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

2
00:00:07,320 --> 00:00:13,920
Speaker 1: How Stuff Works. Hey there, and welcome to tech Stuff.

3
00:00:13,960 --> 00:00:17,040
Speaker 1: I'm your host, Jonathan Strickland. I'm an executive producer with

4
00:00:17,079 --> 00:00:18,920
Speaker 1: How Stuff Works and I Heart Radio and a love

5
00:00:18,920 --> 00:00:22,280
Speaker 1: of all things tech and his time for another classic

6
00:00:22,440 --> 00:00:27,800
Speaker 1: tech Stuff episode. This episode originally published on July two

7
00:00:27,920 --> 00:00:32,839
Speaker 1: thousand twelve, just days after technology had been invented. This

8
00:00:32,920 --> 00:00:37,479
Speaker 1: particular episode covers solid state drives, something that is far

9
00:00:37,560 --> 00:00:40,640
Speaker 1: more common these days, was already, you know, fairly common

10
00:00:40,680 --> 00:00:43,479
Speaker 1: back in two thousand twelve, but now it's becoming a

11
00:00:43,560 --> 00:00:46,680
Speaker 1: standard feature and it's rare when you come across other

12
00:00:46,680 --> 00:00:49,560
Speaker 1: types of hard drives in a lot of different technologies.

13
00:00:49,840 --> 00:00:54,360
Speaker 1: So let's join past Jonathan Strickland and his co host

14
00:00:54,440 --> 00:00:58,560
Speaker 1: Chris Palette and learn about solid state drives. Hey there,

15
00:00:59,800 --> 00:01:02,640
Speaker 1: So today we wanted to talk a little bit about

16
00:01:03,000 --> 00:01:08,040
Speaker 1: solid state drives and what makes them work and how

17
00:01:08,080 --> 00:01:12,199
Speaker 1: they are different from the traditional forms of storage media

18
00:01:12,319 --> 00:01:15,520
Speaker 1: that we are used to in the world of computers. Now,

19
00:01:15,760 --> 00:01:17,520
Speaker 1: some of you out there may be used to solid

20
00:01:17,520 --> 00:01:19,520
Speaker 1: state drives, and so you're thinking how our solid state

21
00:01:19,600 --> 00:01:21,679
Speaker 1: drives different from other solid state drives. That's not what

22
00:01:21,720 --> 00:01:23,959
Speaker 1: I mean. I mean those of us who have used

23
00:01:24,800 --> 00:01:28,400
Speaker 1: hard disks that use magnetic storage in some form and

24
00:01:28,640 --> 00:01:33,319
Speaker 1: uh there are a lot of differences, mainly mechanical physical differences,

25
00:01:33,400 --> 00:01:36,119
Speaker 1: because when you're talking about a hard disk I gus

26
00:01:36,520 --> 00:01:40,120
Speaker 1: a traditional hard hard disk drive, you're talking about a

27
00:01:40,160 --> 00:01:44,960
Speaker 1: device that has moving parts and UM, it has platters

28
00:01:44,959 --> 00:01:49,080
Speaker 1: that are that have information stored on the magnetically use

29
00:01:49,160 --> 00:01:53,160
Speaker 1: magnetic fields to change uh informations into zeros and ones,

30
00:01:53,600 --> 00:01:56,560
Speaker 1: those bits that we use to create the data that

31
00:01:56,600 --> 00:02:01,560
Speaker 1: computers can understand. And you have a physical device that

32
00:02:01,680 --> 00:02:04,320
Speaker 1: reads that information off of the platters. The players have

33
00:02:04,360 --> 00:02:08,000
Speaker 1: to spin for this to happen. Uh. Turns out those

34
00:02:08,040 --> 00:02:11,480
Speaker 1: devices they take up they take a lot of time

35
00:02:11,880 --> 00:02:15,680
Speaker 1: to pull that information up relatively speaking. Right When I

36
00:02:15,680 --> 00:02:19,880
Speaker 1: say a lot of time, we're talking milliseconds. But still, yeah,

37
00:02:19,880 --> 00:02:24,320
Speaker 1: there are When you think about getting information into and

38
00:02:24,360 --> 00:02:28,840
Speaker 1: out of uh your computers basically what you're doing right now,

39
00:02:28,880 --> 00:02:32,840
Speaker 1: there there's several different ways it does that. There's uh,

40
00:02:32,919 --> 00:02:38,640
Speaker 1: some information that is available for you know, ready retrieval. UM.

41
00:02:38,680 --> 00:02:41,040
Speaker 1: You know in the cash that's built in you have

42
00:02:41,160 --> 00:02:44,160
Speaker 1: level one, level two cash. You may not have necessarily

43
00:02:44,200 --> 00:02:47,400
Speaker 1: known what that was. I didn't, uh, you know, and

44
00:02:48,320 --> 00:02:50,800
Speaker 1: it's sort of uh, it's sort of intuitive when you

45
00:02:50,800 --> 00:02:53,200
Speaker 1: think about it. It keeps it this information in this

46
00:02:53,320 --> 00:02:58,120
Speaker 1: cash on hand. So stuff that you're doing right now

47
00:02:58,800 --> 00:03:01,320
Speaker 1: is is kept right there close by in a very

48
00:03:01,400 --> 00:03:05,960
Speaker 1: fast um uh, in a very fast retrieval system, so

49
00:03:06,120 --> 00:03:08,960
Speaker 1: I can pull it back at at at a nanoseconds.

50
00:03:09,000 --> 00:03:11,720
Speaker 1: Notice right, in fact, if we if we go, it's

51
00:03:11,720 --> 00:03:15,080
Speaker 1: easy to imagine this if we think of the CPU first,

52
00:03:15,400 --> 00:03:17,840
Speaker 1: just look at the central processing of your computer and

53
00:03:17,880 --> 00:03:21,280
Speaker 1: think of that as this is the place where operations

54
00:03:21,320 --> 00:03:25,160
Speaker 1: are executed upon data. Right. It pulls data in, it

55
00:03:25,240 --> 00:03:27,799
Speaker 1: executes an operation on it, it gets a result. That's

56
00:03:27,840 --> 00:03:31,560
Speaker 1: the purpose of the CPU. Now, a CPU has something

57
00:03:31,600 --> 00:03:35,960
Speaker 1: called registers. Registers are where the CPU can hold data.

58
00:03:36,080 --> 00:03:40,440
Speaker 1: But registers hold a very small amount of data comparatively speaking,

59
00:03:41,160 --> 00:03:44,760
Speaker 1: usually just a few hundred bytes. You can build larger registers,

60
00:03:44,760 --> 00:03:47,440
Speaker 1: and you can build more registers for your CPU, but

61
00:03:47,520 --> 00:03:50,280
Speaker 1: that tends to be pretty expensive. Now, the benefit of

62
00:03:50,320 --> 00:03:53,800
Speaker 1: having information and registers on the CPU is that you

63
00:03:53,880 --> 00:03:57,640
Speaker 1: have next to no time at all between when you

64
00:03:57,680 --> 00:04:00,520
Speaker 1: pull the information and when you can execute an operation

65
00:04:00,640 --> 00:04:04,480
Speaker 1: upon that information. So that means that we say that

66
00:04:04,520 --> 00:04:09,920
Speaker 1: the latency for the information within a CPUs registers is zero.

67
00:04:10,360 --> 00:04:13,440
Speaker 1: Latency is that time between when you retrieve information and

68
00:04:13,480 --> 00:04:15,600
Speaker 1: when you can or when you send a request for

69
00:04:15,720 --> 00:04:18,560
Speaker 1: information and when you actually retrieve it. So there's no

70
00:04:18,720 --> 00:04:23,040
Speaker 1: latency when you're talking about a the CPUs registers. However,

71
00:04:23,480 --> 00:04:26,880
Speaker 1: only a few hundred bytes are exist within those registers.

72
00:04:26,880 --> 00:04:29,280
Speaker 1: It's not a lot of data. Now you were talking

73
00:04:29,279 --> 00:04:31,680
Speaker 1: about the cash is the level one cash is kind

74
00:04:31,680 --> 00:04:35,360
Speaker 1: of the information that the CPU uses frequently. It's going

75
00:04:35,400 --> 00:04:37,600
Speaker 1: to in whatever application you have to be in at

76
00:04:37,640 --> 00:04:40,160
Speaker 1: that time, right, So this is information that the CPU

77
00:04:40,240 --> 00:04:42,680
Speaker 1: is having to go to again and again, and it

78
00:04:42,720 --> 00:04:45,400
Speaker 1: needs it to be as close as possible so that

79
00:04:45,839 --> 00:04:49,080
Speaker 1: it keeps that latency down because of course, if you

80
00:04:49,120 --> 00:04:51,479
Speaker 1: have to go further out for your information, it's going

81
00:04:51,520 --> 00:04:53,720
Speaker 1: to take longer for it to get back. So at

82
00:04:53,800 --> 00:04:58,279
Speaker 1: level one cash, you're talking about around bytes of data

83
00:04:58,480 --> 00:05:04,880
Speaker 1: per CPU core. If you're running an ivy Bridge processor

84
00:05:04,920 --> 00:05:08,600
Speaker 1: and I seven core ivy Bridge processor from Mentel, so

85
00:05:08,680 --> 00:05:14,640
Speaker 1: that's three per UH core and as far as that

86
00:05:14,720 --> 00:05:18,440
Speaker 1: level one cash goes, and that takes a nanosecond to

87
00:05:18,520 --> 00:05:23,080
Speaker 1: pull that data up. So that's one billionth of a second. Yep.

88
00:05:23,440 --> 00:05:26,279
Speaker 1: And uh, you know, just on the just to comment

89
00:05:26,360 --> 00:05:30,880
Speaker 1: on the ivy bridge notation that the cash is themselves, Um,

90
00:05:30,920 --> 00:05:34,400
Speaker 1: you know they exist on all these different processors. That's

91
00:05:34,480 --> 00:05:38,559
Speaker 1: just that figure was just specific to the ivy bridge,

92
00:05:38,560 --> 00:05:41,080
Speaker 1: But other processors do have this as well. Yes, yes,

93
00:05:41,240 --> 00:05:43,680
Speaker 1: just in case there was any confusion. And then you

94
00:05:43,720 --> 00:05:46,400
Speaker 1: could have a level to cash just slightly further out.

95
00:05:46,680 --> 00:05:49,320
Speaker 1: You know, think of think of it like concentric circles, right,

96
00:05:49,680 --> 00:05:52,599
Speaker 1: So the level two cash is a concentric circle further

97
00:05:52,680 --> 00:05:55,560
Speaker 1: out from the CPU. It can hold a little more data.

98
00:05:55,640 --> 00:05:57,880
Speaker 1: It's a little slower to pull that information up, it's

99
00:05:57,880 --> 00:06:00,840
Speaker 1: a few nanoseconds. Then it may even have a level

100
00:06:00,880 --> 00:06:04,840
Speaker 1: three cash. Not all CPUs do, but many do, and

101
00:06:04,920 --> 00:06:07,520
Speaker 1: then that is even larger and holds even more data

102
00:06:07,839 --> 00:06:11,640
Speaker 1: and takes even longer again in relative terms, to get

103
00:06:11,640 --> 00:06:14,719
Speaker 1: that data to the CPU. And beyond that, that's when

104
00:06:14,720 --> 00:06:18,919
Speaker 1: you go to your computer's actual memory, because all the

105
00:06:18,920 --> 00:06:21,760
Speaker 1: stuff we've been talking about right now is all located

106
00:06:22,360 --> 00:06:25,680
Speaker 1: on the CPU die itself, so it's all part of

107
00:06:25,720 --> 00:06:29,320
Speaker 1: that that chip. Yeah, it's not you know, it's not separate.

