1
00:00:04,240 --> 00:00:07,240
Speaker 1: Welcome to tex Stuff, a production of I Heart Radios,

2
00:00:07,320 --> 00:00:13,880
Speaker 1: How Stuff Works. Hey there, and welcome to tech Stuff.

3
00:00:13,920 --> 00:00:17,159
Speaker 1: I'm your host, Jonathan Strickland. I'm an executive producer with

4
00:00:17,200 --> 00:00:19,320
Speaker 1: How Stuff Works and I heart Radio and I love

5
00:00:19,400 --> 00:00:23,200
Speaker 1: all things tech. And today's episode comes to us due

6
00:00:23,320 --> 00:00:29,040
Speaker 1: to a listener request. Martin many many moons ago asked

7
00:00:29,080 --> 00:00:31,880
Speaker 1: if I could do an episode talking about the evolution

8
00:00:32,000 --> 00:00:35,520
Speaker 1: of virtualization, and so that's what this episode is all about.

9
00:00:35,920 --> 00:00:41,520
Speaker 1: But first, what is virtualization? Well, as the name implies,

10
00:00:41,640 --> 00:00:46,600
Speaker 1: it's all about creating a virtual representation of something else. Now,

11
00:00:46,640 --> 00:00:50,479
Speaker 1: in computing, we typically use it to mean creating virtual

12
00:00:50,640 --> 00:00:54,360
Speaker 1: computer platforms, and there are a lot of reasons why

13
00:00:54,560 --> 00:00:56,560
Speaker 1: you would want to do this. It would be a

14
00:00:56,640 --> 00:01:01,160
Speaker 1: virtual computer platform that exists on top of actual physical hardware.

15
00:01:01,600 --> 00:01:04,040
Speaker 1: And you might want to do it for security reasons,

16
00:01:04,319 --> 00:01:08,440
Speaker 1: to separate different programs and storage systems from one another.

17
00:01:08,920 --> 00:01:11,160
Speaker 1: You might want to do it to provide different stable

18
00:01:11,280 --> 00:01:16,120
Speaker 1: development environments so that a mistake in one partition doesn't

19
00:01:16,120 --> 00:01:19,119
Speaker 1: bring everything else down on the machine. You might want

20
00:01:19,160 --> 00:01:21,560
Speaker 1: to do it to run different operating systems on the

21
00:01:21,600 --> 00:01:24,200
Speaker 1: same physical hardware, or maybe you want to do it

22
00:01:24,240 --> 00:01:28,080
Speaker 1: to maximize efficiency in your computer system, or perhaps some

23
00:01:28,160 --> 00:01:31,520
Speaker 1: combination of some or all of the above, but we'll

24
00:01:31,520 --> 00:01:33,920
Speaker 1: get into that later. For now, the important thing, no,

25
00:01:34,480 --> 00:01:37,000
Speaker 1: is that the goal is to create a virtual version

26
00:01:37,120 --> 00:01:41,160
Speaker 1: of something that runs on top of actual physical hardware. Now,

27
00:01:41,160 --> 00:01:45,119
Speaker 1: the history of virtualization stretches back to the nineteen sixties,

28
00:01:45,480 --> 00:01:48,400
Speaker 1: well before the days when the average person had any

29
00:01:48,440 --> 00:01:52,720
Speaker 1: sort of interactions with a computer. Early computers were useful

30
00:01:53,120 --> 00:01:56,400
Speaker 1: but had some limitations. This is like the main frame

31
00:01:56,600 --> 00:02:00,240
Speaker 1: age of computers. So one of those limitations was that

32
00:02:00,360 --> 00:02:04,040
Speaker 1: early computers could really only run one process at a time,

33
00:02:04,440 --> 00:02:06,800
Speaker 1: and there had to be a person there to initiate

34
00:02:07,000 --> 00:02:10,000
Speaker 1: a new program. At first, that wasn't that big of

35
00:02:10,000 --> 00:02:13,679
Speaker 1: a problem because computers were such a niche gadget. Very

36
00:02:13,680 --> 00:02:16,080
Speaker 1: few people had any access to them in the first place,

37
00:02:16,080 --> 00:02:18,960
Speaker 1: so it wasn't that big of a limitation. The scientific

38
00:02:19,000 --> 00:02:23,280
Speaker 1: and academic communities were using them, but beyond them, very

39
00:02:23,400 --> 00:02:26,880
Speaker 1: few people had any experience with a computer. They were

40
00:02:26,880 --> 00:02:31,639
Speaker 1: mostly unknown. But as computers were becoming more available, as

41
00:02:31,680 --> 00:02:35,160
Speaker 1: businesses were starting to use them, and they were becoming

42
00:02:35,160 --> 00:02:39,160
Speaker 1: more capable of handling processes that went beyond academic interests,

43
00:02:39,520 --> 00:02:43,080
Speaker 1: it became clear that the limitations were a liability, and

44
00:02:43,120 --> 00:02:46,399
Speaker 1: so you had companies like IBM researching ways to get

45
00:02:46,400 --> 00:02:50,480
Speaker 1: around these limitations. One such way was an approach later

46
00:02:50,720 --> 00:02:54,680
Speaker 1: called batch processing. This dates from the time when programs

47
00:02:54,680 --> 00:02:58,920
Speaker 1: were represented on physical cards with holes punched in them,

48
00:02:59,000 --> 00:03:02,919
Speaker 1: known as punch cards for obvious reasons. So in an

49
00:03:02,919 --> 00:03:07,240
Speaker 1: early computer, you'd have a stack of cards representing a

50
00:03:07,440 --> 00:03:10,840
Speaker 1: single program, and you would feed those cards into the

51
00:03:10,880 --> 00:03:15,280
Speaker 1: computer's hopper, which is essentially its intake for cards, and

52
00:03:15,320 --> 00:03:19,280
Speaker 1: the computer would analyze the cards, which would have instructions

53
00:03:19,320 --> 00:03:23,080
Speaker 1: on them. The computer would then follow those instructions and

54
00:03:23,120 --> 00:03:26,480
Speaker 1: produce a result and either printed out or create a

55
00:03:26,480 --> 00:03:29,840
Speaker 1: new punch card or stack of punch cards, or later

56
00:03:29,919 --> 00:03:33,880
Speaker 1: on it would display the results on a monitor. IBM

57
00:03:33,919 --> 00:03:37,680
Speaker 1: developed computers that could accept stacks of cards that represented

58
00:03:37,760 --> 00:03:40,920
Speaker 1: more than one process. The computer would be able to

59
00:03:40,960 --> 00:03:44,160
Speaker 1: read out the stacks and complete each process in turn,

60
00:03:44,520 --> 00:03:47,840
Speaker 1: so you could feed batches of cards to the computer.

61
00:03:48,200 --> 00:03:51,640
Speaker 1: Thus batch processing it cut down on the amount of

62
00:03:51,680 --> 00:03:55,200
Speaker 1: time people had to spend babysitting a computer or waiting

63
00:03:55,200 --> 00:03:57,520
Speaker 1: for their turn to run a process, freeing them up

64
00:03:57,520 --> 00:04:00,560
Speaker 1: to work on other stuff. But was still limited, and

65
00:04:00,560 --> 00:04:03,800
Speaker 1: you were still stuck running just one process at a time.

66
00:04:03,840 --> 00:04:06,600
Speaker 1: You would just run them in sequence, and you couldn't

67
00:04:06,640 --> 00:04:09,440
Speaker 1: easily have multiple people using the same computer at the

68
00:04:09,480 --> 00:04:12,600
Speaker 1: same time. At best, you could use batch processing to

69
00:04:12,720 --> 00:04:15,640
Speaker 1: run a second job right after the first one, but

70
00:04:15,760 --> 00:04:18,800
Speaker 1: it was restrictive, and as computers were becoming more important,

71
00:04:19,040 --> 00:04:22,160
Speaker 1: this was a problem in the late fifties and early

72
00:04:22,320 --> 00:04:27,120
Speaker 1: nineteen sixties some engineers or computer scientists, though that term

73
00:04:27,240 --> 00:04:30,520
Speaker 1: was hardly in use yet, and many in academic fields

74
00:04:30,520 --> 00:04:33,080
Speaker 1: were very snooty. They didn't quite yet view the subject

75
00:04:33,160 --> 00:04:36,640
Speaker 1: as worthy of standing on its own anyway, Some of

76
00:04:36,680 --> 00:04:39,760
Speaker 1: these innovators began to experiment with ways to get around

77
00:04:39,960 --> 00:04:44,160
Speaker 1: these limitations. A fellow named John McCarthy, a professor at

78
00:04:44,160 --> 00:04:47,960
Speaker 1: the Massachusetts Institute of Technology better known as m i T,

79
00:04:48,520 --> 00:04:52,080
Speaker 1: proposed in nineteen fifty nine that an IBM seven O

80
00:04:52,279 --> 00:04:55,279
Speaker 1: nine main frame could be tweaked to allow for a

81
00:04:55,360 --> 00:04:59,320
Speaker 1: few people to use the machine more or less at

82
00:04:59,360 --> 00:05:02,120
Speaker 1: the same time time. He then began to recommend some

83
00:05:02,240 --> 00:05:05,920
Speaker 1: changes to the university's IBM seven oh four main frame

84
00:05:06,200 --> 00:05:10,400
Speaker 1: to allow what he called a time stealing mode. In

