1
00:00:00,060 --> 00:00:01,080
In this lesson,

2
00:00:01,080 --> 00:00:03,390
we start designing redundant networks.

3
00:00:03,390 --> 00:00:04,680
First, you need to ask yourself,

4
00:00:04,680 --> 00:00:06,689
are you going to use redundancy in the network,

5
00:00:06,689 --> 00:00:09,390
and if so, where and how?

6
00:00:09,390 --> 00:00:12,360
So are you going to do it from a module or a parts perspective?

7
00:00:12,360 --> 00:00:14,460
For instance, are you going to have multiple power supplies,

8
00:00:14,460 --> 00:00:17,940
multiple network interface devices, multiple hard drives,

9
00:00:17,940 --> 00:00:20,310
or are you going to look at more from a chassis redundancy

10
00:00:20,310 --> 00:00:23,520
and have two sets of routers or two sets of switches?

11
00:00:23,520 --> 00:00:25,350
These are things you have to think about.

12
00:00:25,350 --> 00:00:27,120
Which one of these are you going to use,

13
00:00:27,120 --> 00:00:29,760
because each one is going to affect the cost of your network

14
00:00:29,760 --> 00:00:31,440
based on the decisions you make.

15
00:00:31,440 --> 00:00:33,180
You have to be able to make a good business case

16
00:00:33,180 --> 00:00:35,370
for which one you're going to use and why.

17
00:00:35,370 --> 00:00:37,110
For instance, if you could just have a second network

18
00:00:37,110 --> 00:00:39,360
interface card or a second power supply,

19
00:00:39,360 --> 00:00:41,340
that's going to be a lot cheaper than having to have

20
00:00:41,340 --> 00:00:44,160
an entire switch or an entire extra router there.

21
00:00:44,160 --> 00:00:45,900
Now, each of those switches and routers,

22
00:00:45,900 --> 00:00:48,720
some of these can cost 3 or 4 or $5,000,

23
00:00:48,720 --> 00:00:50,130
and so it might be a lot cheaper

24
00:00:50,130 --> 00:00:51,597
to have a redundant power supply, right?

25
00:00:51,597 --> 00:00:53,280
And so these are the things you have to think about

26
00:00:53,280 --> 00:00:55,500
and weigh as you're building your networks.

27
00:00:55,500 --> 00:00:58,200
Another thing you have to think about is software redundancy

28
00:00:58,200 --> 00:01:00,450
and which features of those are going to be appropriate?

29
00:01:00,450 --> 00:01:03,060
Sometimes you can solve a lot of these redundancy problems

30
00:01:03,060 --> 00:01:05,550
by using software as opposed to hardware.

31
00:01:05,550 --> 00:01:08,130
For example, if you have a virtual network setup,

32
00:01:08,130 --> 00:01:09,570
you could just put in a virtual switch

33
00:01:09,570 --> 00:01:11,010
or a virtual router in there,

34
00:01:11,010 --> 00:01:12,240
and that way you don't have to bring another

35
00:01:12,240 --> 00:01:14,070
real router or real switch in.

36
00:01:14,070 --> 00:01:16,260
That can save you a lot of money.

37
00:01:16,260 --> 00:01:18,630
There's also a lot of other software solutions out there,

38
00:01:18,630 --> 00:01:20,340
like a software RAID, that will give you

39
00:01:20,340 --> 00:01:22,440
additional redundancy for your storage devices

40
00:01:22,440 --> 00:01:25,170
as opposed to putting in an extra hard drive chassis

41
00:01:25,170 --> 00:01:27,900
or another RAID array or storage area network.

42
00:01:27,900 --> 00:01:29,130
Also, these are the types of things

43
00:01:29,130 --> 00:01:30,150
you have to be thinking about

44
00:01:30,150 --> 00:01:32,010
as you're building out your network, right?

45
00:01:32,010 --> 00:01:33,450
When you think about your protocols,

46
00:01:33,450 --> 00:01:35,580
what protocol characteristics are going to affect

47
00:01:35,580 --> 00:01:36,990
your design requirements?

