:02,990
 >> Okay. 
 >> So we were building the printer. 

27
00:02:02,990 --> 00:02:06,090
We were building the PCs all at the same 
time. 

28
00:02:07,450 --> 00:02:12,779
So, I started worked on this, and there 
was a preceding effort before mine called 

29
00:02:12,779 --> 00:02:17,604
Signet at Park. 
And it was being done by Charles Simony, 

30
00:02:17,604 --> 00:02:22,641
my friend, and but he wasn't a networking 
guy, so they sent me in to pick up Signet 

31
00:02:22,641 --> 00:02:27,363
from Charles. 
And he was going to go off and do 

32
00:02:27,363 --> 00:02:30,720
something else. 
And by the way the other thing he did, is 

33
00:02:30,720 --> 00:02:35,220
he wrote a word a text editor, a word 
processor, called Bravo, which then 

34
00:02:35,220 --> 00:02:41,628
became Microsoft Office. 
So, Charles is a billionarie and he has 

35
00:02:41,628 --> 00:02:47,620
been to the space station. 
Twice, the International Spa. 

36
00:02:47,620 --> 00:02:51,460
So maybe I shouldn't have kicked him off 
Signet, maybe I should have done Bravo. 

37
00:02:51,460 --> 00:02:54,986
I don't know. 
Anyway, so I, immediately decided that 

38
00:02:54,986 --> 00:03:00,834
Signet, by the way, SIGNet stands for 
Simonyi's Infinitely Glorious Network was 

39
00:03:00,834 --> 00:03:09,048
in fact infinitely glorious and much had 
too many moving parts to be a LAN. 

40
00:03:09,048 --> 00:03:13,144
By the way, the word LAN wasn't invented 
until 1990, and this is still 1973, so 

41
00:03:13,144 --> 00:03:18,504
the word LAN was still way in the future, 
so that's an anachronism. 

42
00:03:18,504 --> 00:03:22,627
Anyway, the Signet had too many moving 
parts for a LAN. 

43
00:03:22,627 --> 00:03:27,181
And it was in the course of investigating 
how to organize this LAN that I ran into 

44
00:03:27,181 --> 00:03:31,735
a packet, quite by accident, a packet 
radio network at the University of Hawaii 

45
00:03:31,735 --> 00:03:37,811
that was called the Aloha Net. 
And what was beautiful about the Aloha 

46
00:03:37,811 --> 00:03:40,922
Network was it had a solved a distributed 
problem. 

47
00:03:40,922 --> 00:03:44,982
That is, how would we share a radio 
channel back to the mainframe in, in 

48
00:03:44,982 --> 00:03:49,906
Hawaii, at the university? 
If we're just a bunch of terminals 

49
00:03:49,906 --> 00:03:54,736
scattered among the Hawaiian islands and 
they couldn't really easily talk to each 

50
00:03:54,736 --> 00:03:59,890
other to get coordinated? 
How would they coordinate their sharing 

51
00:03:59,890 --> 00:04:04,052
of this inbound radio channel? 
And Norm Abramson, there at the 

52
00:04:04,052 --> 00:04:07,643
University of Hawaii devised this very 
simple randomized retransmission 

53
00:04:07,643 --> 00:04:12,542
procedure, where a person would type. 
By the way what they would type was a 

54
00:04:12,542 --> 00:04:16,399
card image. 
It was 80 columns wide, back from the 

55
00:04:16,399 --> 00:04:21,850
days of card batch processing. 
So, you would type in your card image and 

56
00:04:21,850 --> 00:04:24,853
hit send. 
And then your terminal would send it in 

57
00:04:24,853 --> 00:04:28,821
toward the main frame, and then would 
wait a short time to see if there was an 

58
00:04:28,821 --> 00:04:34,660
acknowledgement that returned on the 
outbound channel. 

59
00:04:34,660 --> 00:04:37,787
And if there was, everything was 
hunky-dory, but if there wasn't, that 

60
00:04:37,787 --> 00:04:42,059
probably meant that two terminals decided 
to send at the same time. 