85
00:05:10,440 --> 00:05:12,839
Speaker 1: this mode, you could have a batch of jobs running

86
00:05:12,880 --> 00:05:15,760
Speaker 1: on a computer, then someone else with an unrelated job

87
00:05:15,800 --> 00:05:19,000
Speaker 1: shows up, and, using time stealing, the new person could

88
00:05:19,080 --> 00:05:21,960
Speaker 1: interrupt the batch process to run this other job, and

89
00:05:22,000 --> 00:05:25,920
Speaker 1: then the batch process could resume as per normal. Another

90
00:05:26,080 --> 00:05:29,760
Speaker 1: m i T professor named Fernando Corbato worked on a

91
00:05:29,800 --> 00:05:33,599
Speaker 1: similar project, tweaking the university's IBM seven oh nine main

92
00:05:33,640 --> 00:05:36,400
Speaker 1: frame so that four people could use the system at

93
00:05:36,400 --> 00:05:39,599
Speaker 1: the same time. But these were all solutions using systems

94
00:05:39,600 --> 00:05:44,120
Speaker 1: that weren't natively designed to support multiple users. They were workarounds,

95
00:05:44,400 --> 00:05:48,479
Speaker 1: and m i T computer scientists found manufacturers were uninterested

96
00:05:48,520 --> 00:05:52,200
Speaker 1: in changing that because there seemed to be very little

97
00:05:52,240 --> 00:05:55,839
Speaker 1: call for it. On July one, nineteen sixty three, m

98
00:05:55,880 --> 00:05:59,960
Speaker 1: i T launched a project originally called Mathematics and Computation

99
00:06:00,040 --> 00:06:04,240
Speaker 1: in or MAC, but later the acronym would be retroactively

100
00:06:04,320 --> 00:06:08,800
Speaker 1: applied to the phrase multiple access computer. Funding for the

101
00:06:08,839 --> 00:06:12,080
Speaker 1: project came courtesy of ar PA, which would later get

102
00:06:12,120 --> 00:06:15,560
Speaker 1: its own acronym update to DARPA. And in case you

103
00:06:15,560 --> 00:06:19,200
Speaker 1: are unfamiliar with that organization, it's a division within the

104
00:06:19,279 --> 00:06:22,800
Speaker 1: United States Department of Defense, and its mission is to

105
00:06:22,920 --> 00:06:26,400
Speaker 1: fund research and development into technologies that contribute to the

106
00:06:26,440 --> 00:06:29,719
Speaker 1: defense of the country in some way. So why was

107
00:06:29,760 --> 00:06:33,480
Speaker 1: the Department of Defense interested in this? It largely had

108
00:06:33,520 --> 00:06:36,719
Speaker 1: to do with Russia and Sputnik. Now, I've talked about

109
00:06:36,720 --> 00:06:40,160
Speaker 1: how spot Nick helped spur on a ton of innovation

110
00:06:40,240 --> 00:06:44,120
Speaker 1: in the United States. It scared the daylights out of

111
00:06:44,240 --> 00:06:47,280
Speaker 1: people in the US. It suggested that Russia was much

112
00:06:47,360 --> 00:06:51,520
Speaker 1: further along technologically speaking than the US had suspected. And

113
00:06:51,560 --> 00:06:54,680
Speaker 1: it lit a fire under the proverbial backside of the

114
00:06:54,760 --> 00:06:57,960
Speaker 1: US military. And so there was a strong incentive to

115
00:06:58,040 --> 00:07:00,960
Speaker 1: advance computer science and technolog g in the US to

116
00:07:01,080 --> 00:07:06,080
Speaker 1: outpace the Russians. Project max primary purpose was to advance

117
00:07:06,120 --> 00:07:09,880
Speaker 1: computer science in several ways, including the development of new

118
00:07:09,960 --> 00:07:14,080
Speaker 1: operating systems and computational theory. It largely grew out of

119
00:07:14,160 --> 00:07:17,560
Speaker 1: a meeting between m i T Professor Robert Fano and

120
00:07:17,640 --> 00:07:21,080
Speaker 1: Joseph C. R. Lick Lighter, who had previously established a

121
00:07:21,120 --> 00:07:24,240
Speaker 1: psychology group in the Electrical Engineering department of m i T.

122
00:07:24,840 --> 00:07:27,840
Speaker 1: Then gone on to join a research firm called Bolt,

123
00:07:27,960 --> 00:07:31,240
Speaker 1: Baroneck and Newman better known as B B N, and

124
00:07:31,280 --> 00:07:35,200
Speaker 1: then was named the first director of ARPA's Information Processing

125
00:07:35,280 --> 00:07:39,520
Speaker 1: Techniques Office or I p t O. Lick Lighter convinced

126
00:07:39,520 --> 00:07:43,240
Speaker 1: Fano to head up Project Mac, which would receive funding

127
00:07:43,400 --> 00:07:46,800
Speaker 1: from the I p t O through ARPA. But standing

128
00:07:46,800 --> 00:07:49,600
Speaker 1: in the way of this goal were the limitations I've

129
00:07:49,600 --> 00:07:52,840
Speaker 1: mentioned already. It's hard to make real progress if you're

130
00:07:52,880 --> 00:07:55,920
Speaker 1: limited to just one person working on one computer at

131
00:07:55,960 --> 00:07:58,680
Speaker 1: any given time, and so M I T sought out

132
00:07:58,680 --> 00:08:02,520
Speaker 1: proposals from companies like General Electric and IBM to develop

133
00:08:02,560 --> 00:08:05,800
Speaker 1: a computer system that could support multiple users through this

134
00:08:06,080 --> 00:08:10,480
Speaker 1: time sharing idea. Now, the basic idea behind time sharing

135
00:08:10,880 --> 00:08:14,080
Speaker 1: is that you have multiple workstations that all connect back

136
00:08:14,080 --> 00:08:17,760
Speaker 1: to the same computer. The workstations are all dummy terminals

137
00:08:17,840 --> 00:08:21,080
Speaker 1: or thin clients. The computer can only work on one

138
00:08:21,160 --> 00:08:23,360
Speaker 1: job at a time, but it can do so at

139
00:08:23,360 --> 00:08:26,480
Speaker 1: a pretty darn fast pace, and so as you run

140
00:08:26,480 --> 00:08:29,040
Speaker 1: a job at your workstation, the computer waits for a

141
00:08:29,080 --> 00:08:32,040
Speaker 1: break and its processing of other jobs, then slots your

142
00:08:32,120 --> 00:08:35,160
Speaker 1: job in that break and runs your job to you.

143
00:08:35,320 --> 00:08:37,439
Speaker 1: It's like you've got the full attention of the computer,

144
00:08:37,559 --> 00:08:40,400
Speaker 1: but in reality, the computer is actually switching back and

145
00:08:40,440 --> 00:08:44,280
Speaker 1: forth between users, finishing each job quickly before moving on

146
00:08:44,320 --> 00:08:46,920
Speaker 1: to the next one, rather than doing all the jobs

147
00:08:46,920 --> 00:08:49,480
Speaker 1: at the same time. The benefit of time sharing is

148
00:08:49,480 --> 00:08:52,320
Speaker 1: that it's as if you've increased the number of computers

149
00:08:52,360 --> 00:08:55,480
Speaker 1: by the number of workstations, and so more people can

150
00:08:55,520 --> 00:08:58,480
Speaker 1: take advantage of the computer than with older systems. You

151
00:08:58,559 --> 00:09:02,520
Speaker 1: reduce downtime, you increase efficiency, and more people can actually

152
00:09:02,520 --> 00:09:07,400
Speaker 1: get stuff done. IBM was initially not interested in working

153
00:09:07,400 --> 00:09:10,000
Speaker 1: with m I T. The company wasn't convinced that a

154
00:09:10,080 --> 00:09:12,520
Speaker 1: multi user computer was going to be that big of

155
00:09:12,559 --> 00:09:15,320
Speaker 1: a deal and that most of their customers would have

156
00:09:15,400 --> 00:09:18,840
Speaker 1: no need for such a computer. So rather than dedicate

157
00:09:18,880 --> 00:09:21,480
Speaker 1: the time and resources needed to develop something that the

158
00:09:21,520 --> 00:09:24,439
Speaker 1: company wasn't convinced would ever be an actual product, it

159
00:09:24,480 --> 00:09:27,080
Speaker 1: would just be a one off, they bowed out and

160
00:09:27,120 --> 00:09:30,040
Speaker 1: so ge became the vendor for the early days of

161
00:09:30,080 --> 00:09:33,720
Speaker 1: Project Mac. But over at IBM, minds were slowly changing.