108
00:06:29,320 --> 00:06:31,760
Speaker 1: It's not another element that's on the motherboard. This is

109
00:06:31,800 --> 00:06:34,039
Speaker 1: all part of the CPU. Yeah. If you if you

110
00:06:34,040 --> 00:06:36,359
Speaker 1: pull your hard drive out of the computer toss it

111
00:06:36,360 --> 00:06:39,520
Speaker 1: on the floor, don't don't do that. Um. And uh.

112
00:06:39,560 --> 00:06:42,080
Speaker 1: And you you pull the the RAM chips out of

113
00:06:42,120 --> 00:06:43,960
Speaker 1: your computer and toss them on the floor, don't do

114
00:06:44,000 --> 00:06:47,279
Speaker 1: that either. Um. Then I'm glad you did that, because

115
00:06:47,320 --> 00:06:49,760
Speaker 1: I was thinking it. It was just funny that you

116
00:06:49,760 --> 00:06:54,520
Speaker 1: said it. Um. Then the the registry and the cashes

117
00:06:54,960 --> 00:06:57,720
Speaker 1: will still be there on the computer and I won't

118
00:06:57,720 --> 00:07:01,400
Speaker 1: be able to do a whole lot with them. But still, um,

119
00:07:01,520 --> 00:07:04,479
Speaker 1: so you know that, just to illustrate that. So, yes,

120
00:07:04,560 --> 00:07:07,840
Speaker 1: the next part will pick this stuff gently back up

121
00:07:07,880 --> 00:07:10,280
Speaker 1: off the floor and put it inside. The RAM is

122
00:07:11,200 --> 00:07:13,720
Speaker 1: essentially the next ring out if you will, from the

123
00:07:13,760 --> 00:07:16,080
Speaker 1: cash is however many that you happen to have on

124
00:07:16,120 --> 00:07:20,040
Speaker 1: your CPU, right, And these these memory chips have been

125
00:07:20,080 --> 00:07:23,880
Speaker 1: optimized so that the latency is really really low. However,

126
00:07:24,640 --> 00:07:29,520
Speaker 1: they are not located on the CPU. They are connected

127
00:07:29,560 --> 00:07:34,600
Speaker 1: by circuits by by pathways to the CPU. Well, because

128
00:07:34,640 --> 00:07:36,840
Speaker 1: they are not located on the CPU, and because that

129
00:07:36,920 --> 00:07:40,360
Speaker 1: information does have to actually travel a physical distance, that

130
00:07:40,480 --> 00:07:44,600
Speaker 1: increases the latency time. So when your CPU has to

131
00:07:44,640 --> 00:07:47,960
Speaker 1: pull information out and that information is not in the

132
00:07:48,080 --> 00:07:50,880
Speaker 1: cash for the CPU, it has to but it resides

133
00:07:50,920 --> 00:07:54,520
Speaker 1: within the memory of your computer. Then it has to

134
00:07:54,560 --> 00:07:58,800
Speaker 1: travel this pathway and for the request and retrieval that

135
00:07:58,840 --> 00:08:01,800
Speaker 1: can take between forty an eighty nanoseconds. So we're still

136
00:08:01,840 --> 00:08:05,040
Speaker 1: talking a fraction of a second. Yeah, I mean this

137
00:08:05,120 --> 00:08:07,840
Speaker 1: is these are are times that you or I will

138
00:08:07,880 --> 00:08:10,400
Speaker 1: not even be able to notice. Yeah, we can't. We

139
00:08:10,520 --> 00:08:16,000
Speaker 1: we have no ability to to register that using our senses.

140
00:08:16,040 --> 00:08:21,520
Speaker 1: We would have to use incredibly sensitive measurement devices in

141
00:08:21,600 --> 00:08:24,520
Speaker 1: order to be able to tell the difference between forty

142
00:08:24,600 --> 00:08:27,880
Speaker 1: and eighty nano seconds. To us, there's no meaningful difference

143
00:08:27,880 --> 00:08:31,560
Speaker 1: at all. But the but the thing about RAM is

144
00:08:32,080 --> 00:08:35,480
Speaker 1: when you shut off the computer, all that information that's

145
00:08:35,520 --> 00:08:41,040
Speaker 1: in RAM. Yeah, it's it's held there by the electrical

146
00:08:41,120 --> 00:08:44,920
Speaker 1: charge UM, which is relative to the computer. Now you're

147
00:08:44,960 --> 00:08:48,199
Speaker 1: hard drive with the information stored we're talking traditional when

148
00:08:48,200 --> 00:08:52,040
Speaker 1: the information stored man magnetically on those platters. Um, it's

149
00:08:52,080 --> 00:08:53,880
Speaker 1: able to save that stuff so that when you turn

150
00:08:53,920 --> 00:08:56,440
Speaker 1: the computer back on, you can read the hard drive

151
00:08:56,480 --> 00:09:00,839
Speaker 1: and get it back. Um. The thing is that hard

152
00:09:00,920 --> 00:09:04,880
Speaker 1: drives have different rotating speeds UM typical to see a

153
00:09:04,960 --> 00:09:12,400
Speaker 1: laptop with a rpm uh drive or or even revolutions

154
00:09:12,440 --> 00:09:14,440
Speaker 1: per minute just in case you aren't familiar with the

155
00:09:14,559 --> 00:09:17,160
Speaker 1: term UM, and you're more likely to see fat those

156
00:09:17,200 --> 00:09:22,079
Speaker 1: and faster in desktop computers UM and the faster these rotate.

157
00:09:22,360 --> 00:09:25,320
Speaker 1: In general, that means the faster the information can be

158
00:09:25,400 --> 00:09:29,360
Speaker 1: pulled from the hard drive and sent to memory and

159
00:09:29,760 --> 00:09:33,200
Speaker 1: then onto your CPU. UM. So if you if you

160
00:09:33,240 --> 00:09:36,280
Speaker 1: took your hard drive. The thing is that these systems

161
00:09:36,320 --> 00:09:41,040
Speaker 1: are are delicate their their machine to very precise tolerances.

162
00:09:41,080 --> 00:09:44,559
Speaker 1: That the head UM that reads the disc, it looks

163
00:09:44,600 --> 00:09:46,720
Speaker 1: like it looks like a record player for those of

164
00:09:46,760 --> 00:09:49,880
Speaker 1: us who remember that, But the the head doesn't actually

165
00:09:50,040 --> 00:09:52,440
Speaker 1: touch the disc. If it does, that's what they call

166
00:09:52,760 --> 00:09:55,680
Speaker 1: a bad thing. Yeah, it looks like it's in contact

167
00:09:55,679 --> 00:09:58,480
Speaker 1: because it's so close to the platter, but in actuality

168
00:09:58,520 --> 00:10:02,560
Speaker 1: there is like a millimeter's difference between where it is

169
00:10:02,840 --> 00:10:05,920
Speaker 1: or even less. It's amazing, And that's the thing. If

170
00:10:06,000 --> 00:10:08,480
Speaker 1: you did take your if you did take your hard

171
00:10:08,559 --> 00:10:10,360
Speaker 1: drive and throw it on the floor, it is very

172
00:10:10,360 --> 00:10:15,400
Speaker 1: possible that the head crashed into the platters, which is

173
00:10:16,559 --> 00:10:19,720
Speaker 1: very bad right. If you ever hear a clicking noise

174
00:10:19,760 --> 00:10:22,200
Speaker 1: from your hard drive, that usually means that the platters

175
00:10:22,240 --> 00:10:24,480
Speaker 1: are out of alignment or that the head is actually

176
00:10:24,520 --> 00:10:28,160
Speaker 1: coming into contact, something is hitting against something else within

177
00:10:28,200 --> 00:10:31,680
Speaker 1: that physical mechanical device, and that means that it is

178
00:10:31,720 --> 00:10:35,400
Speaker 1: breaking down. That that also means you should take that

179
00:10:35,640 --> 00:10:38,640
Speaker 1: hard drive as soon as you can to a professional

180
00:10:38,679 --> 00:10:42,280
Speaker 1: who can pull the data off of the hard drive,

181
00:10:42,600 --> 00:10:44,679
Speaker 1: because the the hard drive itself may or may not

182
00:10:44,760 --> 00:10:47,440
Speaker 1: be uh you may or may not be able to

183
00:10:47,440 --> 00:10:49,760
Speaker 1: repair it. So you definitely want to be able to

184
00:10:49,760 --> 00:10:53,000
Speaker 1: retrieve the information. And the reason why we're even talking

185
00:10:53,040 --> 00:10:56,240
Speaker 1: about this physical device is because you may have guessed,

186
00:10:56,280 --> 00:10:59,480
Speaker 1: because you've got this mechanical element, it's going to take

187
00:10:59,559 --> 00:11:03,160
Speaker 1: a lot longer to retrieve that information. Comparatively speaking, we're

188
00:11:03,160 --> 00:11:06,920
Speaker 1: talking about milliseconds now as opposed to nanoseconds, and in

189
00:11:06,960 --> 00:11:10,199
Speaker 1: the world of computers, that's a long time. You know,

190
00:11:10,240 --> 00:11:12,800
Speaker 1: a you're talking about these other fractions of a second,

191
00:11:12,800 --> 00:11:15,520
Speaker 1: billions of a second, and then several or you know,

192
00:11:15,559 --> 00:11:18,640
Speaker 1: you go to a couple orders of magnitude up you

193
00:11:18,679 --> 00:11:20,679
Speaker 1: realize this is this is a lot longer, and it's

194
00:11:20,720 --> 00:11:24,040
Speaker 1: going to mean that in general, the operations that you

195
00:11:24,080 --> 00:11:26,600
Speaker 1: start to use on your computer are going to take

196
00:11:26,640 --> 00:11:28,920
Speaker 1: more and more time. Well, there are only so many

197
00:11:28,960 --> 00:11:32,840
Speaker 1: ways we can limit how much time it takes to

198
00:11:32,880 --> 00:11:36,479
Speaker 1: retrieve information from a hard drive. Some of that includes

199
00:11:37,080 --> 00:11:41,720
Speaker 1: creating better interfaces, which is when we went from the

200
00:11:42,000 --> 00:11:45,959
Speaker 1: two SATA interfaces s A T A interfaces. Uh, that

201
00:11:46,040 --> 00:11:48,040
Speaker 1: was actually a big improvement. It meant that we could

202
00:11:48,040 --> 00:11:52,920
Speaker 1: move data much more quickly from the hard drive into RAM.

203
00:11:53,040 --> 00:11:57,360
Speaker 1: But there's only so so much faster you can go

204
00:11:57,800 --> 00:12:02,400
Speaker 1: without really turning up that RPM speed to ludicrous amounts.

205
00:12:02,440 --> 00:12:04,160
Speaker 1: And of course the faster it goes, the more likely

206
00:12:04,240 --> 00:12:07,520
Speaker 1: you have mechanical problems down the line. I mean we're

207
00:12:07,520 --> 00:12:10,800
Speaker 1: in tear and things of that nature. So, Um, you

208
00:12:10,880 --> 00:12:14,480
Speaker 1: have the incredibly reliable hard drive. I mean we've been

209
00:12:14,559 --> 00:12:17,440
Speaker 1: using these things for years and years now. Um, they've

210
00:12:17,480 --> 00:12:20,840
Speaker 1: gone up to very large sizes. Uh. And and they're

211
00:12:20,880 --> 00:12:24,200
Speaker 1: they're fairly cheap compared to the way they were just

212
00:12:24,440 --> 00:12:28,000
Speaker 1: a few years ago. But they have they have their

213
00:12:28,000 --> 00:12:30,559
Speaker 1: own problems. I mean, they're they're delicate. You can't necessarily

214
00:12:30,559 --> 00:12:36,320
Speaker 1: take them everywhere, um, you know, and expect them to operate. Uh. Flawlessly.