61
00:04:43,110 --> 00:04:48,074
So, then as many terminals particpated in 
that collision of transmissions would 

62
00:04:48,074 --> 00:04:54,030
then randimoize and then re-transimit at 
a random time in the future. 

63
00:04:54,030 --> 00:04:57,130
That way if they overlapped here they 
would not overlap again in the future 

64
00:04:57,130 --> 00:05:01,248
because thye would choose different 
random numbers to count down. 

65
00:05:01,248 --> 00:05:05,780
So that's randomizory transmission, 
multiple access. 

66
00:05:07,050 --> 00:05:11,250
Since I was trying to avoid this big 
rat's nest of wires and only wanted to 

67
00:05:11,250 --> 00:05:15,800
have one wire, just one wire and not 16 
or 32 or whatever the alternatives were, 

68
00:05:15,800 --> 00:05:21,118
just one. 
I wanted a distributed solution to how to 

69
00:05:21,118 --> 00:05:25,438
share this cable, and the Aloha Network 
produced that using randomized 

70
00:05:25,438 --> 00:05:29,262
retransmissions. 
So then, I, 

71
00:05:29,262 --> 00:05:33,070
Actually y'know, there's sort of two 
stories here. 

72
00:05:33,070 --> 00:05:35,420
One is a hardware story. 
 >> I was wonderin' about the hardware. 

73
00:05:35,420 --> 00:05:37,910
 >> There's a hardware story, and 
there's a software story. 

74
00:05:37,910 --> 00:05:42,308
And the hardware story is , not that 
interesting but it's there. 

75
00:05:42,308 --> 00:05:46,598
So one of the first things I did was to 
buy, a kilometer of cable, or it may have 

76
00:05:46,598 --> 00:05:51,577
been a mile I don't know if I'd gone 
metric yet, but I. 

77
00:05:51,577 --> 00:05:56,259
So a spool of coax cable about this big, 
with the two ends sticking up. 

78
00:05:56,259 --> 00:05:59,913
So I got a, pulse generator and I hooked 
it up to one end and hooked it around 

79
00:05:59,913 --> 00:06:03,915
back into the oscilloscope, and started 
launching square waves down the cable and 

80
00:06:03,915 --> 00:06:08,808
watching what came out. 
I thought that would be good preparation 

81
00:06:08,808 --> 00:06:11,745
for building a network. 
And what came out the other side wasn't a 

82
00:06:11,745 --> 00:06:15,890
square wave. 
It was sort of lazy rise times and lazy 

83
00:06:15,890 --> 00:06:21,210
fall times. 
But if you put a digital gate, you could 

84
00:06:21,210 --> 00:06:25,347
recover the square wave. 
That is just set, use the digital 

85
00:06:25,347 --> 00:06:28,900
threshold, so out of this gate came a 
square wave. 

86
00:06:28,900 --> 00:06:33,460
So you could recover, at the end of a 
mile of cable this square wave. 

87
00:06:33,460 --> 00:06:36,630
So I kind of knew. 
I kind of had some confidence then that 

88
00:06:36,630 --> 00:06:40,982
we, that various stations connected to 
this cable could inject their square 

89
00:06:40,982 --> 00:06:45,612
waves and, the other guys could recover 
them. 

90
00:06:45,612 --> 00:06:53,370
So that hardware wasn't that hard that 
did it was, a, straightforward. 

91
00:06:53,370 --> 00:06:55,146
 >> This was the first hardware. 
was it kind of relatively 

92
00:06:55,146 --> 00:06:56,990
straightforward? 
 >> Very straightforward. 

93
00:06:56,990 --> 00:06:59,730
And the and the square ware , by the way 
was called Manchester. 

94
00:06:59,730 --> 00:07:03,570
What we'd do is we'd take the bit, we'd 
make a packet of bits, and then we'd send 

95
00:07:03,570 --> 00:07:08,450
them one at a time down this cable, and 
we would encode them. 

96
00:07:08,450 --> 00:07:12,103
And the encoding was also simple, it was 
Manchester coding. 