162
00:09:33,800 --> 00:09:37,120
Speaker 1: The Project Mac contract was a big deal with a

163
00:09:37,120 --> 00:09:40,079
Speaker 1: lot of funding at that time, and then it became

164
00:09:40,120 --> 00:09:42,480
Speaker 1: known that Bell Labs was also looking out for a

165
00:09:42,520 --> 00:09:45,400
Speaker 1: system that would grant access to multiple users to a

166
00:09:45,400 --> 00:09:48,360
Speaker 1: computer at the same time, and this was enough for

167
00:09:48,400 --> 00:09:51,920
Speaker 1: the folks at IBM to say, huh, maybe we were wrong,

168
00:09:52,240 --> 00:09:54,040
Speaker 1: and they changed their minds and they put some work

169
00:09:54,040 --> 00:09:57,840
Speaker 1: into developing such a system. They created a prototype called

170
00:09:58,040 --> 00:10:03,120
Speaker 1: the c P forty mainframe computer. The computer never became

171
00:10:03,120 --> 00:10:06,720
Speaker 1: a product sold by IBM, but was used internally in

172
00:10:06,800 --> 00:10:10,360
Speaker 1: IBM S labs. But the CP forty will be really

173
00:10:10,400 --> 00:10:13,280
Speaker 1: important in our story because it was the starting point

174
00:10:13,360 --> 00:10:15,280
Speaker 1: for a journey that would take us to the first

175
00:10:15,320 --> 00:10:20,360
Speaker 1: commercial mainframe computer system that could support virtualization. The time

176
00:10:20,400 --> 00:10:25,000
Speaker 1: sharing approach involved sharing parts of a computer's processing capabilities,

177
00:10:25,280 --> 00:10:29,280
Speaker 1: such as its system memory or its storage, and sharing

178
00:10:29,320 --> 00:10:32,240
Speaker 1: that with each user. But this was not just useful,

179
00:10:32,280 --> 00:10:35,160
Speaker 1: it was also limiting. If something went wrong for one user,

180
00:10:35,600 --> 00:10:39,360
Speaker 1: it would affect everyone on that computer. Creating more meaningful

181
00:10:39,400 --> 00:10:43,360
Speaker 1: partitions that would isolate each user would be more useful. Now.

182
00:10:43,400 --> 00:10:47,839
Speaker 1: Typically we described this process as a hypervisor, or sometimes

183
00:10:47,880 --> 00:10:51,559
Speaker 1: as a control program. A hypervisor's job is to separate

184
00:10:51,600 --> 00:10:55,079
Speaker 1: the applications and operating system running on a computer from

185
00:10:55,120 --> 00:10:58,440
Speaker 1: the computer's actual hardware. There are a couple of different

186
00:10:58,440 --> 00:11:02,400
Speaker 1: types of hypervisors. Type one hypervisors are also known as

187
00:11:02,520 --> 00:11:06,400
Speaker 1: bare metal hypervisors because they exist in a layer that's

188
00:11:06,480 --> 00:11:09,600
Speaker 1: directly on top of a machine's hardware. Then you have

189
00:11:09,679 --> 00:11:13,880
Speaker 1: type two hypervisors that's a software layer that exists over

190
00:11:13,960 --> 00:11:18,200
Speaker 1: top the existing machines operating system, so there's an extra

191
00:11:18,360 --> 00:11:22,360
Speaker 1: layer of abstraction there. Now, through a hypervisor, a single

192
00:11:22,440 --> 00:11:26,760
Speaker 1: machine can host multiple virtual guest machines. This idea would

193
00:11:26,800 --> 00:11:30,880
Speaker 1: be further explored in the next phase, that of true virtualization,

194
00:11:31,320 --> 00:11:33,160
Speaker 1: which I'll get to in just a moment. But first

195
00:11:33,240 --> 00:11:43,800
Speaker 1: let's take a quick break. Okay, so we just got

196
00:11:43,800 --> 00:11:48,439
Speaker 1: into what you could think of as the prehistoric virtualization era,

197
00:11:48,679 --> 00:11:53,240
Speaker 1: with hypervisor technology just coming into development. IBM was at

198
00:11:53,240 --> 00:11:56,640
Speaker 1: the forefront of this research with the CP forty mainframe

199
00:11:56,720 --> 00:11:59,560
Speaker 1: system in the nineteen sixties. The system ran on a

200
00:11:59,600 --> 00:12:05,000
Speaker 1: special operating system called CP slash c MS, which originally

201
00:12:05,040 --> 00:12:10,240
Speaker 1: stood for Control Program slash Cambridge Monitor System, though later

202
00:12:10,280 --> 00:12:13,760
Speaker 1: CMS would change to mean Conversational Monitor System and I've

203
00:12:13,800 --> 00:12:16,200
Speaker 1: even seen a couple of other variants out there. The

204
00:12:16,240 --> 00:12:20,280
Speaker 1: control program part was what allowed for this early virtualization

205
00:12:20,440 --> 00:12:24,760
Speaker 1: in the CP forty. It effectively created a full virtualization

206
00:12:24,920 --> 00:12:28,200
Speaker 1: of the underlying hardware of the CP forty for each

207
00:12:28,400 --> 00:12:32,560
Speaker 1: virtual machine, so people on dumb terminals terminals that did

208
00:12:32,600 --> 00:12:36,319
Speaker 1: not have their own solely dedicated computer could have a

209
00:12:36,440 --> 00:12:40,280
Speaker 1: virtual computer at their disposal. All the work was actually

210
00:12:40,360 --> 00:12:43,719
Speaker 1: being done on top of the single CP forty prototype,

211
00:12:43,960 --> 00:12:45,800
Speaker 1: but to each user it was as if they had

212
00:12:45,880 --> 00:12:51,239
Speaker 1: access to their own individual computer. The virtualization software partitioned

213
00:12:51,320 --> 00:12:55,120
Speaker 1: computer assets to each user. This wasn't just a matter

214
00:12:55,240 --> 00:12:59,280
Speaker 1: of convenience. It also allowed for rapid innovation. Programmers could

215
00:12:59,320 --> 00:13:03,320
Speaker 1: work on their own virtual machines, creating applications without having

216
00:13:03,320 --> 00:13:07,120
Speaker 1: to worry about a bug affecting everyone else. The partitions

217
00:13:07,160 --> 00:13:10,880
Speaker 1: provided protection, so if you were plugging away on a

218
00:13:10,960 --> 00:13:14,679
Speaker 1: difficult section of code and your coworker fouled something up

219
00:13:14,679 --> 00:13:17,320
Speaker 1: on their virtual machine, you didn't have to worry about

220
00:13:17,400 --> 00:13:20,840
Speaker 1: their mistake bringing you down with them. Each virtual machine

221
00:13:21,080 --> 00:13:25,640
Speaker 1: acted like its own individual physical computer. You can see

222
00:13:25,679 --> 00:13:29,080
Speaker 1: some similarities between this model of computing and what would

223
00:13:29,120 --> 00:13:32,200
Speaker 1: come much later in the era of cloud computing and

224
00:13:32,240 --> 00:13:36,800
Speaker 1: the brief age of netbooks. Technically, netbooks are still around,

225
00:13:37,000 --> 00:13:41,760
Speaker 1: but they are no longer the heyday type device that

226
00:13:41,800 --> 00:13:44,920
Speaker 1: they were for a very short while. The netbook essentially

227
00:13:44,960 --> 00:13:47,360
Speaker 1: had a five year run from two thousand seven to

228
00:13:47,400 --> 00:13:51,520
Speaker 1: two thousand twelve, and it's a category of lightweight laptop

229
00:13:51,559 --> 00:13:55,800
Speaker 1: computers that had modest specifications because the real purpose of

230
00:13:55,800 --> 00:13:58,760
Speaker 1: the computer was to use cloud based services. You didn't

231
00:13:58,800 --> 00:14:01,960
Speaker 1: have to have a b E fee computer because most

232
00:14:01,960 --> 00:14:05,320
Speaker 1: of the processing was taking place somewhere else. The netbooks

233
00:14:05,320 --> 00:14:07,920
Speaker 1: were a type of thin client, kind of like those

234
00:14:08,000 --> 00:14:11,920
Speaker 1: dummy terminals used in time sharing mainframe systems. You didn't

235
00:14:11,960 --> 00:14:14,000
Speaker 1: have to have a lot of horsepower. You just needed

236
00:14:14,040 --> 00:14:18,719
Speaker 1: input devices like a keyboard and output devices like a display,

237
00:14:18,760 --> 00:14:21,520
Speaker 1: and the actual computing was happening on some other machine

238
00:14:21,560 --> 00:14:24,080
Speaker 1: out on the internet. We'll talk about that more a

239
00:14:24,120 --> 00:14:27,920
Speaker 1: bit later in this episode. So while work on and

240
00:14:28,040 --> 00:14:31,640
Speaker 1: with the CP forty system was going, IBM researchers were

241
00:14:31,680 --> 00:14:34,840
Speaker 1: looking into other breakthroughs that would become important for computers

242
00:14:34,840 --> 00:14:39,120
Speaker 1: in general and virtualization in particular. In nineteen sixty five,

243
00:14:39,200 --> 00:14:42,240
Speaker 1: IBM announced it was developing a thirty two bit central

244
00:14:42,280 --> 00:14:46,360
Speaker 1: processing unit that included virtual memory hardware in it. This

245
00:14:46,480 --> 00:14:50,920
Speaker 1: was called the IBM System SLASH three sixty. IBM researchers

246
00:14:51,000 --> 00:14:54,560
Speaker 1: used what they learned developing the CP forty to create

247
00:14:54,640 --> 00:14:58,000
Speaker 1: the CP sixty seven, which was built on this IBM

248
00:14:58,040 --> 00:15:02,560
Speaker 1: system Slash three sixty DASH sixties seven hardware. I love

249
00:15:02,640 --> 00:15:06,200
Speaker 1: these names, by the way, there's so much fun to say. Anyway,

250
00:15:06,360 --> 00:15:09,040
Speaker 1: this would be the first computer to support virtualization that

251
00:15:09,040 --> 00:15:12,880
Speaker 1: would actually be widely available as a commercial product, not

252
00:15:12,960 --> 00:15:16,880
Speaker 1: just a prototype. This was an important component for mainframe users,

253
00:15:17,160 --> 00:15:20,760
Speaker 1: but it wouldn't really transfer to micro or mini computers.