215
00:12:36,880 --> 00:12:39,720
Speaker 1: Um and uh you know, as you pointed out, they're

216
00:12:39,880 --> 00:12:45,040
Speaker 1: they're only so fast. So uh, flash memory in general

217
00:12:45,440 --> 00:12:48,920
Speaker 1: is you know, an alternative, a very pleasant alternative. It

218
00:12:49,000 --> 00:12:53,240
Speaker 1: works in our in our smartphones and our music players. UM,

219
00:12:53,400 --> 00:12:56,640
Speaker 1: it works in in memory sticks, I mean d some drives.

220
00:12:56,800 --> 00:12:58,720
Speaker 1: You know, kids of all ages now take them to

221
00:12:58,760 --> 00:13:00,800
Speaker 1: school with them because you know that you can keep

222
00:13:01,280 --> 00:13:05,000
Speaker 1: um I remember one gigabyte hard drives that were huge,

223
00:13:05,120 --> 00:13:08,160
Speaker 1: and now you can keep sixteen gigs on a tiny

224
00:13:08,200 --> 00:13:11,240
Speaker 1: thumb drive that cost you know, a very tiny fraction

225
00:13:11,280 --> 00:13:14,320
Speaker 1: of the price. Spies, Spies use them to uh to

226
00:13:14,640 --> 00:13:20,240
Speaker 1: put malware onto secure systems. Because both both stucks net

227
00:13:20,360 --> 00:13:25,000
Speaker 1: and Flame appear to have been injected into target computers

228
00:13:25,120 --> 00:13:29,520
Speaker 1: using UH and an offsite sort of storage device, So

229
00:13:29,559 --> 00:13:33,040
Speaker 1: some sort of well offsite from the computer system. UH

230
00:13:33,080 --> 00:13:35,720
Speaker 1: so something like a thumb drive. So you can imagine

231
00:13:35,760 --> 00:13:38,240
Speaker 1: there's a guy who might have paid a little visit

232
00:13:38,280 --> 00:13:43,400
Speaker 1: to an Iranian UH uranium enrichment plant and happened to

233
00:13:43,440 --> 00:13:45,640
Speaker 1: have this thumb drive, plug it into a computer system

234
00:13:45,640 --> 00:13:48,679
Speaker 1: and infected it that way. That's just one potential way

235
00:13:48,720 --> 00:13:51,960
Speaker 1: that scenario could have unfolded. But yeah, I mean, these

236
00:13:52,000 --> 00:13:57,120
Speaker 1: things have have become ubiquitous in all areas of computing,

237
00:13:57,440 --> 00:14:01,960
Speaker 1: So why not a hard drive out of the same

238
00:14:02,120 --> 00:14:06,320
Speaker 1: sort of approach, Because it means that you're using instead

239
00:14:06,360 --> 00:14:09,840
Speaker 1: of a mechanical system, you're using an integrated circuit in

240
00:14:09,960 --> 00:14:13,480
Speaker 1: order to store information. You no longer have to worry

241
00:14:13,559 --> 00:14:17,800
Speaker 1: about spinning platters or reading heads or anything like that.

242
00:14:17,880 --> 00:14:21,640
Speaker 1: You can really decrease the amount of time that that

243
00:14:21,720 --> 00:14:25,560
Speaker 1: latency time, so that when you are pulling information from

244
00:14:25,640 --> 00:14:29,240
Speaker 1: the hard drive, it's much closer to the speeds that

245
00:14:29,280 --> 00:14:32,160
Speaker 1: you would see on the CPU die itself, or at

246
00:14:32,200 --> 00:14:35,280
Speaker 1: least in the computer's memory, as opposed to on a

247
00:14:35,320 --> 00:14:38,320
Speaker 1: traditional hard drive. And before we get too far into this,

248
00:14:38,440 --> 00:14:41,400
Speaker 1: I do want to say something about what some of

249
00:14:41,440 --> 00:14:44,280
Speaker 1: the sources of information we pulled from. We do have

250
00:14:44,320 --> 00:14:47,880
Speaker 1: a great article on how stuff works about how flash

251
00:14:47,960 --> 00:14:50,480
Speaker 1: drives work, and a lot of the information applies to

252
00:14:51,040 --> 00:14:55,080
Speaker 1: solid state drives. Yeah they're not exactly the same, but yeah,

253
00:14:55,120 --> 00:14:57,880
Speaker 1: they're they're kind of principles, are there. Yeah, they're close

254
00:14:57,920 --> 00:15:01,080
Speaker 1: cousins because a lot of the thing that go into

255
00:15:01,280 --> 00:15:04,560
Speaker 1: what makes flash drives work apply to solid state drives.

256
00:15:04,560 --> 00:15:08,840
Speaker 1: But an excellent resource on the Web is an Ours

257
00:15:08,960 --> 00:15:12,200
Speaker 1: Technica series, one of which one of the articles in

258
00:15:12,200 --> 00:15:15,640
Speaker 1: that series is called Solid State Revolution in Depth on

259
00:15:15,720 --> 00:15:19,160
Speaker 1: how s s d s Really work by Lee Hutchinson.

260
00:15:19,800 --> 00:15:22,880
Speaker 1: And I can't say enough good things about this article.

261
00:15:23,000 --> 00:15:27,080
Speaker 1: It really is a comprehensive approach to how solid state

262
00:15:27,160 --> 00:15:31,600
Speaker 1: drives work. And there's a little, uh, little latitude in

263
00:15:31,640 --> 00:15:33,560
Speaker 1: the article, so it makes it it's not it's not

264
00:15:33,680 --> 00:15:37,200
Speaker 1: dry reading. No, well, our our Technica is like that. Um.

265
00:15:37,200 --> 00:15:40,280
Speaker 1: It's also just just as a note, it's also very

266
00:15:40,360 --> 00:15:43,560
Speaker 1: technical in spots too, so if you know, it's it's

267
00:15:43,560 --> 00:15:46,360
Speaker 1: a they there site. It takes a little bit different

268
00:15:46,360 --> 00:15:50,360
Speaker 1: approach to technology than than we do in a good way.

269
00:15:50,840 --> 00:15:53,000
Speaker 1: It's at a different level. So it's it's definitely more

270
00:15:53,000 --> 00:15:56,880
Speaker 1: serious if you if you're already familiar, Yeah, if you're

271
00:15:56,880 --> 00:16:01,120
Speaker 1: already familiar with UM with computer art, texture and data

272
00:16:01,160 --> 00:16:04,480
Speaker 1: and that sort of thing. Uh, it's an excellent resource. Otherwise,

273
00:16:04,480 --> 00:16:08,560
Speaker 1: it may it may feel a little advanced for someone

274
00:16:08,600 --> 00:16:11,520
Speaker 1: who is just curious about this but doesn't have any

275
00:16:11,520 --> 00:16:16,880
Speaker 1: real background. However, acts well worthy. Chris and I will

276
00:16:16,920 --> 00:16:19,400
Speaker 1: return in just a moment to talk more about solid

277
00:16:19,400 --> 00:16:29,640
Speaker 1: state drives, but first let's take a quick break. So

278
00:16:29,720 --> 00:16:33,600
Speaker 1: getting back to solid state drives. So the idea of

279
00:16:33,640 --> 00:16:37,880
Speaker 1: creating a solid state drive is was really really attractive

280
00:16:38,280 --> 00:16:42,280
Speaker 1: because of that decreased latency. Um, there were some challenges

281
00:16:42,360 --> 00:16:45,640
Speaker 1: of course, because solid state drives they do not store

282
00:16:45,760 --> 00:16:51,280
Speaker 1: information magnetically the way traditional hard drives do. Now, actually

283
00:16:51,320 --> 00:16:56,240
Speaker 1: it sort of reminds me of electronic ink in a way. Interesting. Well,

284
00:16:56,640 --> 00:16:58,760
Speaker 1: if you know something about electronic ink, you know that

285
00:16:59,240 --> 00:17:03,600
Speaker 1: the capsules white or black generally are stored in between

286
00:17:03,640 --> 00:17:08,639
Speaker 1: a sandwich of h material that holds a positive or

287
00:17:08,720 --> 00:17:11,280
Speaker 1: negative charge, and that's how it reads a page. But

288
00:17:11,320 --> 00:17:14,359
Speaker 1: once the page is there, it stays there. I see.

289
00:17:14,400 --> 00:17:17,600
Speaker 1: So you're thinking of that as for it. Let's just say,

290
00:17:17,640 --> 00:17:20,720
Speaker 1: for example, that the black parts of the screen are

291
00:17:20,800 --> 00:17:24,480
Speaker 1: ones and the white parts are zeros, and they retain

292
00:17:24,640 --> 00:17:28,680
Speaker 1: that even when the power is off, so they're non volatile. Right.

293
00:17:28,800 --> 00:17:31,720
Speaker 1: That means that when you remove the power source from

294
00:17:31,760 --> 00:17:34,920
Speaker 1: this system, it keeps that information. Of that of course

295
00:17:35,000 --> 00:17:37,439
Speaker 1: is extremely important when it comes to computers because, like

296
00:17:37,520 --> 00:17:42,200
Speaker 1: Chris was saying, RAM is volatile memory. If you lose

297
00:17:42,280 --> 00:17:46,400
Speaker 1: that power, then that information goes away. There's no longer

298
00:17:46,440 --> 00:17:49,360
Speaker 1: a charge to maintain the information that's stored in your

299
00:17:49,359 --> 00:17:51,959
Speaker 1: computer's memory. You don't want that to happen to your

300
00:17:52,000 --> 00:17:53,879
Speaker 1: hard drive because that means that every time you would

301
00:17:53,880 --> 00:17:56,680
Speaker 1: turn off your computer or lose power, you would lose

302
00:17:56,800 --> 00:17:59,080
Speaker 1: all the data stored there. You have to have non

303
00:17:59,200 --> 00:18:04,800
Speaker 1: volatile memory to keep storage a possibility. Well, nice, yeah,

304
00:18:04,880 --> 00:18:07,720
Speaker 1: because I mean the old, the old, old, old computers

305
00:18:07,760 --> 00:18:10,679
Speaker 1: that that Chris and I worked on didn't have hard drives.

306
00:18:11,200 --> 00:18:14,560
Speaker 1: You had to store things on magnetic discs. If you

307
00:18:14,640 --> 00:18:17,800
Speaker 1: turned your computer on, it just went to its initial state.

308
00:18:17,960 --> 00:18:20,080
Speaker 1: The only thing that was stored on there was the

309
00:18:20,119 --> 00:18:22,879
Speaker 1: operating system because it was written in read only memory,

310
00:18:22,920 --> 00:18:26,359
Speaker 1: which was non volatile, but it was also unchangeable. You

311
00:18:26,400 --> 00:18:29,919
Speaker 1: couldn't write to it. So that meant that, you know,

312
00:18:30,040 --> 00:18:31,760
Speaker 1: if you wanted to write a program, you had to

313
00:18:31,760 --> 00:18:33,600
Speaker 1: store it on a disk, because if you tried to

314
00:18:33,640 --> 00:18:35,359
Speaker 1: write it just on your computer, you didn't have a

315
00:18:35,400 --> 00:18:37,040
Speaker 1: disk in there, and you turned the computer off, all

316
00:18:37,080 --> 00:18:39,520
Speaker 1: your work has gone. This is also why if you've

317
00:18:39,560 --> 00:18:41,679
Speaker 1: ever worked on a computer, and you've ever heard anyone

318
00:18:41,720 --> 00:18:45,800
Speaker 1: say save your work, often that's why when you save

319
00:18:45,840 --> 00:18:48,600
Speaker 1: your work, it's being saved to your hard drive not

320
00:18:48,680 --> 00:18:50,920
Speaker 1: to your computer's memory. So if you're working on something

321
00:18:50,920 --> 00:18:53,040
Speaker 1: and you haven't saved it in a while, it may

322
00:18:53,200 --> 00:18:55,920
Speaker 1: only exist in your computer's memory. If power goes out,

323
00:18:56,000 --> 00:18:58,520
Speaker 1: you may lose all that work, as I have done

324
00:18:58,960 --> 00:19:02,280
Speaker 1: on multiple asians. I was actually in my college's computer

325
00:19:02,359 --> 00:19:07,040
Speaker 1: lab during a storm and the screams power went out.