97
00:07:12,103 --> 00:07:16,129
Meaning that for each bit the first half 
of the bit would be the complement of the 

98
00:07:16,129 --> 00:07:19,009
bit. 
And the second half of the bit would be 

99
00:07:19,009 --> 00:07:21,650
the bit. 
So you would have a transition in the 

100
00:07:21,650 --> 00:07:26,130
middle of each bit cell, which is a very 
simple modulation scheme. 

101
00:07:26,130 --> 00:07:28,796
So as you're sending the packet the 
cables wiggling and you can recover the 

102
00:07:28,796 --> 00:07:32,449
signal at the other end then. 
Clock those bits into a shift register, 

103
00:07:32,449 --> 00:07:35,029
and then click the sh, clock the shift 
register into the memory, and then 

104
00:07:35,029 --> 00:07:39,150
collect a packet that way. 
 >> And that all took just some 

105
00:07:39,150 --> 00:07:41,470
soldering to make that all happen 
basically. 

106
00:07:41,470 --> 00:07:44,921
 >> Well I don't wanna[LAUGH]. 
It got a little bit complicated when you 

107
00:07:44,921 --> 00:07:47,960
wanted to put 255 of them on a mile of 
cable. 

108
00:07:47,960 --> 00:07:50,880
You had to be a little bit careful about 
the impedance on the taps. 

109
00:07:50,880 --> 00:07:52,249
 >> Okay. 
 >> because the square wave by a tap 

110
00:07:52,249 --> 00:07:55,726
might generate reflections that would 
then interfere. 

111
00:07:56,740 --> 00:08:00,304
But the beauty of Manchester encoding was 
that while you were sending a packet, 

112
00:08:00,304 --> 00:08:04,365
there were constant transitions. 
So if you listened, you could tell 

113
00:08:04,365 --> 00:08:08,530
whether a packet was going by and you 
didn't have to listen for long. 

114
00:08:08,530 --> 00:08:12,338
You only had to listen for about a bit 
time, which turned out to be 340 

115
00:08:12,338 --> 00:08:15,785
nanoseconds. 
So you could wait that long, and tell 

116
00:08:15,785 --> 00:08:20,782
whether there was a packet going by. 
So one of the first differences between 

117
00:08:20,782 --> 00:08:25,198
the Ethernet, and the Aloha net, we got a 
lot of differences, but one of the first 

118
00:08:25,198 --> 00:08:30,210
ones was carrier sense. 
In the Aloha network. 

119
00:08:30,210 --> 00:08:32,784
You couldn't tell if somebody else was 
transmitting at the same time as you, but 

120
00:08:32,784 --> 00:08:36,243
on the Ethernet, you could. 
And the advantage of that was, you might 

121
00:08:36,243 --> 00:08:38,990
as well, if you're sending and somebody 
else is sending at the same timeyou might 

122
00:08:38,990 --> 00:08:43,010
as well give up, because you've destroyed 
each other's packets, so punt. 

123
00:08:43,010 --> 00:08:47,365
Then that recovers that bandwidth that 
otherwise would be lost just continuing 

124
00:08:47,365 --> 00:08:52,131
to transmit a damaged packet. 
And the the other was this, this 

125
00:08:52,131 --> 00:08:57,946
Manchester code meant that the cable was 
on half the time and of half the time. 

126
00:08:57,946 --> 00:09:02,033
That is the, the driver, so the open, 
what we used to call an open collector 

127
00:09:02,033 --> 00:09:07,500
driver would either yank the cable up to 
three, four, five volts. 

128
00:09:07,500 --> 00:09:11,270
I even forget the voltage now. 
but it was under five, I'm sure of that. 

129
00:09:11,270 --> 00:09:14,035
I had a rule. 
I never went above five volts. 

130
00:09:14,035 --> 00:09:17,615
you would yank the cable up. 
And then during the other half of the bit 

131
00:09:17,615 --> 00:09:21,493
when you weren't yanking you let go. 
So, so for example in each bit cell you 

132
00:09:21,493 --> 00:09:26,194
got to look to see if anyone else was 
transmitting when you had stopped. 

133
00:09:26,194 --> 00:09:29,580
And if there was somebody else, then you 
had a collision. 

134
00:09:29,580 --> 00:09:32,690
So that was the second feature. 
 >> So it was really a digital signal. 