48
00:01:36,990 --> 00:01:39,150
This is really important if you're designing things

49
00:01:39,150 --> 00:01:39,990
and you're using something like

50
00:01:39,990 --> 00:01:42,240
TCP versus UDP in your designs,

51
00:01:42,240 --> 00:01:45,180
because TCP has that additional redundancy

52
00:01:45,180 --> 00:01:47,820
by resending packets where UDP doesn't.

53
00:01:47,820 --> 00:01:50,190
This is something you have to consider as well.

54
00:01:50,190 --> 00:01:52,050
As you design all these different things,

55
00:01:52,050 --> 00:01:54,600
all these different factors are going to work together

56
00:01:54,600 --> 00:01:55,890
just like gears.

57
00:01:55,890 --> 00:01:59,310
Each one turns another, each one is going to feed another one,

58
00:01:59,310 --> 00:02:01,237
and the more reliability and availability

59
00:02:01,237 --> 00:02:02,070
you get in your networks

60
00:02:02,070 --> 00:02:04,380
by adding all these components together.

61
00:02:04,380 --> 00:02:05,340
In addition to all this,

62
00:02:05,340 --> 00:02:06,810
there are other design considerations

63
00:02:06,810 --> 00:02:08,550
that we have to think about as well.

64
00:02:08,550 --> 00:02:10,560
Like what redundancy features should we use

65
00:02:10,560 --> 00:02:13,470
in terms of powering the infrastructure devices?

66
00:02:13,470 --> 00:02:15,030
Are we going to have internal power supplies

67
00:02:15,030 --> 00:02:17,370
and have two of those and have them redundant?

68
00:02:17,370 --> 00:02:19,860
Or, are we going to have battery backups or UPSes?

69
00:02:19,860 --> 00:02:21,330
Are we going to have generators?

70
00:02:21,330 --> 00:02:23,430
All of these things are things you have to think about.

71
00:02:23,430 --> 00:02:25,530
And I don't have necessarily the right answers for you,

72
00:02:25,530 --> 00:02:28,500
because it all comes down to a case-by-case basis.

73
00:02:28,500 --> 00:02:30,240
Every network is going to be different

74
00:02:30,240 --> 00:02:31,740
and everyone has its own needs

75
00:02:31,740 --> 00:02:34,200
and its own business case associated with it.

76
00:02:34,200 --> 00:02:36,060
The network that I had at former employers

77
00:02:36,060 --> 00:02:38,100
were serving hundreds of thousands of clients,

78
00:02:38,100 --> 00:02:39,660
and those are vastly different

79
00:02:39,660 --> 00:02:40,770
than the ones that are servicing

80
00:02:40,770 --> 00:02:42,300
my training company right now

81
00:02:42,300 --> 00:02:44,130
with just a handful of employees.

82
00:02:44,130 --> 00:02:45,960
Because when you're dealing with your network design

83
00:02:45,960 --> 00:02:47,070
and your redundancies,

84
00:02:47,070 --> 00:02:49,800
you have to think about the business case first.

85
00:02:49,800 --> 00:02:51,000
Each one is going to be different

86
00:02:51,000 --> 00:02:53,610
based on your needs and your considerations.

87
00:02:53,610 --> 00:02:55,290
What redundancy features should be used

88
00:02:55,290 --> 00:02:57,780
to maintain the environmental conditions of your space?

89
00:02:57,780 --> 00:03:00,030
If you have good power and space and cooling,

90
00:03:00,030 --> 00:03:01,380
you need to make sure that you're thinking about

91
00:03:01,380 --> 00:03:04,110
air conditioning and do you have one unit or two?

92
00:03:04,110 --> 00:03:05,700
Do you have generators on site?

93
00:03:05,700 --> 00:03:08,550
Do you have additional thermal heating or thermal cooling?

94
00:03:08,550 --> 00:03:10,680
All of these things are things you have to think about.

95
00:03:10,680 --> 00:03:12,510
What do you do when power goes down?