326
00:19:07,200 --> 00:19:10,920
Speaker 1: There's nothing like working in any sort of computer environment

327
00:19:11,000 --> 00:19:13,280
Speaker 1: when the power goes out and then you hear there's

328
00:19:13,320 --> 00:19:17,040
Speaker 1: there's usually about a second and a half delay between

329
00:19:17,040 --> 00:19:21,120
Speaker 1: the power going out and every single person making essentially

330
00:19:21,119 --> 00:19:26,800
Speaker 1: the same noise which sounds like this. Ah. I leaned

331
00:19:26,800 --> 00:19:28,640
Speaker 1: back as I did that, so there was a little

332
00:19:28,680 --> 00:19:32,760
Speaker 1: Doppler effect. But anyway, so yeah, this this sort of

333
00:19:32,760 --> 00:19:36,879
Speaker 1: of non volatile memory means that that information is going

334
00:19:36,920 --> 00:19:39,240
Speaker 1: to stay there even when you turn the power off.

335
00:19:39,280 --> 00:19:41,720
Speaker 1: This is the exact same sort of stuff we find

336
00:19:41,720 --> 00:19:46,320
Speaker 1: in our MP three players and other mobile devices that

337
00:19:46,600 --> 00:19:49,240
Speaker 1: because again, if we didn't have that, then every time

338
00:19:49,240 --> 00:19:51,800
Speaker 1: you turned off your MP three player, you would lose

339
00:19:51,800 --> 00:19:54,000
Speaker 1: your entire library of songs. You have to reload it

340
00:19:54,040 --> 00:19:56,280
Speaker 1: the next time you turn it on. Right, right, Well,

341
00:19:56,320 --> 00:20:01,159
Speaker 1: there is um to to complete my analogy along with

342
00:20:01,200 --> 00:20:07,040
Speaker 1: the electronic ink thing. The flash memory also uses cells.

343
00:20:07,760 --> 00:20:11,679
Speaker 1: Um it stores information in in cells, and that cell

344
00:20:12,040 --> 00:20:14,640
Speaker 1: is either a one or a zero. Right to think

345
00:20:14,680 --> 00:20:17,359
Speaker 1: of the cells like a sheet of grid paper. Yes,

346
00:20:17,640 --> 00:20:22,040
Speaker 1: and you've got rows, and you've got columns. Right, So

347
00:20:22,119 --> 00:20:25,640
Speaker 1: the rows of cells. If you took one row of cells,

348
00:20:25,760 --> 00:20:28,919
Speaker 1: we would call that in in solid state drive terminology,

349
00:20:29,000 --> 00:20:32,200
Speaker 1: that would be a page. So one row of the

350
00:20:32,600 --> 00:20:34,560
Speaker 1: cells would be a page, and then you would have

351
00:20:34,920 --> 00:20:37,840
Speaker 1: several rows of cells and several columns of cells that

352
00:20:37,920 --> 00:20:41,560
Speaker 1: would form what is called a block. And this is

353
00:20:41,560 --> 00:20:44,560
Speaker 1: really important because it comes down to the way information

354
00:20:44,960 --> 00:20:48,760
Speaker 1: is written and erased in solid state drives. It turns

355
00:20:48,760 --> 00:20:54,440
Speaker 1: out that you cannot individually change the cells within that

356
00:20:54,960 --> 00:20:56,680
Speaker 1: grid paper. For example, if you if you had a

357
00:20:56,680 --> 00:20:58,480
Speaker 1: sheet of grid paper in front of you, you could

358
00:20:58,480 --> 00:21:01,159
Speaker 1: write a one or a zero and every single grid

359
00:21:01,520 --> 00:21:03,320
Speaker 1: and if you wanted to, if you were writing in pencil,

360
00:21:03,359 --> 00:21:07,520
Speaker 1: you could erase a single cell and change that one

361
00:21:07,600 --> 00:21:10,400
Speaker 1: to a zero or zero to a one. You cannot

362
00:21:10,440 --> 00:21:13,359
Speaker 1: do that with a solid state drive. We'll get into

363
00:21:13,359 --> 00:21:15,040
Speaker 1: that in a little bit, but that's an important thing

364
00:21:15,080 --> 00:21:18,600
Speaker 1: to think about from the start. Well, if you know

365
00:21:18,720 --> 00:21:22,760
Speaker 1: something about hard drives. The the magnetic platter hard drives.

366
00:21:23,280 --> 00:21:26,200
Speaker 1: UM when you when your computer, and that pretty much

367
00:21:26,240 --> 00:21:29,960
Speaker 1: goes for all modern operating systems. Let's say you have

368
00:21:30,640 --> 00:21:34,119
Speaker 1: a document and you realize that well, you worked on

369
00:21:34,119 --> 00:21:37,199
Speaker 1: it three years ago, You've turned it in. You know,

370
00:21:37,240 --> 00:21:38,840
Speaker 1: I don't need to save it for anything, So I'm

371
00:21:38,840 --> 00:21:41,560
Speaker 1: going to delete it, and I'm gonna tell my computer delete.

372
00:21:41,640 --> 00:21:45,000
Speaker 1: Well that first of all, the computer doesn't delete it

373
00:21:45,040 --> 00:21:47,400
Speaker 1: deleted if you just tell it to you know, throw

374
00:21:47,400 --> 00:21:49,040
Speaker 1: it in the trash can, empty the trash can or

375
00:21:49,080 --> 00:21:51,639
Speaker 1: recycle bin or whatever. It's actually still there on your

376
00:21:51,680 --> 00:21:55,040
Speaker 1: hard drive, but it's been marked for deletion. So basically,

377
00:21:55,080 --> 00:21:58,360
Speaker 1: when uh, something else, Hey I've created a new document,

378
00:21:58,520 --> 00:22:01,640
Speaker 1: and the computer goes or can I say, ah, this

379
00:22:01,720 --> 00:22:06,560
Speaker 1: is marked for deletion, I'll right over that old one. Um.

380
00:22:06,640 --> 00:22:09,359
Speaker 1: That's one thing to note is that it can do that.

381
00:22:09,359 --> 00:22:13,040
Speaker 1: That's the way that that computers work with magnetic platter

382
00:22:13,080 --> 00:22:16,320
Speaker 1: hard drives. Now it also can do something else. Let's

383
00:22:16,320 --> 00:22:19,200
Speaker 1: say you have five different documents and you've deleted these

384
00:22:19,280 --> 00:22:24,160
Speaker 1: five documents. Well, those gaps are different sizes, but you're

385
00:22:24,320 --> 00:22:26,640
Speaker 1: storing a brand new file and it's larger than all

386
00:22:26,680 --> 00:22:29,800
Speaker 1: five of those. It can fill in sections sort of

387
00:22:29,840 --> 00:22:32,600
Speaker 1: like packets when you send, when you break up an

388
00:22:32,600 --> 00:22:34,359
Speaker 1: email file into a bunch of packets and they go

389
00:22:34,440 --> 00:22:37,560
Speaker 1: and they reassemble themselves on the other side on somebody

390
00:22:37,560 --> 00:22:41,800
Speaker 1: else's computer. These different gaps can be used to store

391
00:22:41,960 --> 00:22:45,639
Speaker 1: parts of this file, which the computer will then reassemble

392
00:22:45,720 --> 00:22:48,199
Speaker 1: as you need it. Um. That's when you need to

393
00:22:48,320 --> 00:22:51,879
Speaker 1: fragment your hard drive. You know, they have sections and

394
00:22:51,920 --> 00:22:54,600
Speaker 1: they're all scattered out. You've got applications, and then they're

395
00:22:54,640 --> 00:22:58,639
Speaker 1: all in different places, in different sectors, on different platters, um,

396
00:22:58,680 --> 00:23:02,080
Speaker 1: And they say, okay, I going to reorganize everything. And

397
00:23:02,160 --> 00:23:07,240
Speaker 1: so the computer basically uses storage empty storage to reshuffle

398
00:23:07,320 --> 00:23:09,679
Speaker 1: everything and put it back into sectors where all the

399
00:23:09,760 --> 00:23:12,800
Speaker 1: parts of the filer together. And that makes a computer

400
00:23:12,920 --> 00:23:15,159
Speaker 1: run a little bit faster when it's accessing those files

401
00:23:15,160 --> 00:23:16,720
Speaker 1: because they're on one place and they don't have to

402
00:23:16,760 --> 00:23:21,199
Speaker 1: reassemble them. Now, you can't do that with uh, solid

403
00:23:21,200 --> 00:23:23,600
Speaker 1: state drives. Yeah. In fact, when it gets down to

404
00:23:23,760 --> 00:23:29,400
Speaker 1: erasing data off a solid state drive, it's pretty pretty complex.

405
00:23:29,400 --> 00:23:31,400
Speaker 1: But before we get into that, let's talk a little

406
00:23:31,400 --> 00:23:33,399
Speaker 1: bit a little a little bit more about the way

407
00:23:33,600 --> 00:23:36,399
Speaker 1: information is stored within these cells. So these grids on

408
00:23:36,440 --> 00:23:37,960
Speaker 1: your grid paper, Yeah, I just thought it would be

409
00:23:38,000 --> 00:23:41,560
Speaker 1: interesting to compare that to a platter drive entirely. Yeah,

410
00:23:41,600 --> 00:23:44,240
Speaker 1: And it is important to make the comparisons between the

411
00:23:44,280 --> 00:23:47,119
Speaker 1: two because there are advantages and disadvantages to both. That

412
00:23:47,200 --> 00:23:51,639
Speaker 1: whole erasing thing or overwriting. You can't overwrite in an SSD.

413
00:23:52,080 --> 00:23:55,439
Speaker 1: You can't erase stuff. But it takes a lot of

414
00:23:55,480 --> 00:24:01,640
Speaker 1: effort actually, um. But anyway, so in order to make

415
00:24:01,840 --> 00:24:06,119
Speaker 1: the the cell have a value in it, you have

416
00:24:06,200 --> 00:24:10,440
Speaker 1: to apply a voltage to that cell. Yes, I couldn't.

417
00:24:10,480 --> 00:24:14,720
Speaker 1: I couldn't read that article without thinking of a C

418
00:24:14,880 --> 00:24:18,400
Speaker 1: d C, the band with the Dirty deeds. That they're

419
00:24:18,480 --> 00:24:23,360
Speaker 1: cheap voltage. It uses high voltage. Actually that's important, yes, yeah,

420
00:24:23,440 --> 00:24:26,960
Speaker 1: in fact they're well uses both high and low voltage.

421
00:24:28,760 --> 00:24:31,800
Speaker 1: There are two different ways of wiring these transistors together.