254
00:15:20,960 --> 00:15:23,000
Speaker 1: So this is still in the main frame age where

255
00:15:23,000 --> 00:15:26,480
Speaker 1: you have the big centralized computer, and it's not that

256
00:15:26,560 --> 00:15:29,240
Speaker 1: big of a surprise. You know, a mainframe is a

257
00:15:29,280 --> 00:15:33,960
Speaker 1: centralized machine, but later computers didn't necessarily follow that model. Instead,

258
00:15:34,200 --> 00:15:36,640
Speaker 1: you could have a computer at your own desk. The

259
00:15:36,680 --> 00:15:39,800
Speaker 1: need to partition and separate your machine from everyone else

260
00:15:40,200 --> 00:15:43,200
Speaker 1: wasn't as pressing since you were working on an actual

261
00:15:43,320 --> 00:15:47,800
Speaker 1: separate piece of hardware rather than sharing a centralized computer

262
00:15:47,920 --> 00:15:51,440
Speaker 1: amongst everybody else, and the hardware you were using was

263
00:15:51,480 --> 00:15:54,000
Speaker 1: capable enough to do whatever it was you were doing.

264
00:15:54,400 --> 00:15:57,800
Speaker 1: Majorization reached the point where it was possible to produce

265
00:15:57,840 --> 00:16:01,760
Speaker 1: a relatively small computer, especially compared to the giant mainframes

266
00:16:01,840 --> 00:16:04,480
Speaker 1: of the nineteen fifties and nineteen sixties, and it could

267
00:16:04,520 --> 00:16:07,440
Speaker 1: handle all the tasks that your average user could throw

268
00:16:07,480 --> 00:16:09,760
Speaker 1: at it. So we started to see a shift away

269
00:16:09,880 --> 00:16:13,480
Speaker 1: from this centralized model of the mainframe strategy to a

270
00:16:13,680 --> 00:16:17,000
Speaker 1: decentralized approach in which everyone would work on their own

271
00:16:17,080 --> 00:16:20,800
Speaker 1: separate machine. This would continue and accelerate in the mid

272
00:16:20,840 --> 00:16:24,440
Speaker 1: to late nineteen seventies when the personal computer emerged and

273
00:16:24,480 --> 00:16:27,400
Speaker 1: the general public began to dip their proverbial toe in

274
00:16:27,440 --> 00:16:31,600
Speaker 1: the computational pool or something. Now that's not to say

275
00:16:31,640 --> 00:16:35,880
Speaker 1: that all work on virtualization stopped completely in the nineteen seventies.

276
00:16:36,120 --> 00:16:38,960
Speaker 1: It certainly didn't speed up, and it pretty much remained

277
00:16:39,000 --> 00:16:43,480
Speaker 1: in the domain of mainframe computers, which still had their uses,

278
00:16:43,520 --> 00:16:47,760
Speaker 1: but in niche cases, in fact increasingly niche cases. So

279
00:16:47,840 --> 00:16:50,960
Speaker 1: whether they were legacy systems that companies depended upon to

280
00:16:51,040 --> 00:16:54,600
Speaker 1: keep business going, or they were more powerful machines that

281
00:16:54,640 --> 00:16:57,560
Speaker 1: could be realized in the smaller formats. You know, they

282
00:16:57,560 --> 00:17:00,400
Speaker 1: were these powerful machines that were being used are very

283
00:17:00,440 --> 00:17:04,159
Speaker 1: specific tasks. They still had a place, it was just

284
00:17:04,320 --> 00:17:08,400
Speaker 1: a reduced place, and that was pretty much where virtualization

285
00:17:08,480 --> 00:17:12,439
Speaker 1: was stuck until the mid nineteen eighties. That's when a

286
00:17:12,440 --> 00:17:17,280
Speaker 1: company called Locust Computing Corporation developed special software in collaboration

287
00:17:17,320 --> 00:17:20,000
Speaker 1: with A T and T for a computer called the

288
00:17:20,080 --> 00:17:24,399
Speaker 1: A T and T sixty three hundred plus. The computers

289
00:17:24,440 --> 00:17:28,560
Speaker 1: operating system was Unix s v R two. Now, the

290
00:17:28,600 --> 00:17:31,800
Speaker 1: thing about Unix is. It's an operating system that can

291
00:17:31,800 --> 00:17:36,520
Speaker 1: support Unix programs, but it couldn't run DOSS based programs

292
00:17:36,560 --> 00:17:38,600
Speaker 1: on its own. And for those of you who aren't

293
00:17:38,600 --> 00:17:42,320
Speaker 1: familiar with DOSS, it was the text based operating system

294
00:17:42,400 --> 00:17:45,399
Speaker 1: used by many computers. There were lots of different flavors

295
00:17:45,440 --> 00:17:49,200
Speaker 1: of DOSS um Apple had its own sort of version,

296
00:17:49,280 --> 00:17:52,359
Speaker 1: but the one that everyone really knew was IMMAs DOSS,

297
00:17:52,480 --> 00:17:55,920
Speaker 1: the Microsoft flavor of DOSS that was the most popular

298
00:17:56,040 --> 00:17:58,719
Speaker 1: version of the text based operating system out there for

299
00:17:58,760 --> 00:18:02,399
Speaker 1: IBM and IBM comp udible machines. So there are a

300
00:18:02,400 --> 00:18:06,880
Speaker 1: lot of programs developed for MS DOSS and DAWs in general,

301
00:18:07,240 --> 00:18:10,120
Speaker 1: but you couldn't run them on a Unix based device.

302
00:18:10,520 --> 00:18:14,119
Speaker 1: So Locus developed some software that would handle the interaction

303
00:18:14,480 --> 00:18:18,840
Speaker 1: between DOSS programs designed for the eight eight six instruction

304
00:18:18,920 --> 00:18:24,280
Speaker 1: set that DOSS programs were reliant upon and the underlying

305
00:18:24,400 --> 00:18:27,560
Speaker 1: Unix operating system, so it's kind of like a liaison

306
00:18:27,640 --> 00:18:30,879
Speaker 1: between those. The software have become known as MERGE and

307
00:18:30,920 --> 00:18:33,960
Speaker 1: would be one of the first virtual machine managing or

308
00:18:34,080 --> 00:18:37,800
Speaker 1: VMM products on the market. So now you could have

309
00:18:37,840 --> 00:18:40,560
Speaker 1: a Unix machine and still run DOSS programs on top

310
00:18:40,600 --> 00:18:44,320
Speaker 1: of it. By having this virtual DOSS machine running on

311
00:18:44,400 --> 00:18:47,840
Speaker 1: the Unix hardware. This would mark a new era in

312
00:18:47,880 --> 00:18:51,119
Speaker 1: which companies would make software that would allow computers running

313
00:18:51,160 --> 00:18:54,120
Speaker 1: on one type of operating system in hardware to run

314
00:18:54,160 --> 00:18:57,600
Speaker 1: an instance of a different operating system, and it opened

315
00:18:57,680 --> 00:19:00,879
Speaker 1: up access to other types of programs. Most of us

316
00:19:01,359 --> 00:19:05,040
Speaker 1: don't tend to work on Unix systems. But what about

317
00:19:05,119 --> 00:19:09,040
Speaker 1: max versus PCs. You know that a Mac computer can't

318
00:19:09,160 --> 00:19:12,800
Speaker 1: run a DOSS or later on a Windows program, and

319
00:19:12,920 --> 00:19:17,560
Speaker 1: vice versa Windows computer can't run a Mac program. But

320
00:19:17,640 --> 00:19:21,600
Speaker 1: developers created virtual machine software that would allow a Macintosh

321
00:19:21,600 --> 00:19:25,600
Speaker 1: computer to run a virtual instance of DOSS on a Mac,

322
00:19:25,880 --> 00:19:29,080
Speaker 1: meaning you could access those DOSS programs on a Macintosh

323
00:19:29,240 --> 00:19:32,400
Speaker 1: while running this virtual machine software. Suddenly you could take

324
00:19:32,400 --> 00:19:35,959
Speaker 1: advantage of programs meant for another type of computer on

325
00:19:36,000 --> 00:19:39,439
Speaker 1: your own machine. By the way, it probably comes as

326
00:19:39,520 --> 00:19:41,760
Speaker 1: very little surprised to those of you who are familiar

327
00:19:41,800 --> 00:19:44,840
Speaker 1: with Apple that the company was not a fan of

328
00:19:44,880 --> 00:19:49,240
Speaker 1: anyone running the Mac operating system on a non Apple machine.