96
00:03:12,510 --> 00:03:13,500
What are some of those things

97
00:03:13,500 --> 00:03:14,333
that you're going to have to deal with

98
00:03:14,333 --> 00:03:15,750
if you're running a server farm

99
00:03:15,750 --> 00:03:18,270
that has to have units running all the time,

100
00:03:18,270 --> 00:03:19,620
because it can't afford to go down

101
00:03:19,620 --> 00:03:22,380
because it's going to affect thousands and thousands of people

102
00:03:22,380 --> 00:03:25,020
instead of just your one office with 20 people?

103
00:03:25,020 --> 00:03:26,610
All of these are things you have to consider

104
00:03:26,610 --> 00:03:27,840
as you think about it.

105
00:03:27,840 --> 00:03:29,760
In my office, we made the decision

106
00:03:29,760 --> 00:03:31,860
that one air conditioning unit was enough,

107
00:03:31,860 --> 00:03:34,380
because if it goes down, we might just not work today

108
00:03:34,380 --> 00:03:35,760
and we'll come to work tomorrow.

109
00:03:35,760 --> 00:03:36,990
We can get over that.

110
00:03:36,990 --> 00:03:38,400
But in a server farm,

111
00:03:38,400 --> 00:03:40,770
we need to make sure we have multiple air conditioners,

112
00:03:40,770 --> 00:03:41,670
because if that goes down,

113
00:03:41,670 --> 00:03:43,710
it can actually burn up all the components, right?

114
00:03:43,710 --> 00:03:46,020
So we have to have additional power and space and cooling

115
00:03:46,020 --> 00:03:47,310
that are fully redundant

116
00:03:47,310 --> 00:03:49,050
because of that server infrastructure

117
00:03:49,050 --> 00:03:50,520
that we're supporting there.

118
00:03:50,520 --> 00:03:53,400
These are the things you have to balance in your practices.

119
00:03:53,400 --> 00:03:55,590
And so when you start looking at the best practices,

120
00:03:55,590 --> 00:03:57,600
I want you to examine your technical goals

121
00:03:57,600 --> 00:03:59,250
and your operational goals.

122
00:03:59,250 --> 00:04:00,600
Now, what I mean by that is

123
00:04:00,600 --> 00:04:02,700
what is the function of this network?

124
00:04:02,700 --> 00:04:04,680
What are you actually trying to accomplish?

125
00:04:04,680 --> 00:04:06,390
Are you trying to get to 90% uptime?

126
00:04:06,390 --> 00:04:08,880
Or 95% or 99%?

127
00:04:08,880 --> 00:04:10,440
Or are you going for that gold standard

128
00:04:10,440 --> 00:04:12,570
of five nines of availability?

129
00:04:12,570 --> 00:04:14,790
Every company has a different technical goal,

130
00:04:14,790 --> 00:04:16,440
and that technical goal is going to determine

131
00:04:16,440 --> 00:04:17,970
the design of your network.

132
00:04:17,970 --> 00:04:19,410
And you need to identify that

133
00:04:19,410 --> 00:04:20,850
inside of your budgeting as well,

134
00:04:20,850 --> 00:04:23,340
because funding these high availability-features

135
00:04:23,340 --> 00:04:24,900
is really expensive.

136
00:04:24,900 --> 00:04:27,120
As I said, if I want to put a second router in there,

137
00:04:27,120 --> 00:04:30,300
that might cost me another 3,000 or $5,000.

138
00:04:30,300 --> 00:04:32,640
In my own personal network, we have a file server,

139
00:04:32,640 --> 00:04:34,530
and it's a small NAS device.

140
00:04:34,530 --> 00:04:36,540
We're not comfortable just having all of those

141
00:04:36,540 --> 00:04:37,620
on our devices there.

142
00:04:37,620 --> 00:04:38,940
So we decided we weren't comfortable

143
00:04:38,940 --> 00:04:41,580
having all of our file storage on a single hard drive,

144
00:04:41,580 --> 00:04:43,530
and we built this NAS array instead.