422
00:24:31,840 --> 00:24:35,679
Speaker 1: That's what each of the cells actually represents. Um. There

423
00:24:35,920 --> 00:24:39,840
Speaker 1: is the nor approach, which is a little simpler but

424
00:24:40,160 --> 00:24:44,640
Speaker 1: less useful. Really, it's typical of flash drives, however, these

425
00:24:44,680 --> 00:24:48,520
Speaker 1: smaller drives. So think of think of those those rows

426
00:24:48,640 --> 00:24:55,000
Speaker 1: there and the uh, the columns as having a a

427
00:24:55,000 --> 00:24:57,920
Speaker 1: a circuit line going through each one. All right, so

428
00:24:58,000 --> 00:25:03,000
Speaker 1: there are rows of circuit and columns of circuits, uh, connections,

429
00:25:03,000 --> 00:25:05,639
Speaker 1: really electronic connections, I should say, not just circuits, but

430
00:25:05,720 --> 00:25:10,720
Speaker 1: electronic connections. So the rows would be word lines, the

431
00:25:10,840 --> 00:25:13,600
Speaker 1: columns would be bitlines. So you would have this grid

432
00:25:13,800 --> 00:25:17,280
Speaker 1: of word lines and bitlines. Sort of it would look

433
00:25:17,359 --> 00:25:19,600
Speaker 1: kind of like a city block, like if you were

434
00:25:19,640 --> 00:25:21,959
Speaker 1: looking for a city landscape if you were looking at

435
00:25:21,960 --> 00:25:24,880
Speaker 1: it from the air. So you have these these streets

436
00:25:24,880 --> 00:25:29,560
Speaker 1: that are criss crossing um and you would kind of

437
00:25:30,680 --> 00:25:35,400
Speaker 1: tell each cell what its contents were based upon applying

438
00:25:35,920 --> 00:25:38,920
Speaker 1: voltages across these lines. I'm not going to get too

439
00:25:38,920 --> 00:25:42,399
Speaker 1: far into this because it really gets kind of complex

440
00:25:42,400 --> 00:25:46,119
Speaker 1: and also involves a concept called quantum tunneling, which we

441
00:25:46,160 --> 00:25:49,000
Speaker 1: have talked about here on tech stuff before. But it

442
00:25:49,040 --> 00:25:53,280
Speaker 1: makes my brain hurt because it's Quantum Tunneling is one

443
00:25:53,320 --> 00:25:57,040
Speaker 1: of those things that is crazy to me. Tunneling is

444
00:25:57,080 --> 00:26:00,399
Speaker 1: this concept that and it's it's a real thing, otherwise

445
00:26:00,400 --> 00:26:03,879
Speaker 1: our electronics wouldn't work. Tunneling is this concept where you

446
00:26:03,960 --> 00:26:07,560
Speaker 1: have a barrier, let's say, and you've got an electron.

447
00:26:07,680 --> 00:26:10,320
Speaker 1: That part of me an electron that's on one side

448
00:26:10,359 --> 00:26:14,760
Speaker 1: of that barrier. With the right kind of energy, that

449
00:26:14,880 --> 00:26:17,520
Speaker 1: electron can pass from one side of the barrier to

450
00:26:17,560 --> 00:26:19,119
Speaker 1: the other side of the barrier as if it has

451
00:26:19,160 --> 00:26:23,320
Speaker 1: tunneled through without actually physically tunneling through. And this all

452
00:26:23,359 --> 00:26:25,679
Speaker 1: has to do with the potential for the electron to

453
00:26:25,800 --> 00:26:29,399
Speaker 1: be in one position versus another. UH. There's sort of

454
00:26:29,480 --> 00:26:31,080
Speaker 1: you can think of it as a there's a radius

455
00:26:31,119 --> 00:26:35,159
Speaker 1: around an electron that represents all the different locations that

456
00:26:35,240 --> 00:26:39,399
Speaker 1: electron could be in. UH. If that electrons at the

457
00:26:39,480 --> 00:26:42,439
Speaker 1: proper energy state, and there is a barrier next to

458
00:26:42,440 --> 00:26:46,399
Speaker 1: the electron, that radius might extend beyond the other side

459
00:26:46,600 --> 00:26:49,760
Speaker 1: of that barrier. That means that there's the potential for

460
00:26:49,840 --> 00:26:52,520
Speaker 1: that electron to be on the other side of that barrier,

461
00:26:53,040 --> 00:26:55,960
Speaker 1: which means if there is a potential for it, sometimes

462
00:26:56,040 --> 00:26:58,439
Speaker 1: the electron is on the other side of that barrier,

463
00:26:59,520 --> 00:27:03,919
Speaker 1: as if the area weren't there. This drives me insane.

464
00:27:04,320 --> 00:27:06,240
Speaker 1: It's like saying, if I'm if I'm running fast enough,

465
00:27:06,240 --> 00:27:07,640
Speaker 1: there's a chance I'm going to be on the other

466
00:27:07,680 --> 00:27:09,160
Speaker 1: side of the wall, not on this side of the wall.

467
00:27:09,200 --> 00:27:10,720
Speaker 1: But every time I try that, I end up with

468
00:27:10,720 --> 00:27:17,120
Speaker 1: the bloody knows I'm not quantum enough, is what it. Yeah,

469
00:27:17,440 --> 00:27:20,320
Speaker 1: we'll return to this classic episode of tech stuff in

470
00:27:20,400 --> 00:27:22,560
Speaker 1: just a moment after we take this break to thank

471
00:27:22,560 --> 00:27:33,439
Speaker 1: our sponsor. In the nand one, the bitlines are actually

472
00:27:33,520 --> 00:27:36,880
Speaker 1: kind of um daisy chained in a way from one

473
00:27:36,920 --> 00:27:41,399
Speaker 1: cell down to the next, and this becomes important when

474
00:27:41,400 --> 00:27:44,000
Speaker 1: you were actually reading from the memory in order to

475
00:27:44,040 --> 00:27:48,679
Speaker 1: determine what bit is in each cell uh. And the

476
00:27:48,720 --> 00:27:52,800
Speaker 1: way that works is that you apply a weak voltage

477
00:27:53,640 --> 00:27:57,240
Speaker 1: across these lines to try and determine if a full

478
00:27:57,320 --> 00:28:01,120
Speaker 1: circuit is being made, and if there, you get two

479
00:28:01,160 --> 00:28:05,920
Speaker 1: different outcomes depending on if there's a one or a zero. Right.

480
00:28:07,119 --> 00:28:10,639
Speaker 1: So if you get one outcome, for example, if the

481
00:28:10,640 --> 00:28:14,120
Speaker 1: circuit is made, you know that the value is uh,

482
00:28:14,520 --> 00:28:16,560
Speaker 1: you know what the value is inside that cell because

483
00:28:16,560 --> 00:28:18,840
Speaker 1: it can only be that value. And if the circuit

484
00:28:18,920 --> 00:28:21,720
Speaker 1: is not made, then you know it's the opposite value, right.

485
00:28:22,480 --> 00:28:25,400
Speaker 1: And so you're you're thinking, well, if the charge goes through,

486
00:28:25,440 --> 00:28:27,399
Speaker 1: it means this. If the charge doesn't go through, it

487
00:28:27,400 --> 00:28:31,040
Speaker 1: means that you collect that with all of these cells,

488
00:28:31,160 --> 00:28:33,439
Speaker 1: and that's what builds up data. Remember, each of these

489
00:28:33,440 --> 00:28:36,880
Speaker 1: cells represents one bit, so a zero or a one,

490
00:28:37,000 --> 00:28:40,640
Speaker 1: unless that's a well, I guess I should say that

491
00:28:40,640 --> 00:28:43,800
Speaker 1: that would be an s l C, a single level

492
00:28:43,840 --> 00:28:47,040
Speaker 1: cell which can represent either a zero or a one.

493
00:28:47,320 --> 00:28:49,520
Speaker 1: You could also have a multi level cell, and in fact,

494
00:28:49,560 --> 00:28:52,200
Speaker 1: most ssd s are multi level cells. Now, these can

495
00:28:52,240 --> 00:28:55,560
Speaker 1: contain UH two or more bits. Usually it's two bits

496
00:28:55,640 --> 00:28:57,800
Speaker 1: or three bits, which means that if it's a two

497
00:28:57,800 --> 00:29:01,360
Speaker 1: bit system, there are four show values that you could

498
00:29:01,360 --> 00:29:05,120
Speaker 1: find within that cell. It would either be zero, zero, zero, one,

499
00:29:05,280 --> 00:29:10,280
Speaker 1: one zero, or one one. But with a multi level cell, UH,

500
00:29:10,400 --> 00:29:12,520
Speaker 1: it's a little more complicated as well, because you know,

501
00:29:12,600 --> 00:29:14,440
Speaker 1: like I said, you could use the weak voltage to

502
00:29:14,520 --> 00:29:18,160
Speaker 1: determine whether or not the content of a single layer

503
00:29:18,200 --> 00:29:20,840
Speaker 1: cell is a zero or a one. With multi level

504
00:29:20,960 --> 00:29:25,480
Speaker 1: level cell, because there are four potential UH outcomes, you

505
00:29:25,520 --> 00:29:28,440
Speaker 1: have to use different voltages and essentially you work from

506
00:29:28,440 --> 00:29:30,280
Speaker 1: the weakest and you work your way up and as

507
00:29:30,280 --> 00:29:32,680
Speaker 1: soon as that circuit is complete, then you know what

508
00:29:32,800 --> 00:29:38,320
Speaker 1: the value of that cell is. Did I just Chris's

509
00:29:38,360 --> 00:29:40,760
Speaker 1: brains are actually leaking out of his ears right now.

510
00:29:42,080 --> 00:29:44,560
Speaker 1: I just find the whole thing revolting. Yeah, I thought

511
00:29:44,600 --> 00:29:47,960
Speaker 1: you would get a charge out of it, So anyway. Yeah,

512
00:29:47,960 --> 00:29:50,600
Speaker 1: it all has to do with these voltages and uh

513
00:29:50,760 --> 00:29:55,840
Speaker 1: and that that's all the reading information to write to

514
00:29:56,000 --> 00:30:01,320
Speaker 1: a to it is um even more complex actually. Well.

515
00:30:01,880 --> 00:30:04,880
Speaker 1: One of the things to note though is that making

516
00:30:04,920 --> 00:30:11,320
Speaker 1: these changes, uh the voltage changes can be dangerous um

517
00:30:11,400 --> 00:30:14,400
Speaker 1: using the high voltage to to do these changes, which

518
00:30:14,440 --> 00:30:17,200
Speaker 1: is one of the reasons why it is so difficult

519
00:30:17,240 --> 00:30:23,280
Speaker 1: to uh to erase and rewrite on this flash memory

520
00:30:23,320 --> 00:30:25,600
Speaker 1: that that's used in SMCS. That's one of the reasons.

521
00:30:25,640 --> 00:30:29,080
Speaker 1: The other reason is that when you are writing information

522
00:30:29,320 --> 00:30:33,000
Speaker 1: to a an SSD, you have to write it in

523
00:30:33,000 --> 00:30:36,360
Speaker 1: in pages so rows. So think of think of this

524
00:30:36,440 --> 00:30:39,720
Speaker 1: grid paper again. You can only write to a single

525
00:30:39,800 --> 00:30:43,640
Speaker 1: row at a time, all right, when you're erasing, you

526
00:30:43,720 --> 00:30:46,800
Speaker 1: have to erase these in blocks of pages. You can't

527
00:30:46,840 --> 00:30:50,200
Speaker 1: erase one row, you can't erase one page. You have

528
00:30:50,240 --> 00:30:52,520
Speaker 1: to erase a block of pages, which tends to be

529
00:30:52,520 --> 00:30:58,040
Speaker 1: about eight rows total, which equals about five and twelve kilobytes.

530
00:30:59,120 --> 00:31:03,240
Speaker 1: There's actually some extra information there too, because there are

531
00:31:03,280 --> 00:31:06,840
Speaker 1: a few, uh few cells that are dedicated to things

532
00:31:06,880 --> 00:31:10,840
Speaker 1: like error correction and other information. So there's technically a

533
00:31:10,880 --> 00:31:13,600
Speaker 1: little more than that. But the data that you're actually

534
00:31:13,640 --> 00:31:17,040
Speaker 1: writing to the s s D or erasing from the

535
00:31:17,120 --> 00:31:21,560
Speaker 1: SSD is either in four or eight kilobyte pages. Again,

536
00:31:21,560 --> 00:31:25,080
Speaker 1: that depends upon the format of the solid state drive.