135
00:09:32,690 --> 00:09:36,490
It was really a digital signal, it wasn't 
a modulated signal in any way. 

136
00:09:36,490 --> 00:09:37,946
 >> Right? 
 >> No,well it was, 

137
00:09:37,946 --> 00:09:41,045
 >> At best you could send additional 
signal over[UNKNOWN]. 

138
00:09:41,045 --> 00:09:43,240
 >> On/off, with the Manchester 
encoding. 

139
00:09:43,240 --> 00:09:47,020
So Manchester encoding is akin to a 
modulation scheme. 

140
00:09:47,020 --> 00:09:48,560
But it's the simplest one you can think 
of. 

141
00:09:48,560 --> 00:09:52,060
 >> It's very base bane. 
 >> Yeah, it's very base bane. 

142
00:09:52,060 --> 00:09:54,920
But it's the sort of modulation scheme 
that a computer scientist would come up 

143
00:09:54,920 --> 00:09:57,850
with as apposed to one of those fancy 
 >> Radio people. 

144
00:09:57,850 --> 00:09:58,650
 >> Radio people. 
 >> Right. 

145
00:09:58,650 --> 00:10:02,340
So, and, and by the way, took a lot of 
gas from those radio people. 

146
00:10:02,340 --> 00:10:04,810
Why do you use such a simple scheme? 
 >> Right. 

147
00:10:04,810 --> 00:10:05,990
 >> You're wasting. 
 >> [CROSSTALK] . 

148
00:10:05,990 --> 00:10:08,370
 >> You would have wasted all that, you 
wasted all that bandwidth. 

149
00:10:08,370 --> 00:10:12,234
That capable, that cable was capable of 
carrying hundreds of megabits per seconds 

150
00:10:12,234 --> 00:10:15,970
and you only carried 2.9 4. 
What a waste of the cable. 

151
00:10:15,970 --> 00:10:19,353
 >> Exactly[/g] . 
[LAUGH] Well, the --[LAUGH] there was 

152
00:10:19,353 --> 00:10:24,278
plenty of cable to waste. 
So we talk about carrier sense inclusion 

153
00:10:24,278 --> 00:10:30,130
detection, and the scheme would be that 
each packet would carry two addresses. 

154
00:10:30,130 --> 00:10:34,840
The address of the destination and the 
address of the source. 

155
00:10:34,840 --> 00:10:38,100
And each of these would be 8 bits. 
And so on the backplane of these little 

156
00:10:38,100 --> 00:10:42,894
personal computers with wire app[SOUND]. 
You would wire app in a code between zero 

157
00:10:42,894 --> 00:10:46,860
and 255 and that would be the serial 
number of the machine. 

158
00:10:46,860 --> 00:10:50,940
And that, and then you'd read that off 
the back plane and put it in the pack. 

159
00:10:50,940 --> 00:10:53,975
So each packet had two addresses. 
This is different from the Aloha incident 

160
00:10:53,975 --> 00:10:57,738
which had one address. 
Because the channel was only going. 

161
00:10:57,738 --> 00:11:02,864
Through two channels. 
So two addresses, and we added a cyclic 

162
00:11:02,864 --> 00:11:07,830
redundancy checksum on the end of the 
packet, which we implemented in hardware. 

163
00:11:07,830 --> 00:11:14,610
So, that you in addition so you could 
tell if the packet had been damaged. 

164
00:11:14,610 --> 00:11:19,032
So if there was a collision and the 
terminal, the contending stations backed 

165
00:11:19,032 --> 00:11:21,602
off. 
There would be a hunk of garbage on the 

166
00:11:21,602 --> 00:11:24,618
cables sitting around. 
But when it was received, the check sum 

167
00:11:24,618 --> 00:11:27,450
wouldn't add up and you'd just throw that 
chunk of garbage away. 

168
00:11:27,450 --> 00:11:32,427
The day that I was launching pulses down 
this spool of cable, I was doing some 

169
00:11:32,427 --> 00:11:37,325
soddering, and I was doing some knife 
work to get the insulation off the 

170
00:11:37,325 --> 00:11:43,935
copper. 
And across the room was a young grad 

171
00:11:43,935 --> 00:11:49,045
student, who was doing something else and 
he noticed me not being good at this. 