145
00:04:43,530 --> 00:04:45,240
So if one of those drives goes out,

146
00:04:45,240 --> 00:04:47,700
we have three others that are carrying the load.

147
00:04:47,700 --> 00:04:49,050
This is the idea here.

148
00:04:49,050 --> 00:04:51,750
Now, eventually we decided we didn't need that NAS anymore,

149
00:04:51,750 --> 00:04:54,990
and so we replaced that NAS enclosure with a full RAID 5.

150
00:04:54,990 --> 00:04:56,880
Later on, we took that full RAID 5

151
00:04:56,880 --> 00:04:58,650
and we switched it over to a cloud server

152
00:04:58,650 --> 00:04:59,760
that has redundant backups

153
00:04:59,760 --> 00:05:01,500
in two different cloud environments.

154
00:05:01,500 --> 00:05:03,210
And so all of these things work together

155
00:05:03,210 --> 00:05:04,590
based on our decisions.

156
00:05:04,590 --> 00:05:06,060
But as we moved up that scale

157
00:05:06,060 --> 00:05:07,650
and got more and more redundancy,

158
00:05:07,650 --> 00:05:09,630
we have more and more cost associated.

159
00:05:09,630 --> 00:05:10,830
It was a lot cheaper just to have

160
00:05:10,830 --> 00:05:13,230
an eight terabyte hard drive with all of our files on it.

161
00:05:13,230 --> 00:05:14,400
Then we went to a NAS array,

162
00:05:14,400 --> 00:05:16,140
and that cost two or three times that money.

163
00:05:16,140 --> 00:05:17,370
Then we went to a full RAID 5,

164
00:05:17,370 --> 00:05:18,723
and that cost a couple more times than that.

165
00:05:18,723 --> 00:05:21,480
Then we went to a cloud and we have to pay more for that.

166
00:05:21,480 --> 00:05:23,190
Remember, all your decisions here

167
00:05:23,190 --> 00:05:24,840
are going to cost you more money,

168
00:05:24,840 --> 00:05:27,960
but if it's worth it to you, that would be important, right?

169
00:05:27,960 --> 00:05:29,880
And so these are the things you have to balance

170
00:05:29,880 --> 00:05:32,070
as you're designing these fully redundant networks

171
00:05:32,070 --> 00:05:34,260
based on those technical goals.

172
00:05:34,260 --> 00:05:35,430
You also need to categorize

173
00:05:35,430 --> 00:05:38,070
all of your business applications into profiles

174
00:05:38,070 --> 00:05:39,660
to help with this redundancy mission

175
00:05:39,660 --> 00:05:41,550
that you're trying to go and accomplish here.

176
00:05:41,550 --> 00:05:42,383
This will really help you

177
00:05:42,383 --> 00:05:45,090
as you start going into the quality of service as well.

178
00:05:45,090 --> 00:05:46,350
Now, if I said, for instance,

179
00:05:46,350 --> 00:05:50,010
that web is considered category one, email is category two,

180
00:05:50,010 --> 00:05:52,050
and streaming video is going to be category three,

181
00:05:52,050 --> 00:05:53,520
then we can apply profiles

182
00:05:53,520 --> 00:05:55,140
and give certain levels of service

183
00:05:55,140 --> 00:05:56,760
to each of those categories.

184
00:05:56,760 --> 00:05:58,890
Now we'll talk specifically of how that works

185
00:05:58,890 --> 00:06:01,680
when we talk about quality of service in a future lesson.

186
00:06:01,680 --> 00:06:02,700
Another thing we want to do

187
00:06:02,700 --> 00:06:04,320
is establish performance standards

188
00:06:04,320 --> 00:06:06,480
for our high-availability networks.

189
00:06:06,480 --> 00:06:08,850
What are the standards that we're going to have to have?

190
00:06:08,850 --> 00:06:10,140
These standards are going to drive

191
00:06:10,140 --> 00:06:12,030
how success is measured for us.