537
00:31:26,000 --> 00:31:29,040
Speaker 1: And uh, when you're racing's twenty eight pages, so five

538
00:31:29,240 --> 00:31:34,360
Speaker 1: and twelve kilobites so or for the four kilobyte pages anyway,

539
00:31:34,680 --> 00:31:39,160
Speaker 1: so you've got you can write in a page, you

540
00:31:39,240 --> 00:31:42,880
Speaker 1: erase in a block. This is why it's really hard

541
00:31:43,000 --> 00:31:46,400
Speaker 1: to You can't. This is why you cannot overwrite information

542
00:31:46,560 --> 00:31:50,080
Speaker 1: because your file that you know that this has nothing

543
00:31:50,120 --> 00:31:53,440
Speaker 1: to do with file size, is just the individual bits

544
00:31:53,680 --> 00:31:58,640
Speaker 1: that are found within those cells. Right, So a block

545
00:31:58,840 --> 00:32:01,600
Speaker 1: might contain the end of one file in the beginning

546
00:32:01,640 --> 00:32:05,840
Speaker 1: of another file. So you can't erase an entire block, uh,

547
00:32:06,280 --> 00:32:10,000
Speaker 1: just because you erased one particular file, because you if

548
00:32:10,040 --> 00:32:11,840
Speaker 1: you did that, then you would lose the beginning of

549
00:32:11,920 --> 00:32:16,120
Speaker 1: a of an unrelated file that you did not delete, right,

550
00:32:16,400 --> 00:32:20,240
Speaker 1: And and it treats old files that have been marked

551
00:32:20,280 --> 00:32:24,960
Speaker 1: for deletion as information that should be saved. It doesn't

552
00:32:25,080 --> 00:32:29,520
Speaker 1: make there's on every drive there is a controller that

553
00:32:29,640 --> 00:32:32,000
Speaker 1: provides instruction for the drive, and it doesn't know the

554
00:32:32,120 --> 00:32:37,560
Speaker 1: difference between the file that I just uh quote unquote

555
00:32:37,640 --> 00:32:41,720
Speaker 1: on my computer deleted and the computer marks for delete. No,

556
00:32:41,960 --> 00:32:45,800
Speaker 1: it's okay to overwrite this sector of the drive versus

557
00:32:46,480 --> 00:32:49,719
Speaker 1: a file that I want to keep, So it treats

558
00:32:49,800 --> 00:32:52,360
Speaker 1: anything that's written in there as well. I better save

559
00:32:52,480 --> 00:32:56,560
Speaker 1: this and just so that you guys can kind of

560
00:32:56,720 --> 00:32:59,120
Speaker 1: envision what is going on here. So let's go back

561
00:32:59,160 --> 00:33:02,040
Speaker 1: to that grid paper example. Let's say you've got a

562
00:33:02,400 --> 00:33:05,680
Speaker 1: sheet of grid paper. If this were like a solid

563
00:33:05,760 --> 00:33:07,800
Speaker 1: state drive, every single one of those cells, if this

564
00:33:07,920 --> 00:33:11,360
Speaker 1: was a brand new sheet, nothing had been saved to

565
00:33:11,520 --> 00:33:14,680
Speaker 1: the sheet. Yet, actually every single one of those cells

566
00:33:14,760 --> 00:33:17,160
Speaker 1: would have a one in them. And when you were

567
00:33:17,240 --> 00:33:20,200
Speaker 1: writing information, what you do is you apply a certain

568
00:33:20,280 --> 00:33:24,160
Speaker 1: voltage and you would switch that one to a zero. Switch.

569
00:33:24,240 --> 00:33:26,560
Speaker 1: Switching the one to a zero is not such a

570
00:33:26,640 --> 00:33:30,160
Speaker 1: big deal. Switching the zero back to one is a

571
00:33:30,360 --> 00:33:34,920
Speaker 1: huge deal. And here's here's why you cannot overwrite specific

572
00:33:35,040 --> 00:33:39,320
Speaker 1: parts of this this page. You have to use enough

573
00:33:39,440 --> 00:33:43,680
Speaker 1: voltage to switch that back to a one that if

574
00:33:43,720 --> 00:33:46,600
Speaker 1: you were to try and target a single cell, that

575
00:33:46,960 --> 00:33:51,840
Speaker 1: energy could overflow into neighboring cells, which would make those flip.

576
00:33:53,160 --> 00:33:55,240
Speaker 1: And if you're making all of those flip. That means

577
00:33:55,280 --> 00:33:58,440
Speaker 1: you've just corrupted the data, right because not all those

578
00:33:58,760 --> 00:34:01,080
Speaker 1: not all of those rows need to be turned back

579
00:34:01,120 --> 00:34:06,080
Speaker 1: into one's sound effective. Students going yeah, so that's so,

580
00:34:06,240 --> 00:34:08,800
Speaker 1: that's why you can't target a specific cell. You have

581
00:34:08,920 --> 00:34:11,799
Speaker 1: to do it in these blocks. And uh. The way,

582
00:34:12,400 --> 00:34:15,480
Speaker 1: the way a solid state drive actually does this handles

583
00:34:15,520 --> 00:34:18,440
Speaker 1: this because eventually you will have to have that information

584
00:34:19,360 --> 00:34:22,600
Speaker 1: erased or else you'll run out of space. You'll run

585
00:34:22,640 --> 00:34:25,120
Speaker 1: out of space. Just even every time you save a

586
00:34:25,280 --> 00:34:28,120
Speaker 1: new version of that document, if it's a document that

587
00:34:28,160 --> 00:34:29,960
Speaker 1: you're working on as opposed to like we'll use that

588
00:34:30,000 --> 00:34:34,680
Speaker 1: as an example, um as a type of file you're creating.

589
00:34:34,760 --> 00:34:37,239
Speaker 1: Let's say that you've created a document. Every time you

590
00:34:37,280 --> 00:34:39,600
Speaker 1: save a new version of it, it's writing that information

591
00:34:39,800 --> 00:34:43,600
Speaker 1: to more pages on your solid state drive. Well, if

592
00:34:43,760 --> 00:34:46,880
Speaker 1: you never erased, if if it never had the opportunity

593
00:34:46,960 --> 00:34:49,279
Speaker 1: to erase the information on that drive, you would run

594
00:34:49,320 --> 00:34:53,960
Speaker 1: out of space eventually. Yes, So the way it tries

595
00:34:54,040 --> 00:34:57,160
Speaker 1: to handle this is that eventually there's a there's got

596
00:34:57,239 --> 00:35:01,040
Speaker 1: to be a connection between the operating system and the

597
00:35:01,120 --> 00:35:04,640
Speaker 1: solid state drive that lets the solid state drive no

598
00:35:05,520 --> 00:35:10,760
Speaker 1: this particular information that is stored within your pages is stale.

599
00:35:11,480 --> 00:35:15,880
Speaker 1: This information does not really um this, this isn't pertinent anymore.

600
00:35:16,160 --> 00:35:19,919
Speaker 1: You can get rid of this. What will happen is rubbish. Yeah,

601
00:35:20,160 --> 00:35:24,160
Speaker 1: and there's this is called garbage collection. Actually, what happens

602
00:35:24,280 --> 00:35:27,560
Speaker 1: is the solid state drive will take a block that

603
00:35:27,719 --> 00:35:32,680
Speaker 1: contains the pages that have stale information, and we'll copy

604
00:35:32,800 --> 00:35:37,640
Speaker 1: that entire block and save it again within the drive.

605
00:35:38,000 --> 00:35:41,759
Speaker 1: So now you're saying, wait a minute, now you just uh, well,

606
00:35:41,840 --> 00:35:45,440
Speaker 1: technically it only copy the stuff that is um that

607
00:35:45,520 --> 00:35:49,040
Speaker 1: does not stale. So you've got a block of of pages.

608
00:35:49,360 --> 00:35:51,160
Speaker 1: Some of those pages are stale, some of the pages

609
00:35:51,239 --> 00:35:54,760
Speaker 1: aren't stale. The solid state drive will copy the stuff

610
00:35:54,760 --> 00:35:57,400
Speaker 1: that's not stale and paste into a new block and

611
00:35:57,600 --> 00:36:01,319
Speaker 1: a new series of pages. So you've just doubled all

612
00:36:01,440 --> 00:36:04,920
Speaker 1: the non stale content that is on your solid state drive.

613
00:36:04,960 --> 00:36:06,800
Speaker 1: And I hear you screaming, but you said this is

614
00:36:06,840 --> 00:36:09,760
Speaker 1: to conserve space. How can you conserve space by copying

615
00:36:09,840 --> 00:36:13,520
Speaker 1: and pasting? Here's how. After that information has been copied

616
00:36:13,560 --> 00:36:17,359
Speaker 1: and pasted into the new section of the solid state drive,

617
00:36:17,920 --> 00:36:20,840
Speaker 1: the old block that has both the stale and not

618
00:36:21,080 --> 00:36:25,000
Speaker 1: stale information in it can be erased. You can apply

619
00:36:25,160 --> 00:36:28,080
Speaker 1: that high voltage flip those zeros back to one's, and

620
00:36:28,160 --> 00:36:31,160
Speaker 1: you can do it safely because you've already duplicated the

621
00:36:31,320 --> 00:36:35,640
Speaker 1: non stale data into new pages. The stale data does

622
00:36:35,719 --> 00:36:38,120
Speaker 1: not get duplicated, so it gets erased, which means that

623
00:36:38,600 --> 00:36:43,439
Speaker 1: that block is now available to write to again. There's

624
00:36:43,440 --> 00:36:47,040
Speaker 1: another downside here, which is that every single time you're

625
00:36:47,080 --> 00:36:51,560
Speaker 1: writing to those cells, you're actually breaking down the system

626
00:36:51,600 --> 00:36:54,400
Speaker 1: a little bit. There's only so many times you can

627
00:36:54,520 --> 00:36:58,279
Speaker 1: do this and the cells will remain viable. Eventually, the

628
00:36:58,360 --> 00:37:01,200
Speaker 1: cells will no longer be able to hold a charge

629
00:37:01,239 --> 00:37:03,520
Speaker 1: because they've been broken down too many times with this

630
00:37:03,640 --> 00:37:08,680
Speaker 1: voltage being applied to them. Um SSD vendors have gone

631
00:37:08,840 --> 00:37:12,400
Speaker 1: to some effort to prevent that from being an issue,

632
00:37:12,800 --> 00:37:15,520
Speaker 1: at least for a while. Um In a lot of cases,

633
00:37:16,320 --> 00:37:20,480
Speaker 1: there will be uhm extra space on the drive of

634
00:37:20,560 --> 00:37:23,120
Speaker 1: which you are unaware. Right. You might have a say,

635
00:37:24,040 --> 00:37:26,880
Speaker 1: let's say that you get I don't know, a sixty

636
00:37:26,960 --> 00:37:32,200
Speaker 1: gigabit gigabyte rather hard drive space, and there's actually sixty

637
00:37:32,280 --> 00:37:34,200
Speaker 1: eight gigabytes in there. You just don't know about this

638
00:37:34,320 --> 00:37:38,560
Speaker 1: other eight because they've been included to take into account

639
00:37:38,680 --> 00:37:43,320
Speaker 1: this issue so that one uh, this whole garbage collection

640
00:37:43,480 --> 00:37:47,200
Speaker 1: process has some space to work in and you won't

641
00:37:47,440 --> 00:37:49,520
Speaker 1: end up filling up your hard drive before it can

642
00:37:50,040 --> 00:37:55,000
Speaker 1: take advantage of that and to as cells die and

643
00:37:55,080 --> 00:37:58,080
Speaker 1: are unusable, it can open up new pages of cells

644
00:37:58,160 --> 00:38:02,439
Speaker 1: that have not been written to an x number of times.