329
00:19:49,920 --> 00:19:54,119
Speaker 1: Through virtualization. This was technically possible, but Apple maintained that

330
00:19:54,160 --> 00:19:56,960
Speaker 1: the only legal way to do it was to run

331
00:19:57,080 --> 00:20:00,840
Speaker 1: mac os on a virtual platform on top of another

332
00:20:01,000 --> 00:20:04,440
Speaker 1: Apple branded computer. So why would you want to run

333
00:20:04,520 --> 00:20:07,320
Speaker 1: a virtual version of mac OS on top of a

334
00:20:07,359 --> 00:20:11,520
Speaker 1: computer already running Mac OS. One reason would be to

335
00:20:11,600 --> 00:20:15,399
Speaker 1: test a new program against multiple versions of a single

336
00:20:15,520 --> 00:20:18,640
Speaker 1: operating system. So you might want to run your new

337
00:20:18,720 --> 00:20:22,439
Speaker 1: mac program against the latest version of Mac OS and

338
00:20:22,440 --> 00:20:24,800
Speaker 1: then the previous versions of Mac OS to see if

339
00:20:24,840 --> 00:20:28,000
Speaker 1: it's still compatible, making sure you have backwards compatibility written

340
00:20:28,000 --> 00:20:31,280
Speaker 1: in there. That would be handy to know. For about

341
00:20:31,280 --> 00:20:34,680
Speaker 1: a decade, that was the state of virtualization. Companies made

342
00:20:34,680 --> 00:20:38,159
Speaker 1: programs that would allow users to run one operating system

343
00:20:38,240 --> 00:20:41,399
Speaker 1: on top of a machine running a different operating system.

344
00:20:41,400 --> 00:20:44,119
Speaker 1: It was useful for people who were developing software for

345
00:20:44,160 --> 00:20:47,960
Speaker 1: other platforms, but beyond that, there wasn't much general use

346
00:20:48,040 --> 00:20:50,800
Speaker 1: for it, because, again, everyone was relying on their own

347
00:20:50,800 --> 00:20:53,960
Speaker 1: individual computers anyway, you didn't have to worry about creating

348
00:20:53,960 --> 00:20:57,160
Speaker 1: partitions between users for the most part. There were other

349
00:20:57,240 --> 00:21:00,119
Speaker 1: manifestations of virtualization that would come on the scene a

350
00:21:00,160 --> 00:21:03,359
Speaker 1: little bit later. In the early nineteen nineties, a team

351
00:21:03,400 --> 00:21:06,640
Speaker 1: at Sun Microsystems was hard at work developing a new

352
00:21:06,680 --> 00:21:11,600
Speaker 1: programming language that would eventually take on the name Java. Initially,

353
00:21:12,000 --> 00:21:15,000
Speaker 1: the idea was that this programming language would allow developers

354
00:21:15,040 --> 00:21:19,240
Speaker 1: to create programs running on home appliances like television's. The

355
00:21:19,280 --> 00:21:21,800
Speaker 1: need for a language that developers could use to create

356
00:21:21,840 --> 00:21:25,960
Speaker 1: programs for different platforms was obvious because these appliances would

357
00:21:25,960 --> 00:21:28,760
Speaker 1: be coming from different manufacturers who would be working with

358
00:21:28,880 --> 00:21:32,760
Speaker 1: different microprocessor companies, so there was no guarantee that televisions

359
00:21:32,760 --> 00:21:36,280
Speaker 1: from two different companies would have similar microchips in them. Now,

360
00:21:36,320 --> 00:21:39,400
Speaker 1: this is a non trivial problem. If you are a programmer,

361
00:21:39,640 --> 00:21:42,080
Speaker 1: you have to make some practical decisions that have little

362
00:21:42,119 --> 00:21:44,879
Speaker 1: to do with the actual purpose of your application, and

363
00:21:44,920 --> 00:21:49,360
Speaker 1: one of those decisions is what platform will I develop for. Frequently,

364
00:21:49,400 --> 00:21:52,440
Speaker 1: the answer that many developers gravitate toward is I want

365
00:21:52,480 --> 00:21:55,680
Speaker 1: to develop for the most popular platform out there because

366
00:21:55,680 --> 00:21:59,000
Speaker 1: it represents the largest potential customer base. To put it

367
00:21:59,000 --> 00:22:01,520
Speaker 1: in another way, let's say you're making a video game,

368
00:22:01,840 --> 00:22:04,119
Speaker 1: and let's say there's only two consoles that are on

369
00:22:04,160 --> 00:22:08,119
Speaker 1: the market. One of those consoles has a market share

370
00:22:08,280 --> 00:22:11,000
Speaker 1: and the other has a ten market share, and they're

371
00:22:11,000 --> 00:22:14,000
Speaker 1: both great game consoles, but you're more likely to focus

372
00:22:14,040 --> 00:22:15,760
Speaker 1: on developing the game for the one that has the

373
00:22:16,680 --> 00:22:19,480
Speaker 1: market share because that's where most of the gamers are.

374
00:22:19,920 --> 00:22:22,320
Speaker 1: But one if you could create a programming language that

375
00:22:22,359 --> 00:22:27,440
Speaker 1: could work on different platforms despite the underlying hardware, That

376
00:22:27,520 --> 00:22:31,760
Speaker 1: was the idea behind Java. So programmer James Gosling and

377
00:22:31,880 --> 00:22:34,359
Speaker 1: his team set out to create a programming language that

378
00:22:34,400 --> 00:22:37,280
Speaker 1: could work on top of any device that was running

379
00:22:37,320 --> 00:22:41,200
Speaker 1: a Java virtual machine. While the resulting language wasn't used

380
00:22:41,200 --> 00:22:44,359
Speaker 1: in TVs at that time, it quickly became recognized as

381
00:22:44,359 --> 00:22:46,919
Speaker 1: a valuable tool for web development by the time the

382
00:22:47,040 --> 00:22:51,760
Speaker 1: Java Development Kit debuted in By then, the web was

383
00:22:51,800 --> 00:22:54,680
Speaker 1: really starting to take off, but it was also pretty limited.

384
00:22:54,720 --> 00:22:58,040
Speaker 1: All sorts of computers were acting as servers on the Internet,

385
00:22:58,240 --> 00:23:01,080
Speaker 1: running on different hardware and using for an operating systems,

386
00:23:01,359 --> 00:23:04,800
Speaker 1: and there was a need for more dynamic, interesting, rich

387
00:23:04,880 --> 00:23:08,320
Speaker 1: experiences online. But with everyone running different systems, it was

388
00:23:08,359 --> 00:23:11,520
Speaker 1: impossible to guarantee that a web app would work for everyone.

389
00:23:12,200 --> 00:23:15,880
Speaker 1: Java helped take that pressure off. It was a right, once,

390
00:23:16,080 --> 00:23:20,359
Speaker 1: run anywhere programming language, and so the language was adopted

391
00:23:20,400 --> 00:23:23,919
Speaker 1: widely on the web, with web browsers building Java applets

392
00:23:23,960 --> 00:23:27,920
Speaker 1: to support it within the browser itself. This virtualization made

393
00:23:28,040 --> 00:23:32,040
Speaker 1: rich Internet experiences possible. As it turns out the Internet

394
00:23:32,119 --> 00:23:35,000
Speaker 1: in general and the Web in particular would create the

395
00:23:35,160 --> 00:23:39,600
Speaker 1: perfect environment for the evolution of virtualization. This appeals to

396
00:23:39,680 --> 00:23:43,480
Speaker 1: common sense. After all, the Internet is a network of networks,

397
00:23:43,520 --> 00:23:47,119
Speaker 1: and each network can potentially include millions of computers running

398
00:23:47,119 --> 00:23:51,600
Speaker 1: on different operating systems and hardware. This makes common sense, right,

399
00:23:51,640 --> 00:23:54,119
Speaker 1: after all, being Internet is a network of networks, and

400
00:23:54,160 --> 00:23:57,240
Speaker 1: each network can potentially include millions of computers running on

401
00:23:57,240 --> 00:24:00,440
Speaker 1: different operating systems and hardware. When we come back, I'll

402
00:24:00,480 --> 00:24:03,440
Speaker 1: talk about how virtualization really took off in the late

403
00:24:03,520 --> 00:24:06,760
Speaker 1: nineties and what's been going on up to today. But

404
00:24:06,880 --> 00:24:16,840
Speaker 1: first let's take a quick break. In a way, you

405
00:24:16,920 --> 00:24:21,000
Speaker 1: could say that the Internet made the evolution of virtualization

406
00:24:21,080 --> 00:24:25,760
Speaker 1: not just possible but necessary. Typically, I T departments like

407
00:24:25,880 --> 00:24:29,600
Speaker 1: to dedicate servers to a specific task. They handle that

408
00:24:29,720 --> 00:24:34,080
Speaker 1: one task and nothing else. This dedication helps keep things

409
00:24:34,200 --> 00:24:37,560
Speaker 1: streamlined and reduces the chance that different tasks will start

410
00:24:37,560 --> 00:24:41,480
Speaker 1: to interfere with one another and cost stability issues. But

411
00:24:41,600 --> 00:24:45,120
Speaker 1: as the Internet was growing faster and faster, it became

412
00:24:45,480 --> 00:24:49,280
Speaker 1: clear that this particular strategy was going to be unsustainable.