172
00:11:49,045 --> 00:11:52,231
And it turns out, he was very good at it, 
because he had worked in a television 

173
00:11:52,231 --> 00:11:55,660
studio and he had worked in cable 
television. 

174
00:11:55,660 --> 00:11:58,660
So he knew all about skinning wires and 
coaxes. 

175
00:11:58,660 --> 00:12:01,460
So he came over to help me, and that was 
da, his name was David Boggs. 

176
00:12:02,880 --> 00:12:08,760
And then we started working together, and 
invented ethernet together. 

177
00:12:08,760 --> 00:12:13,472
So, he was, and he was, he was, slightly 
more hardware, and I was slight more 

178
00:12:13,472 --> 00:12:17,560
software. 
But there was then a third guy. 

179
00:12:17,560 --> 00:12:21,190
Who was even more hardware than David. 
Who was the picofarad guy. 

180
00:12:21,190 --> 00:12:24,786
The guy who would put those last few 
passive components on the end just to be 

181
00:12:24,786 --> 00:12:28,672
sure, that, that connection to the 
coaxial cable would be clean for 

182
00:12:28,672 --> 00:12:33,185
transmissions. 
Every time you tapped into it you didn't 

183
00:12:33,185 --> 00:12:36,530
put a lump of impedance on it. 
 >> Yeah, the tapping seems like a 

184
00:12:36,530 --> 00:12:40,388
counter intuitive thing to me. 
I mean, think the, I mean everyone would 

185
00:12:40,388 --> 00:12:44,610
think that a star is the right thing, but 
you were going to tap into that. 

186
00:12:44,610 --> 00:12:47,662
Just tap into[UNKNOWN] . 
 >> One of the problems we decided to 

187
00:12:47,662 --> 00:12:51,410
solve was the rat's nest problem. 
We did not want a rats nest, and every 

188
00:12:51,410 --> 00:12:54,277
time we installed a new PC we didn't have 
to home run a cable back to the rats 

189
00:12:54,277 --> 00:12:57,263
nest. 
So we wanted to put one cable down the 

190
00:12:57,263 --> 00:13:00,239
middle of the corridor and then every 
time you want to put a PC you just run up 

191
00:13:00,239 --> 00:13:04,696
and tap into it. 
And we didn't want the network to go on, 

192
00:13:04,696 --> 00:13:07,576
go down while you were tapping into it, 
because we wanted 24 by seven access to 

193
00:13:07,576 --> 00:13:10,878
the network. 
So there had to be a way to tap into the 

194
00:13:10,878 --> 00:13:15,470
network without bringing it down. 
So that led to a device we found in the 

195
00:13:15,470 --> 00:13:20,968
cable TV industry. 
It was called gerald tap, and it was 

196
00:13:20,968 --> 00:13:25,482
basically a, a vampire tap. 
You'd, you'd drill a little hole in the 

197
00:13:25,482 --> 00:13:29,325
outer, casing of the coax, and then you 
would screw in this tap that would 

198
00:13:29,325 --> 00:13:34,630
puncture the insulation and go right to 
the copper and tap in. 

199
00:13:34,630 --> 00:13:37,336
And you notice in that operation you're 
not breaking the copper, so the network 

200
00:13:37,336 --> 00:13:40,706
continues sending, you tap in, and you're 
now part of the network. 

201
00:13:42,220 --> 00:13:43,560
 >> So that came from the cable 
industry. 

202
00:13:43,560 --> 00:13:44,930
 >> It did. 
 >> Oh, okay. 