645
00:38:02,440 --> 00:38:04,440
Speaker 1: And we're talking thousands of times here. It's not like,

646
00:38:05,239 --> 00:38:06,719
Speaker 1: you know, it's not like you're gonna fill up your

647
00:38:06,760 --> 00:38:09,080
Speaker 1: hard drive and three days later it's gonna be useless.

648
00:38:09,480 --> 00:38:12,040
Speaker 1: But well, it shouldn't be, No, it shouldn't be. But

649
00:38:12,320 --> 00:38:14,520
Speaker 1: but but you know, your mileage may vary depending upon

650
00:38:15,200 --> 00:38:20,800
Speaker 1: manufacturer and model, but the ideally it would take thousands

651
00:38:20,840 --> 00:38:24,880
Speaker 1: and thousands and thousands of times before it would become obsolete,

652
00:38:24,960 --> 00:38:28,279
Speaker 1: before it would not work anymore. And the thing is

653
00:38:28,320 --> 00:38:30,959
Speaker 1: that most of us use our computers frequently enough where

654
00:38:31,280 --> 00:38:34,279
Speaker 1: eventually that could happen. I mean, if you upgrade on

655
00:38:34,360 --> 00:38:37,360
Speaker 1: a regular basis, you may never notice this problem. But

656
00:38:37,680 --> 00:38:40,080
Speaker 1: if you don't, you might notice that your computer takes

657
00:38:40,200 --> 00:38:42,800
Speaker 1: longer to pull information from the hard drive that it

658
00:38:42,920 --> 00:38:46,560
Speaker 1: used to and you may notice that you are running

659
00:38:46,560 --> 00:38:48,680
Speaker 1: out of hard drive space when you thought that you

660
00:38:48,960 --> 00:38:52,880
Speaker 1: should really have more. Why has all that gone? And

661
00:38:52,960 --> 00:38:55,920
Speaker 1: it's because those cells are no longer viable. Well. UM.

662
00:38:56,480 --> 00:38:59,440
Speaker 1: To prevent this, the controller on the S s D

663
00:39:00,120 --> 00:39:04,239
Speaker 1: is designed to route traffic in a way that will

664
00:39:04,280 --> 00:39:09,160
Speaker 1: try to put a fairly even uh distribution, distribution of

665
00:39:09,320 --> 00:39:13,200
Speaker 1: usage across the different cells on the drive uh, thus

666
00:39:13,480 --> 00:39:18,399
Speaker 1: hopefully ensuring that no set of cells is worn down

667
00:39:18,520 --> 00:39:20,399
Speaker 1: more than the others. They're trying to put put even

668
00:39:20,480 --> 00:39:23,839
Speaker 1: wear and tear on it. Um. But the more full

669
00:39:23,960 --> 00:39:26,160
Speaker 1: you get now they're one of the complaints about S

670
00:39:26,320 --> 00:39:28,680
Speaker 1: s D s is that they seem to grow slower

671
00:39:28,800 --> 00:39:32,359
Speaker 1: as time goes on. That's because that information, uh, those

672
00:39:32,480 --> 00:39:35,960
Speaker 1: those cells are getting full of information. Those pages are

673
00:39:35,960 --> 00:39:39,879
Speaker 1: filling up. And because of the way they work um

674
00:39:40,480 --> 00:39:43,960
Speaker 1: and they have to write and rewrite blank pages at

675
00:39:43,960 --> 00:39:47,800
Speaker 1: a time, UH, it can seem to slow down because

676
00:39:47,880 --> 00:39:51,239
Speaker 1: there isn't as much space to uh for them to

677
00:39:51,440 --> 00:39:54,960
Speaker 1: the controller to route the information and regroup things into

678
00:39:55,040 --> 00:39:58,000
Speaker 1: pages uh fresh pages that can be written and rewritten

679
00:39:58,360 --> 00:40:02,279
Speaker 1: or not rewritten, but um erased and written to UM.

680
00:40:03,160 --> 00:40:05,680
Speaker 1: So UM you know that that's that's sort of a

681
00:40:07,320 --> 00:40:09,920
Speaker 1: I would say an illusion. It's not really an illusion,

682
00:40:09,960 --> 00:40:12,759
Speaker 1: but that's why it's not because the drive is uh

683
00:40:13,560 --> 00:40:16,520
Speaker 1: crapping out generally, I mean, uh yeah, I mean the

684
00:40:18,239 --> 00:40:21,440
Speaker 1: vendors for these devices generally say that they're good for

685
00:40:21,880 --> 00:40:25,200
Speaker 1: you know, tens of thousands of read write cycles, so

686
00:40:25,360 --> 00:40:28,040
Speaker 1: they should be good for for years. Of course, that

687
00:40:28,080 --> 00:40:31,560
Speaker 1: doesn't mean you shouldn't back up your hard drive. However,

688
00:40:31,760 --> 00:40:34,239
Speaker 1: it does mean that defragging like we used to do

689
00:40:34,400 --> 00:40:38,640
Speaker 1: with the magnetic traditional platter drives is not a good

690
00:40:38,680 --> 00:40:43,680
Speaker 1: idea because you're adding to writing and rewriting uh those

691
00:40:43,760 --> 00:40:47,719
Speaker 1: cells right right, You're you're effectively you are decreasing the

692
00:40:47,960 --> 00:40:51,120
Speaker 1: life span of your hard drive. And the controller really

693
00:40:51,160 --> 00:40:53,440
Speaker 1: should be doing that anyway, with the garbage collection and

694
00:40:53,760 --> 00:40:57,120
Speaker 1: um organization of that that work, so it should be

695
00:40:57,239 --> 00:40:58,640
Speaker 1: less of an issue than it used to be. And

696
00:40:58,680 --> 00:41:01,800
Speaker 1: the controller really is kind of like a very small,

697
00:41:02,000 --> 00:41:05,480
Speaker 1: very specialized computer. So in a way, you have a

698
00:41:05,520 --> 00:41:08,960
Speaker 1: computer within your computer because the controller is is taking

699
00:41:09,080 --> 00:41:12,920
Speaker 1: this information and putting it in the most uh the

700
00:41:13,560 --> 00:41:18,239
Speaker 1: the optimized format and layout. So yeah, it's uh, it's

701
00:41:18,280 --> 00:41:22,680
Speaker 1: an interesting approach using this voltage difference instead of magnetism

702
00:41:22,760 --> 00:41:27,040
Speaker 1: in order to store information, and it has become incredibly useful,

703
00:41:27,200 --> 00:41:30,080
Speaker 1: especially for things like portable electronics. I mean, it's really

704
00:41:30,200 --> 00:41:33,600
Speaker 1: decreased the size of what our electronics can be. Plus

705
00:41:33,680 --> 00:41:35,400
Speaker 1: you can go running with it and not worry about

706
00:41:35,560 --> 00:41:37,759
Speaker 1: crashing the platters on your hard drive. Yeah, this is

707
00:41:37,800 --> 00:41:40,200
Speaker 1: why back when I remember when m P three's were

708
00:41:40,480 --> 00:41:42,440
Speaker 1: MP three players were first coming out, and there was

709
00:41:42,480 --> 00:41:44,680
Speaker 1: always the argument of do you get the one with

710
00:41:44,880 --> 00:41:46,799
Speaker 1: the spinning hard drive or do you get the one

711
00:41:46,840 --> 00:41:48,600
Speaker 1: with the flash hard drive? And the flash hard drives

712
00:41:48,640 --> 00:41:51,160
Speaker 1: tended to be more expensive, but they also were the

713
00:41:51,239 --> 00:41:55,040
Speaker 1: ones you could go and exercise with and not worry

714
00:41:55,120 --> 00:41:58,560
Speaker 1: about them, you know, something skipping around or or or

715
00:41:58,680 --> 00:42:03,719
Speaker 1: corrupting a file or crashing. Um, do you would you

716
00:42:03,719 --> 00:42:06,000
Speaker 1: would you like to wrap up? Or should we mention encryption?

717
00:42:07,239 --> 00:42:10,920
Speaker 1: We could mention, well, you can mention encryption because frankly,

718
00:42:11,000 --> 00:42:14,560
Speaker 1: my my research did not cover that topic. UM okay,

719
00:42:14,680 --> 00:42:17,960
Speaker 1: well the uh UM This is another in the series

720
00:42:18,120 --> 00:42:21,520
Speaker 1: of articles UM on our technical about s s d S,

721
00:42:22,040 --> 00:42:24,640
Speaker 1: also written by Lee Hutchinson UM. And it's kind of

722
00:42:24,680 --> 00:42:30,359
Speaker 1: fascinating because uh in in the process of UM compression UM,

723
00:42:30,800 --> 00:42:34,120
Speaker 1: they go through a d D duplication phase. So it's

724
00:42:34,120 --> 00:42:36,560
Speaker 1: sort of you know, if they find two copies of

725
00:42:36,640 --> 00:42:39,759
Speaker 1: the same information, it will essentially get rid of one

726
00:42:39,960 --> 00:42:41,560
Speaker 1: so that it takes up less space in the hard

727
00:42:41,640 --> 00:42:44,560
Speaker 1: rid That's essentially how how these things are done. And

728
00:42:44,640 --> 00:42:48,759
Speaker 1: it's done in UM other types of files to UM

729
00:42:49,760 --> 00:42:53,400
Speaker 1: image and sound and uh, you know those kinds of

730
00:42:53,480 --> 00:42:55,560
Speaker 1: things that they find the same information, they can reduce

731
00:42:55,640 --> 00:42:58,120
Speaker 1: the amount of information in that file. Well, they do

732
00:42:58,239 --> 00:43:01,000
Speaker 1: that with hard drives too. But one of the interesting

733
00:43:01,080 --> 00:43:04,480
Speaker 1: things that that Hutchinson mentions is now that modern operating

734
00:43:04,520 --> 00:43:07,239
Speaker 1: systems are allowing you to encrypt your entire hard drive.

735
00:43:07,600 --> 00:43:10,360
Speaker 1: That actually makes it tougher for s s d s

736
00:43:10,520 --> 00:43:14,520
Speaker 1: because they can't de duplicate that information anymore. Because once

737
00:43:14,560 --> 00:43:17,440
Speaker 1: you encrypted a file, it has its own individual signature.

738
00:43:17,880 --> 00:43:21,239
Speaker 1: So even if they were the same exact document um,

739
00:43:21,880 --> 00:43:26,719
Speaker 1: the computer sees it as two different encrypted documents because

740
00:43:26,760 --> 00:43:30,680
Speaker 1: the encryption information is slightly different, so will no longer

741
00:43:30,800 --> 00:43:34,640
Speaker 1: recognize them as the same information. So you will see

742
00:43:34,680 --> 00:43:38,960
Speaker 1: them as completely different information. Taking Bulky and safe exactly,

743
00:43:39,200 --> 00:43:41,719
Speaker 1: so it takes up the same file will take up

744
00:43:41,760 --> 00:43:43,800
Speaker 1: twice as much space if it's been duplicated, and it

745
00:43:43,840 --> 00:43:46,799
Speaker 1: will not be deleted because there's just enough difference there,

746
00:43:47,520 --> 00:43:50,959
Speaker 1: so that it has essentially fooled the controller into thinking

747
00:43:51,440 --> 00:43:54,719
Speaker 1: it's two files, not one that's been duplicated to two

748
00:43:54,760 --> 00:43:57,840
Speaker 1: files and one now. Um. The other thing, uh, the

749
00:43:57,920 --> 00:44:01,080
Speaker 1: thing to know is, you know, these the devices are

750
00:44:01,080 --> 00:44:03,760
Speaker 1: coming down and cost um, they're showing up more often

751
00:44:04,040 --> 00:44:07,600
Speaker 1: in uh, in laptops, books and and you know, just

752
00:44:07,719 --> 00:44:09,600
Speaker 1: a few years ago, I remember that it was really

753
00:44:09,760 --> 00:44:13,480
Speaker 1: unusual to find a solid state drive in a computer,

754
00:44:13,560 --> 00:44:16,120
Speaker 1: and you were paying a premium for that if you

755
00:44:16,200 --> 00:44:18,319
Speaker 1: wanted it. And it was kind of interesting because at

756
00:44:18,360 --> 00:44:21,560
Speaker 1: the time the solid state drives, while you were paying

757
00:44:21,600 --> 00:44:25,279
Speaker 1: a premium, tended to have a smaller capacity than the

758
00:44:25,320 --> 00:44:28,080
Speaker 1: traditional hard drives did when they first started coming out. Well,

759
00:44:28,160 --> 00:44:32,839
Speaker 1: now we've seen that slowly start to change, and that's

760
00:44:32,880 --> 00:44:35,359
Speaker 1: to be expected. That's the way technology tends to work

761
00:44:35,680 --> 00:44:38,319
Speaker 1: in the market. We tend to see when it first

762
00:44:38,360 --> 00:44:41,680
Speaker 1: comes out, it tends to be pretty expensive and fairly limited.