413
00:24:49,400 --> 00:24:53,119
Speaker 1: It was just getting too big and too complex. Servers

414
00:24:53,280 --> 00:24:57,160
Speaker 1: are expensive and they're not just expensive to purchase, they're

415
00:24:57,160 --> 00:25:01,160
Speaker 1: also expensive to operate and maintain. If you're dedicating a

416
00:25:01,240 --> 00:25:05,240
Speaker 1: single server to every single task and you're adding to

417
00:25:05,320 --> 00:25:08,200
Speaker 1: the number of services you have on offer, your server

418
00:25:08,320 --> 00:25:11,720
Speaker 1: room is going to get really crowded really quickly. Those

419
00:25:11,760 --> 00:25:14,960
Speaker 1: machines take up physical space. Then you have other considerations

420
00:25:14,960 --> 00:25:17,240
Speaker 1: you have to take into account, such as having a

421
00:25:17,280 --> 00:25:21,480
Speaker 1: sufficient cooling system to keep everything operational, because computers don't

422
00:25:21,480 --> 00:25:24,479
Speaker 1: do well when they overheat, and as you add more machines,

423
00:25:24,520 --> 00:25:26,840
Speaker 1: you're generating more and more heat, so you've got to

424
00:25:26,920 --> 00:25:30,480
Speaker 1: cool them down more effectively. Then you have to figure

425
00:25:30,520 --> 00:25:34,720
Speaker 1: out how much electricity these things are gobbling up. It's

426
00:25:35,080 --> 00:25:38,320
Speaker 1: it's getting more expensive as you're adding more servers, and

427
00:25:38,320 --> 00:25:40,960
Speaker 1: and then there's the question of downtime. It's quite possible

428
00:25:41,000 --> 00:25:44,919
Speaker 1: that some servers will be in less demand than others.

429
00:25:45,200 --> 00:25:48,080
Speaker 1: So you might have one server operating at about fifteen

430
00:25:48,160 --> 00:25:51,560
Speaker 1: percent of its overall capacity and it's essentially idling the

431
00:25:51,600 --> 00:25:55,800
Speaker 1: other of the day. That's not an efficient use of resources.

432
00:25:56,040 --> 00:25:58,959
Speaker 1: If you've got fifty servers but all of them are

433
00:25:59,000 --> 00:26:02,080
Speaker 1: working at fifteen percent, you're saying, Wow, I'm I'm really

434
00:26:02,320 --> 00:26:07,360
Speaker 1: inefficient with how I'm using these machines. Virtualization would prove

435
00:26:07,440 --> 00:26:10,600
Speaker 1: to be a solution to this problem, but it wasn't

436
00:26:10,760 --> 00:26:16,359
Speaker 1: an easy solution. One of the big challenges facing developers

437
00:26:16,359 --> 00:26:19,840
Speaker 1: at that time was that the X eight six architecture

438
00:26:20,040 --> 00:26:24,600
Speaker 1: that many modern computers rely upon wasn't designed with virtualization

439
00:26:24,640 --> 00:26:28,199
Speaker 1: in mind, and there were some hurdles to overcome. And

440
00:26:28,240 --> 00:26:31,040
Speaker 1: you may have heard of X eight six and maybe

441
00:26:31,160 --> 00:26:33,720
Speaker 1: you wonder what the heck does that actually mean. It's

442
00:26:33,720 --> 00:26:38,560
Speaker 1: a reference to Intel's eight eighty six microprocessor, which originally

443
00:26:38,600 --> 00:26:43,280
Speaker 1: debuted in ninety eight. The eight eight six architecture is

444
00:26:43,320 --> 00:26:47,639
Speaker 1: the foundation for successive processors such as the two eighty six,

445
00:26:47,920 --> 00:26:50,160
Speaker 1: the e D three eighty six, and the eight four

446
00:26:50,200 --> 00:26:52,480
Speaker 1: eighty six. So if you ever heard someone talking about

447
00:26:52,480 --> 00:26:55,680
Speaker 1: like a three eight six or forty six computer, they

448
00:26:55,680 --> 00:26:58,959
Speaker 1: were actually talking about this particular architecture. It's a computer

449
00:26:59,040 --> 00:27:03,560
Speaker 1: that's based off this microprocessor architecture, which was largely designed

450
00:27:03,600 --> 00:27:06,919
Speaker 1: so that it could be backwards compatible while adding to

451
00:27:07,200 --> 00:27:11,639
Speaker 1: the various features and the processing speed generation to generation.

452
00:27:12,160 --> 00:27:16,600
Speaker 1: More importantly, it became the dominant architecture for computing platforms,

453
00:27:17,080 --> 00:27:22,520
Speaker 1: including Internet servers. But as I said that architecture was

454
00:27:22,640 --> 00:27:26,240
Speaker 1: not created with virtualization in mind, and there were several

455
00:27:26,359 --> 00:27:30,440
Speaker 1: key instructions that, when virtualized, would tend to cause an

456
00:27:30,600 --> 00:27:34,480
Speaker 1: X eight six system to crash. So while there was

457
00:27:34,520 --> 00:27:38,040
Speaker 1: a growing need for virtualization as the Internet was growing,

458
00:27:38,640 --> 00:27:42,960
Speaker 1: there was this challenge of implementing virtualization without actually making

459
00:27:43,000 --> 00:27:47,520
Speaker 1: everything crash all the time. Enter vm Ware. Now I'll

460
00:27:47,560 --> 00:27:50,239
Speaker 1: have to do a full episode on the company a

461
00:27:50,320 --> 00:27:55,760
Speaker 1: vm Ware someday. It's known for its virtualization software. In

462
00:27:55,800 --> 00:27:59,879
Speaker 1: two thousand one, vm Ware would introduce virtualization platforms for

463
00:28:00,000 --> 00:28:03,639
Speaker 1: Internet servers that handled the operations of virtualization in a

464
00:28:03,680 --> 00:28:07,000
Speaker 1: way that wouldn't trigger these system crashes. So now you

465
00:28:07,040 --> 00:28:11,840
Speaker 1: could run virtualization software on X eight six architecture systems

466
00:28:12,200 --> 00:28:16,240
Speaker 1: and not have to worry about them just completely crapping

467
00:28:16,240 --> 00:28:19,480
Speaker 1: out on you at a moment's notice. At first, the

468
00:28:19,480 --> 00:28:22,119
Speaker 1: products were actually slow to catch on It was not

469
00:28:22,240 --> 00:28:25,560
Speaker 1: yet apparent how useful they would be, but that would change,

470
00:28:25,720 --> 00:28:28,920
Speaker 1: and vm Ware's early entry into the space would mean

471
00:28:28,960 --> 00:28:31,720
Speaker 1: that the company would end up holding a dominant portion

472
00:28:32,160 --> 00:28:36,359
Speaker 1: of the virtualization market even as other software companies caught

473
00:28:36,440 --> 00:28:40,400
Speaker 1: onto it. Now, to avoid this episode just becoming a

474
00:28:40,480 --> 00:28:45,160
Speaker 1: list of release dates, for virtualization software. I'm gonna summarize

475
00:28:45,200 --> 00:28:48,360
Speaker 1: this to say that companies like Virtue Tech, A M

476
00:28:48,440 --> 00:28:54,200
Speaker 1: D Connectics, Virtus, Sun, and then later Oracle when it

477
00:28:54,240 --> 00:28:58,600
Speaker 1: would buy Sun Microsystems, also Microsoft itself, all of them

478
00:28:58,640 --> 00:29:03,240
Speaker 1: released various virtual realization solutions over the years. Some were

479
00:29:03,240 --> 00:29:06,240
Speaker 1: built to take advantage of changes in architecture, such as

480
00:29:06,400 --> 00:29:11,120
Speaker 1: sixty four bit instruction sets. But beyond these technical specifications,

481
00:29:11,160 --> 00:29:13,720
Speaker 1: there's not really that much to talk about. So I

482
00:29:13,720 --> 00:29:15,719
Speaker 1: think it's better to go back to a broad picture,

483
00:29:15,760 --> 00:29:19,520
Speaker 1: because otherwise, all I'm telling you is that the virtualization

484
00:29:19,600 --> 00:29:23,680
Speaker 1: software improved so that it could take advantage of improvements

485
00:29:23,920 --> 00:29:29,000
Speaker 1: in microprocessor design. That gets really boring really fast. So

486
00:29:29,440 --> 00:29:33,920
Speaker 1: how did virtualization actually help in the Internet age? Well,

487
00:29:33,920 --> 00:29:37,840
Speaker 1: by using this special software and I T admin could

488
00:29:37,840 --> 00:29:42,520
Speaker 1: take a single existing Internet server and then create multiple

489
00:29:42,640 --> 00:29:46,400
Speaker 1: virtual servers, and then each virtual server would perform as

490
00:29:46,440 --> 00:29:49,120
Speaker 1: if it was its own physical machine, similar to the

491
00:29:49,160 --> 00:29:52,880
Speaker 1: way previous implementations of virtualization would do. But you know,

492
00:29:52,880 --> 00:29:54,960
Speaker 1: in those cases you were talking about running a different

493
00:29:55,000 --> 00:29:57,760
Speaker 1: operating system on top of a machine or something like that.