203
00:13:44,930 --> 00:13:48,464
 >> So a guy named David Ladell, who had 
done cable TV installations when he was 

204
00:13:48,464 --> 00:13:52,746
in grad school in Toledo. 
suggested that we use the gerald tap 

205
00:13:52,746 --> 00:13:56,394
since it was already being made in volume 
and worked just fine, as far as he was 

206
00:13:56,394 --> 00:14:01,214
concerned. 
And it allowed us to solve the rat's 

207
00:14:01,214 --> 00:14:04,294
nest, what we perceived to be an 
important problem, the rat's nest 

208
00:14:04,294 --> 00:14:07,573
problem. 
The Aloha network ran in the 

209
00:14:07,573 --> 00:14:12,543
kilobit-per-second range, like 4800 or I 
don't remember the numbers, but kilobits 

210
00:14:12,543 --> 00:14:16,694
per second. 
The Ethernet then started at 2.94 

211
00:14:16,694 --> 00:14:19,107
megabit. 
And in those days, by the way, T1 was 

212
00:14:19,107 --> 00:14:22,162
1.544 megabytes per second. 
So in 1973, the ethernet was already as 

213
00:14:22,162 --> 00:14:25,350
fast as T1. 
Of course T1 is still around, oddly. 

214
00:14:25,350 --> 00:14:37,505
But then we went from 2.94, we briefly 
went to 20 megabytes per second inside of 

215
00:14:37,505 --> 00:14:44,409
Xerox. 
And then when we bumped into DEC and 

216
00:14:44,409 --> 00:14:49,620
Intel for the standardization process in 
802, we decided on ten. 

217
00:14:49,620 --> 00:14:52,690
We came down from 20 to ten so the chips 
would work. 

218
00:14:54,020 --> 00:14:57,796
And then we went from ten, and later, I 
helped found a company called Grand 

219
00:14:57,796 --> 00:15:02,110
Junction Networks, that introduced the 
100-mg Ethernet. 

220
00:15:02,110 --> 00:15:06,465
And I remember being at my coffee table 
in Palo Alto, I think it was. 

221
00:15:06,465 --> 00:15:11,430
I forget the year. 
No, maybe it was in Woodside. 

222
00:15:11,430 --> 00:15:15,655
But we were trying to think of how we 
would make a fast Ethernet, and there's 

223
00:15:15,655 --> 00:15:19,841
some math that shows that if, as you go 
faster. 

224
00:15:19,841 --> 00:15:24,066
The efficiency depends on the diameter of 
the network, and as you go faster and 

225
00:15:24,066 --> 00:15:29,962
faster, the efficiency goes down. 
The diameter of the network in bits. 

226
00:15:29,962 --> 00:15:33,856
And when you're trying to up a factor of 
ten, and it I don't know who it was, one 

227
00:15:33,856 --> 00:15:37,455
of us observed, that wait a minute, we've 
been assuming that Ethernet is a 

228
00:15:37,455 --> 00:15:42,812
kilometer in diameter. 
But we're going, we're all going to hubs 

229
00:15:42,812 --> 00:15:46,497
now so that we only need 100 meters, not 
1000 meters. 

230
00:15:46,497 --> 00:15:50,214
And that was the factor of ten right 
there, so that by changing the collision 

231
00:15:50,214 --> 00:15:54,047
interval. 
You maintain the same efficiencies, 

232
00:15:54,047 --> 00:15:58,603
theoretical efficiencies by just assuming 
you're going 100 meters instead of a 

233
00:15:58,603 --> 00:16:02,814
kilometer. 
So that got us to 100 megabits per second 

234
00:16:02,814 --> 00:16:07,230
and not gigabit, ten gigabit which I 
guess is the mainstream now, and then 100 

235
00:16:07,230 --> 00:16:12,911
gigabit is now beginning to run. 
100 gigabits per second. 

236
00:16:12,911 --> 00:16:16,817
You can't, you can't be a computer 
scientist and build that kind of digital 

237
00:16:16,817 --> 00:16:19,946
now. 
You have to actually be an engineer, a 

238
00:16:19,946 --> 00:16:24,443
hardware engineer to run at a 100. 
 >> I think at that point, you're pretty 

239
00:16:24,443 --> 00:16:28,045
much a radio person. 
 >> [LAUGH] But then after 100 gigabits 

240
00:16:28,045 --> 00:16:33,042
comes terrabits, so i've already begun 
giving talks on terrabit ethernet. 

241
00:16:33,042 --> 00:16:53,075
[MUSIC]