763
00:44:42,040 --> 00:44:44,600
Speaker 1: And as it advances and we get better at the

764
00:44:45,120 --> 00:44:50,560
Speaker 1: production approach, these prices start to fall and then next

765
00:44:50,640 --> 00:44:55,520
Speaker 1: thing you know, it's everywhere. Yes, so um, you know

766
00:44:55,680 --> 00:44:59,120
Speaker 1: it seems like they're they're becoming more common. Um, you know,

767
00:44:59,280 --> 00:45:03,080
Speaker 1: even even in run of the mill laptops. However, um,

768
00:45:03,239 --> 00:45:06,560
Speaker 1: you know, the cost is still not as as inexpensive

769
00:45:06,719 --> 00:45:09,920
Speaker 1: as traditional hard drives UM. And you know you do

770
00:45:10,080 --> 00:45:13,480
Speaker 1: have those trade offs to be made versus the traditional

771
00:45:13,560 --> 00:45:15,600
Speaker 1: So if you were getting, say you really wanted a

772
00:45:15,680 --> 00:45:19,719
Speaker 1: nice workstation to use at home, you already have a laptop, UM,

773
00:45:19,840 --> 00:45:21,600
Speaker 1: and you were choosing whether you wanted to spend that

774
00:45:21,680 --> 00:45:25,560
Speaker 1: extra hundred dollars two hundred dollars for an SSD, you

775
00:45:25,640 --> 00:45:28,440
Speaker 1: know it does you would get some uh savings in

776
00:45:28,760 --> 00:45:32,440
Speaker 1: in UH cost if you went with the magnetic drive, UM,

777
00:45:32,760 --> 00:45:35,840
Speaker 1: but you would trade off speed for that and the

778
00:45:35,920 --> 00:45:38,800
Speaker 1: number of read write cycles. Of course, magnetic drives have

779
00:45:38,920 --> 00:45:42,799
Speaker 1: their own idiosyncrasies and you may or may not lose

780
00:45:42,840 --> 00:45:45,680
Speaker 1: your your hard drive, right. Yeah. No, it's not not

781
00:45:45,840 --> 00:45:49,479
Speaker 1: to say that that the older hard drives are any better.

782
00:45:49,600 --> 00:45:53,839
Speaker 1: They just have a different set of pros and cons. Yeah, exactly. So, yeah,

783
00:45:53,880 --> 00:45:56,719
Speaker 1: it all depends on what your use case scenario is.

784
00:45:56,880 --> 00:45:59,600
Speaker 1: And I mean, like, i have machines at home that

785
00:46:00,440 --> 00:46:02,920
Speaker 1: of both types. So I've got machines that have a

786
00:46:02,960 --> 00:46:05,279
Speaker 1: spinning hard drive, I gut machines that have solid state drives.

787
00:46:05,320 --> 00:46:07,640
Speaker 1: I have an external drive. That's a solid state drive

788
00:46:07,719 --> 00:46:10,360
Speaker 1: that I used for backups. Uh. You know, there's a

789
00:46:10,400 --> 00:46:13,560
Speaker 1: lot of different ways of going about this, and I

790
00:46:14,120 --> 00:46:18,680
Speaker 1: think that both approaches have their own advantages and disadvantages

791
00:46:18,760 --> 00:46:21,279
Speaker 1: that will apply to you based upon the way you

792
00:46:21,400 --> 00:46:24,080
Speaker 1: use your machines. So that's always a good thing to

793
00:46:24,120 --> 00:46:26,800
Speaker 1: think about. Uh. And it may even be that to you.

794
00:46:26,960 --> 00:46:29,400
Speaker 1: It doesn't really matter other than maybe the fact that

795
00:46:29,520 --> 00:46:31,279
Speaker 1: you can get a solid state drive with a small

796
00:46:31,320 --> 00:46:34,040
Speaker 1: reform factor than you could if it were a physical

797
00:46:34,120 --> 00:46:36,200
Speaker 1: hard drive, you know, the mechanical hard drive. I should

798
00:46:36,200 --> 00:46:39,360
Speaker 1: say mechanical, not physical, because they're both physical. It's not

799
00:46:39,680 --> 00:46:43,440
Speaker 1: it's not a virtual hard drive. Um. So yeah, I

800
00:46:43,480 --> 00:46:46,399
Speaker 1: mean that's it's all up to the way you use

801
00:46:46,440 --> 00:46:48,799
Speaker 1: your machines and what you what your personal preferences are,

802
00:46:48,840 --> 00:46:50,799
Speaker 1: and I guess what your budget is as well. Yeah.

803
00:46:51,320 --> 00:46:54,840
Speaker 1: Um yeah, I'll never forget. I knew that solid state

804
00:46:55,120 --> 00:46:58,320
Speaker 1: memory was going to be a big deal. This flash

805
00:46:58,400 --> 00:47:03,719
Speaker 1: based memory sticks solid state drives slightly different, but I

806
00:47:03,800 --> 00:47:06,239
Speaker 1: knew it was gonna be a big deal when we

807
00:47:06,360 --> 00:47:09,040
Speaker 1: went to I guess is when I went to C E. S.

808
00:47:09,080 --> 00:47:10,719
Speaker 1: It was the year after the two of us went

809
00:47:10,800 --> 00:47:12,920
Speaker 1: because the year that Chris and I both went to

810
00:47:13,000 --> 00:47:16,120
Speaker 1: c S, we picked up lots and lots of press

811
00:47:16,239 --> 00:47:19,400
Speaker 1: kits that were either paper press kits or c D

812
00:47:20,120 --> 00:47:24,000
Speaker 1: comic this space. And then the next year I started

813
00:47:24,080 --> 00:47:30,560
Speaker 1: seeing companies produce their press kits on USB thumb drives.

814
00:47:31,200 --> 00:47:33,680
Speaker 1: And that's when I was I thought, Okay, this is

815
00:47:33,719 --> 00:47:35,839
Speaker 1: a big enough deal, because now it's cheap enough where

816
00:47:35,960 --> 00:47:41,160
Speaker 1: these companies can produce thousands of these things for an exhibition.

817
00:47:41,239 --> 00:47:43,960
Speaker 1: Because you've got lots and lots of people at c S,

818
00:47:44,080 --> 00:47:46,400
Speaker 1: you have to produce tons and tons of these not

819
00:47:46,600 --> 00:47:49,120
Speaker 1: literally tons and tons, but lots and lots of these

820
00:47:49,480 --> 00:47:52,279
Speaker 1: thumb drives, uh, in order to give them out to

821
00:47:52,320 --> 00:47:55,120
Speaker 1: all the people who stopped by. Of course, nowadays they

822
00:47:55,160 --> 00:47:57,960
Speaker 1: don't even do that anymore. Now you get a card

823
00:47:58,080 --> 00:47:59,759
Speaker 1: that has a U R L and you go to

824
00:48:00,160 --> 00:48:02,799
Speaker 1: a website that has the press release, which is even

825
00:48:02,920 --> 00:48:05,920
Speaker 1: better really, although it does mean that I don't end

826
00:48:06,040 --> 00:48:10,200
Speaker 1: up with lots of thumb drives that I can use

827
00:48:10,320 --> 00:48:14,040
Speaker 1: once I erase the VATA that's on there. You mean

828
00:48:14,080 --> 00:48:17,200
Speaker 1: you haven't moved everything into the cloud. Now I'm working

829
00:48:17,280 --> 00:48:19,719
Speaker 1: on it, but you know there, here's the other thing,

830
00:48:19,719 --> 00:48:21,600
Speaker 1: about the cloud. I mean, that's a totally different discussions.

831
00:48:21,640 --> 00:48:23,200
Speaker 1: I'm gonna I'm gonna drop this. I was about to

832
00:48:23,239 --> 00:48:24,960
Speaker 1: go off on a cloud rant about how all my

833
00:48:25,080 --> 00:48:27,920
Speaker 1: information is in different pockets in the cloud. That's my

834
00:48:28,080 --> 00:48:32,360
Speaker 1: problem now, So let's let's that's a totally different podcast,

835
00:48:32,440 --> 00:48:34,719
Speaker 1: which I'm sure we'll do. And we've talked about cloud

836
00:48:34,800 --> 00:48:37,960
Speaker 1: storage in the past anyway, so we're gonna wrap this up. Guys.

837
00:48:38,400 --> 00:48:40,600
Speaker 1: I do recommend you go to ours Technica and look

838
00:48:40,640 --> 00:48:43,520
Speaker 1: at those articles if you are interested in solid state

839
00:48:43,600 --> 00:48:46,359
Speaker 1: drives and what makes them work. And that wraps up

840
00:48:46,520 --> 00:48:51,200
Speaker 1: another episode, another classic episode of tech Stuff, and hope

841
00:48:51,200 --> 00:48:54,440
Speaker 1: you guys enjoyed it. I really miss having those conversations

842
00:48:54,480 --> 00:48:56,960
Speaker 1: with Chris. They were always a lot of fun. If

843
00:48:57,000 --> 00:48:59,319
Speaker 1: you guys have suggestions for topics that I should cover

844
00:48:59,440 --> 00:49:02,400
Speaker 1: in future or episodes of tech Stuff, send me an

845
00:49:02,440 --> 00:49:06,080
Speaker 1: email the addresses tech Stuff at how stuff works dot com,

846
00:49:06,600 --> 00:49:10,440
Speaker 1: or drop on by our website that's tech Stuff podcast

847
00:49:10,560 --> 00:49:12,840
Speaker 1: dot com. There you're going to find an archive of

848
00:49:13,040 --> 00:49:15,560
Speaker 1: all of our past episodes. It's fully searchable, so you

849
00:49:15,600 --> 00:49:18,879
Speaker 1: can find out if I've covered a specific topic that way,

850
00:49:19,560 --> 00:49:22,720
Speaker 1: and you can also find links to our social media

851
00:49:23,280 --> 00:49:27,640
Speaker 1: presence on places like Facebook and Twitter. You can also

852
00:49:27,880 --> 00:49:30,400
Speaker 1: find a link to our online store, where every purchase

853
00:49:30,480 --> 00:49:32,480
Speaker 1: you make goes to help the show and we greatly

854
00:49:32,520 --> 00:49:36,040
Speaker 1: appreciate it, and I'll talk to you again really soon.

855
00:49:40,480 --> 00:49:42,680
Speaker 1: Text Stuff is a production of I Heart Radio's How

856
00:49:42,760 --> 00:49:46,120
Speaker 1: Stuff Works. For more podcasts from my heart Radio, visit

857
00:49:46,160 --> 00:49:49,160
Speaker 1: the i heart Radio app, Apple Podcasts, or wherever you

858
00:49:49,320 --> 00:49:50,640
Speaker 1: listen to your favorite shows.