494
00:29:58,840 --> 00:30:01,320
Speaker 1: But in this case, all of the virtual servers would

495
00:30:01,360 --> 00:30:04,480
Speaker 1: exist on top of a single device, and with careful planning,

496
00:30:04,680 --> 00:30:07,600
Speaker 1: you can build out virtual servers to make each physical

497
00:30:07,640 --> 00:30:12,320
Speaker 1: machine more efficient by run making it run closer to capacity. Right,

498
00:30:12,360 --> 00:30:15,600
Speaker 1: so instead of running at capacity, you might have it

499
00:30:15,720 --> 00:30:18,520
Speaker 1: running at seventy or eighty percent capacity. Now, you never

500
00:30:18,600 --> 00:30:22,000
Speaker 1: really want to run at full capacity because you could

501
00:30:22,000 --> 00:30:24,680
Speaker 1: have response time issues if they're if the demand is

502
00:30:25,000 --> 00:30:29,960
Speaker 1: exceeding the server's ability to serve up data, then you're

503
00:30:29,960 --> 00:30:33,600
Speaker 1: gonna have performance problems. But if you could get it

504
00:30:33,880 --> 00:30:37,560
Speaker 1: closer to capacity, then you have less downtime for your

505
00:30:37,600 --> 00:30:40,239
Speaker 1: machine and you're making better use of your equipment. And

506
00:30:40,280 --> 00:30:44,280
Speaker 1: with the right virtual machine management software, each individual virtual

507
00:30:44,360 --> 00:30:48,240
Speaker 1: server is totally partitioned from the others, which helps you

508
00:30:48,280 --> 00:30:52,640
Speaker 1: maintain security and stability. If one virtual server does go down,

509
00:30:52,720 --> 00:30:55,600
Speaker 1: the other virtual servers on top of that same physical

510
00:30:55,640 --> 00:30:59,520
Speaker 1: machine should be able to continue to operate without being affected.

511
00:30:59,640 --> 00:31:03,720
Speaker 1: So let's say you're a company and you have four

512
00:31:03,760 --> 00:31:07,520
Speaker 1: different apps. This is just a very basic example. So

513
00:31:07,600 --> 00:31:10,160
Speaker 1: you've got a server and you divided it into four

514
00:31:10,280 --> 00:31:13,080
Speaker 1: virtual servers, and then one of your virtual servers for

515
00:31:13,160 --> 00:31:15,880
Speaker 1: one of those apps fails, the other three apps should

516
00:31:15,880 --> 00:31:18,720
Speaker 1: still be able to operate just fine. Meanwhile, you've got

517
00:31:18,720 --> 00:31:22,280
Speaker 1: an I T admin frantically trying to fix the problem

518
00:31:22,320 --> 00:31:24,840
Speaker 1: on the virtual server of the affected app and get

519
00:31:24,880 --> 00:31:28,080
Speaker 1: it back online as fast as possible. But ideally you're

520
00:31:28,120 --> 00:31:30,640
Speaker 1: doing this in a way where your other operations aren't

521
00:31:30,680 --> 00:31:33,400
Speaker 1: being affected at all. So not only could you get

522
00:31:33,440 --> 00:31:36,200
Speaker 1: better use out of your hardware using virtualization, you could

523
00:31:36,200 --> 00:31:39,000
Speaker 1: also reduce the number of physical servers you had in

524
00:31:39,040 --> 00:31:42,600
Speaker 1: your data centers, and this would become increasingly important as

525
00:31:42,600 --> 00:31:45,480
Speaker 1: cloud services came on the scene and grew in popularity.

526
00:31:45,760 --> 00:31:49,600
Speaker 1: A large cloud service, whether it's offering an operational platform,

527
00:31:49,840 --> 00:31:53,400
Speaker 1: or storage space or something else, requires an awful lot

528
00:31:53,440 --> 00:31:56,560
Speaker 1: of hardware, but that requirement would be gargantuan if it

529
00:31:56,600 --> 00:32:00,480
Speaker 1: weren't for virtualization. For example, one of the many important

530
00:32:00,520 --> 00:32:05,680
Speaker 1: concepts in computing is redundancy. That is, making sure the

531
00:32:05,680 --> 00:32:08,440
Speaker 1: service you offer remains available even in the event of

532
00:32:08,480 --> 00:32:10,760
Speaker 1: a failure in the system. And this is true for

533
00:32:10,840 --> 00:32:13,920
Speaker 1: all sorts of services, but it's probably easiest to understand

534
00:32:13,960 --> 00:32:15,680
Speaker 1: if we take something that most of us have had

535
00:32:15,720 --> 00:32:19,000
Speaker 1: experience with. So let's talk about cloud storage. Let's say

536
00:32:19,000 --> 00:32:22,800
Speaker 1: you're using an online data storage service like Google Drive,

537
00:32:22,960 --> 00:32:26,200
Speaker 1: so you've stored some files on Google Drive. You've created

538
00:32:26,280 --> 00:32:30,840
Speaker 1: numerous documents, or you've uploaded files into your personal Google Drive,

539
00:32:30,920 --> 00:32:33,280
Speaker 1: and you expect to be able to access those files

540
00:32:33,880 --> 00:32:36,680
Speaker 1: whenever and wherever you need them, assuming you have some

541
00:32:36,720 --> 00:32:39,360
Speaker 1: sort of Internet connection, and you know that your files

542
00:32:39,400 --> 00:32:42,040
Speaker 1: are not living on your own personal computer or device.

543
00:32:42,360 --> 00:32:45,000
Speaker 1: It's living on the cloud, so you can use whatever

544
00:32:45,040 --> 00:32:48,320
Speaker 1: device you want to connect to those files. So the

545
00:32:48,320 --> 00:32:52,000
Speaker 1: files live somewhere on some server that is owned and

546
00:32:52,040 --> 00:32:55,440
Speaker 1: operated by Google, and that's partly true, but it's more

547
00:32:55,520 --> 00:32:59,560
Speaker 1: accurate to say that those files live on multiple servers

548
00:33:00,000 --> 00:33:03,840
Speaker 1: and and operated by Google, because hardware occasionally fails, and

549
00:33:03,840 --> 00:33:06,120
Speaker 1: if that were to happen to the machine that's holding

550
00:33:06,200 --> 00:33:09,080
Speaker 1: your files, you wouldn't be able to get to your stuff,

551
00:33:09,400 --> 00:33:11,160
Speaker 1: and that would be a problem for you and thus

552
00:33:11,200 --> 00:33:14,200
Speaker 1: a problem for Google. So to avoid that, Google has

553
00:33:14,280 --> 00:33:17,800
Speaker 1: copies of all of your files, and it's spread across

554
00:33:18,120 --> 00:33:22,560
Speaker 1: numerous servers to provide redundancy. If one server should fail,

555
00:33:22,880 --> 00:33:25,440
Speaker 1: Google can switch to a different one to serve up

556
00:33:25,480 --> 00:33:27,760
Speaker 1: your files to you when you request them, and you

557
00:33:27,800 --> 00:33:30,000
Speaker 1: get what you want and you're none the wiser that

558
00:33:30,120 --> 00:33:33,800
Speaker 1: anything is wrong on the back end. Redundancy is important

559
00:33:33,800 --> 00:33:37,600
Speaker 1: for any online service, and virtualization can help reduce the

560
00:33:37,680 --> 00:33:41,880
Speaker 1: physical hardware requirements for redundancy. So now, let's say you're

561
00:33:41,920 --> 00:33:46,120
Speaker 1: the one that's running Google Drive. It's it's under your direction,

562
00:33:46,760 --> 00:33:50,080
Speaker 1: and while you'd be using virtual servers, you would likely

563
00:33:50,080 --> 00:33:53,240
Speaker 1: still want to store the same information from the same

564
00:33:53,280 --> 00:33:57,240
Speaker 1: client on different physical machines. So let's say you've got

565
00:33:57,240 --> 00:34:00,480
Speaker 1: two physical servers. You've got server one and server two,

566
00:34:00,520 --> 00:34:04,160
Speaker 1: and each physical server has four virtual servers on it.

567
00:34:04,360 --> 00:34:07,720
Speaker 1: So Server one has virtual servers A, B, C, and D,

568
00:34:07,960 --> 00:34:11,680
Speaker 1: and server too has virtual servers E, F, G, n H. Now,

569
00:34:11,719 --> 00:34:16,000
Speaker 1: technically you could dedicate virtual servers A and B, both

570
00:34:16,040 --> 00:34:18,719
Speaker 1: of those living on the same physical machine, to act

571
00:34:18,719 --> 00:34:21,120
Speaker 1: as backups for each other. But if something were to

572
00:34:21,160 --> 00:34:23,960
Speaker 1: go wrong in one of the virtual environments, it wouldn't

573
00:34:23,960 --> 00:34:26,760
Speaker 1: affect the other one their partition from each other. However,

574
00:34:26,800 --> 00:34:29,840
Speaker 1: if something were to go wrong to the underlying physical hardware,

575
00:34:30,000 --> 00:34:33,480
Speaker 1: the actual machine that's running everything, you'd be out of

576
00:34:33,560 --> 00:34:37,160
Speaker 1: luck because you'd lose both your copies or you'd at

577
00:34:37,239 --> 00:34:39,520
Speaker 1: least not not be able to access them until you've

578
00:34:39,560 --> 00:34:42,560
Speaker 1: repaired whatever the problem was. So for that reason, you

579
00:34:42,600 --> 00:34:45,759
Speaker 1: want to spread the load around your physical machines, and

580
00:34:45,880 --> 00:34:48,200
Speaker 1: you want to keep everything running at a good efficiency,

581
00:34:48,480 --> 00:34:51,799
Speaker 1: so you want to use virtual servers to reduce downtime.

582
00:34:52,400 --> 00:34:55,040
Speaker 1: And of course virtualization is used for all types of

583
00:34:55,080 --> 00:34:59,759
Speaker 1: cloud services, including app development um. In fact, that's a

584
00:34:59,800 --> 00:35:02,600
Speaker 1: great use of virtualization as it gives developers the chance

585
00:35:02,680 --> 00:35:05,839
Speaker 1: to program in an environment where they are reassured they're

586
00:35:05,840 --> 00:35:08,320
Speaker 1: not gonna mess everything else up if something goes wrong.

587
00:35:08,760 --> 00:35:11,160
Speaker 1: And so there are companies out there that get massive

588
00:35:11,160 --> 00:35:14,480
Speaker 1: amounts of revenue from customers who are essentially renting out

589
00:35:14,520 --> 00:35:17,759
Speaker 1: a virtual server space to develop apps on top of

590
00:35:18,160 --> 00:35:21,520
Speaker 1: before they develop, develop and deploy those apps into the

591
00:35:21,680 --> 00:35:24,560
Speaker 1: into the wild to get them to actual customers. Now,

592
00:35:24,600 --> 00:35:27,719
Speaker 1: I've talked a lot about why virtualization is useful, but

593
00:35:27,800 --> 00:35:31,640
Speaker 1: it also presents challenges. It's not a perfect solution to

594
00:35:31,680 --> 00:35:35,600
Speaker 1: all problems. It also presents opportunities for problems to arise,

595
00:35:36,080 --> 00:35:39,279
Speaker 1: and one of those is in security. If someone is

596
00:35:39,280 --> 00:35:42,920
Speaker 1: able to get access to the underlying hardware, either physically

597
00:35:43,120 --> 00:35:47,800
Speaker 1: or remotely, then they could potentially affect all the virtual

598
00:35:47,920 --> 00:35:50,880
Speaker 1: servers that are running on that physical machine. So instead

599
00:35:50,880 --> 00:35:54,680
Speaker 1: of corrupting one service, a bad actor could have affect

600
00:35:54,719 --> 00:35:58,640
Speaker 1: possibly several services. So security is a big concern in

601
00:35:58,680 --> 00:36:02,840
Speaker 1: the virtualization world and in the information world in general.

602
00:36:03,440 --> 00:36:07,040
Speaker 1: Another challenge is that there's a lack of established standards

603
00:36:07,080 --> 00:36:10,200
Speaker 1: in virtualization, though there's movement in the space to change that,

604
00:36:10,680 --> 00:36:12,960
Speaker 1: and that means that a data center might use several

605
00:36:13,000 --> 00:36:16,120
Speaker 1: different virtualization strategies, which in turn means the I T

606
00:36:16,239 --> 00:36:18,600
Speaker 1: department has to be able to handle all that, and

607
00:36:18,640 --> 00:36:21,839
Speaker 1: this can get pretty tough as organizations and systems grow

608
00:36:21,880 --> 00:36:25,560
Speaker 1: more complicated. There's also the problem that as systems grow,

609
00:36:25,600 --> 00:36:28,239
Speaker 1: things can be forgotten. Just keeping track of all those

610
00:36:28,320 --> 00:36:31,120
Speaker 1: virtual machines and what's running on each of them can

611
00:36:31,160 --> 00:36:34,720
Speaker 1: become challenging in itself, and with a complicated enough system,

612
00:36:34,960 --> 00:36:38,279
Speaker 1: it's possible for virtual servers to go overlooked. And if

613
00:36:38,280 --> 00:36:40,760
Speaker 1: it's overlooked and no one is using it, that brings

614
00:36:40,760 --> 00:36:44,280
Speaker 1: down the operational efficiency of the physical machine that's running

615
00:36:44,280 --> 00:36:47,760
Speaker 1: that server, and it's hard to get around performance issues

616
00:36:47,800 --> 00:36:52,000
Speaker 1: as well. A dedicated physical server that is only running

617
00:36:52,000 --> 00:36:56,000
Speaker 1: one task can more efficiently handle that task than a

618
00:36:56,000 --> 00:36:58,560
Speaker 1: computer that has a virtual layer on top of the

619
00:36:58,600 --> 00:37:01,200
Speaker 1: physical layer. Think of it sort of like getting permission

620
00:37:01,200 --> 00:37:04,520
Speaker 1: from a boss in a smooth operation. You just go

621
00:37:04,600 --> 00:37:07,320
Speaker 1: to your boss and you ask for permission to do something,

622
00:37:07,360 --> 00:37:09,520
Speaker 1: and you get it or you don't get it. But

623
00:37:09,600 --> 00:37:13,240
Speaker 1: in a company that has, say a bloated middle management layer,

624
00:37:13,640 --> 00:37:15,960
Speaker 1: maybe first you have to go through a lower level

625
00:37:16,000 --> 00:37:18,840
Speaker 1: manager who doesn't have that authority, but you still have

626
00:37:18,880 --> 00:37:21,719
Speaker 1: to talk to them before you can get access to

627
00:37:21,760 --> 00:37:24,319
Speaker 1: the boss who actually does have the authority to say

628
00:37:24,440 --> 00:37:28,520
Speaker 1: yes or no to your request. It slows things down. Well,

629
00:37:28,600 --> 00:37:32,799
Speaker 1: virtual platforms can also slow things down a little bit,

630
00:37:33,200 --> 00:37:39,760
Speaker 1: so that's another potential drawback to virtualization. However, overall, virtualization

631
00:37:39,760 --> 00:37:43,560
Speaker 1: has been a huge reason why we've seen such explosive

632
00:37:43,640 --> 00:37:47,800
Speaker 1: growth in online services, and it continues to be important

633
00:37:47,840 --> 00:37:52,680
Speaker 1: for developers who are creating programs for multiple platforms. And

634
00:37:52,760 --> 00:37:56,040
Speaker 1: so I hope this episode has sort of given you

635
00:37:56,080 --> 00:37:59,239
Speaker 1: an idea of what virtualization is all about, where it

636
00:37:59,320 --> 00:38:04,520
Speaker 1: came from, and why it's important. Uh, it's definitely something

637
00:38:04,520 --> 00:38:07,920
Speaker 1: that has enabled us to have a world where we

638
00:38:08,000 --> 00:38:12,120
Speaker 1: have discussions about things like the Internet of Things. Without virtualization,

639
00:38:12,480 --> 00:38:16,200
Speaker 1: we wouldn't have the ability to support all those services

640
00:38:16,640 --> 00:38:18,960
Speaker 1: because you wouldn't have a data center large enough to

641
00:38:19,040 --> 00:38:22,400
Speaker 1: hold all the actual physical servers you would need to

642
00:38:22,480 --> 00:38:24,880
Speaker 1: dedicate to each and every task. It would be a

643
00:38:25,000 --> 00:38:30,640
Speaker 1: terrible waste of space, of energy, of resources, of money.

644
00:38:31,040 --> 00:38:35,880
Speaker 1: So virtualization has really helped create an efficient and cost

645
00:38:35,920 --> 00:38:42,719
Speaker 1: effective approach to deploying numerous services across the Internet. So

646
00:38:42,760 --> 00:38:45,080
Speaker 1: we'll probably see a lot more development in that space,

647
00:38:45,480 --> 00:38:49,560
Speaker 1: will see more competition in it, although the big players

648
00:38:49,640 --> 00:38:52,920
Speaker 1: will probably hold on to a pretty large market share.

649
00:38:53,360 --> 00:38:57,200
Speaker 1: They've got time on their side and their reputations are

650
00:38:57,200 --> 00:39:00,919
Speaker 1: built on that. So I'll probably do more episodes about

651
00:39:01,000 --> 00:39:03,720
Speaker 1: virtualization or the company is involved in it in the future.

652
00:39:04,040 --> 00:39:07,360
Speaker 1: In the meantime, thank you so much for this suggestion.

653
00:39:07,600 --> 00:39:10,799
Speaker 1: I greatly appreciate it. It's the sort of topic that

654
00:39:11,360 --> 00:39:12,960
Speaker 1: I used to cover all the time on tech Stuff,

655
00:39:12,960 --> 00:39:14,839
Speaker 1: but it's been a while since I've covered one like this,

656
00:39:15,440 --> 00:39:18,680
Speaker 1: So Martin, thank you very much. I hope you enjoyed

657
00:39:18,719 --> 00:39:20,960
Speaker 1: this episode, and I hope the rest of you were

658
00:39:21,000 --> 00:39:23,480
Speaker 1: able to learn something from it. If you guys have

659
00:39:23,560 --> 00:39:26,560
Speaker 1: suggestions for future episodes, you can send me an email.

660
00:39:26,640 --> 00:39:30,480
Speaker 1: The address is tech stuff at how stuff works dot com.

661
00:39:30,680 --> 00:39:34,719
Speaker 1: Drop on by our website that's tech stuff podcast dot com.

662
00:39:35,040 --> 00:39:36,960
Speaker 1: There you're gonna find an archive of all of our

663
00:39:36,960 --> 00:39:40,000
Speaker 1: past episodes, as well as links to things like our

664
00:39:40,040 --> 00:39:44,480
Speaker 1: social media presence and our online store, where every purchase

665
00:39:44,520 --> 00:39:46,560
Speaker 1: you make goes to help the show and we greatly

666
00:39:46,560 --> 00:39:50,479
Speaker 1: appreciate it, and I'll talk to you again really soon.

667
00:39:56,360 --> 00:39:58,560
Speaker 1: Text Stuff is a production of I heart Radio's How

668
00:39:58,640 --> 00:40:02,000
Speaker 1: Stuff Works. For more podcasts from I heart Radio, visit

669
00:40:02,040 --> 00:40:05,120
Speaker 1: the I heart Radio app, Apple Podcasts, or wherever you

670
00:40:05,200 --> 00:40:11,080
Speaker 1: listen to your favorite shows. H