192
00:06:12,030 --> 00:06:14,430
And in the case of my file server for instance,

193
00:06:14,430 --> 00:06:17,010
we measure success as it being up and available

194
00:06:17,010 --> 00:06:19,140
when my video editors need access it

195
00:06:19,140 --> 00:06:20,490
and that they don't lose data,

196
00:06:20,490 --> 00:06:21,960
because if we lost all our files,

197
00:06:21,960 --> 00:06:23,580
that would be bad for us, right?

198
00:06:23,580 --> 00:06:25,050
Those are two metrics that we have,

199
00:06:25,050 --> 00:06:27,840
and we have numbers associated with each of those things.

200
00:06:27,840 --> 00:06:29,640
In other organizations, we measure it

201
00:06:29,640 --> 00:06:32,490
based on the uptime of the entire end-to-end service.

202
00:06:32,490 --> 00:06:35,130
So if a client can't get out to the internet for an ISP,

203
00:06:35,130 --> 00:06:36,390
that would be a bad thing.

204
00:06:36,390 --> 00:06:37,980
That's one of their measurements.

205
00:06:37,980 --> 00:06:40,350
Now the other one might be what is their uptime?

206
00:06:40,350 --> 00:06:41,910
All these performance standards are developed

207
00:06:41,910 --> 00:06:44,220
through metrics and key performance indicators.

208
00:06:44,220 --> 00:06:45,390
If you're using something like ITIL

209
00:06:45,390 --> 00:06:47,310
as your IT service management standards,

210
00:06:47,310 --> 00:06:48,600
this is what you're going to be doing

211
00:06:48,600 --> 00:06:49,620
as you're trying to run those

212
00:06:49,620 --> 00:06:51,780
inside your organization as well.

213
00:06:51,780 --> 00:06:54,930
Finally, here we want to define how we manage and measure

214
00:06:54,930 --> 00:06:57,420
the high-availability solutions for ourselves.

215
00:06:57,420 --> 00:07:00,180
Metrics are going to be really useful to quantify success

216
00:07:00,180 --> 00:07:02,310
if you develop those metrics correctly.

217
00:07:02,310 --> 00:07:05,370
Decision makers and leaders love seeing metrics.

218
00:07:05,370 --> 00:07:07,680
They love seeing charts and seeing the performance

219
00:07:07,680 --> 00:07:09,150
and how it's going up over time,

220
00:07:09,150 --> 00:07:10,680
and how our availability is going up,

221
00:07:10,680 --> 00:07:12,450
and how our costs are going down.

222
00:07:12,450 --> 00:07:13,830
Those are all good things.

223
00:07:13,830 --> 00:07:15,270
But if you don't know what you're measuring

224
00:07:15,270 --> 00:07:16,590
or why you're measuring it,

225
00:07:16,590 --> 00:07:18,660
it really goes back to your performance standards.

226
00:07:18,660 --> 00:07:19,920
Then, these are the kind of things

227
00:07:19,920 --> 00:07:21,690
that are wasting your time with metrics.

228
00:07:21,690 --> 00:07:23,580
A lot of people measure a lot of things

229
00:07:23,580 --> 00:07:24,600
and they don't really tell you

230
00:07:24,600 --> 00:07:26,700
if you're getting the outcome you're wanting.

231
00:07:26,700 --> 00:07:28,440
I want to make sure that you think about

232
00:07:28,440 --> 00:07:31,020
how you decide on what metrics you're going to use.

233
00:07:31,020 --> 00:07:33,270
Now, we've covered a lot of different design criteria

234
00:07:33,270 --> 00:07:36,240
in this lesson, but the real big takeaway here

235
00:07:36,240 --> 00:07:38,010
that I want you to think about is this.

236
00:07:38,010 --> 00:07:39,690
If you have an existing network,

237
00:07:39,690 --> 00:07:41,280
you can add availability to it

238
00:07:41,280 --> 00:07:43,080
and you can add redundancy to it.

239
00:07:43,080 --> 00:07:44,790
You can retrofit stuff in,

240
00:07:44,790 --> 00:07:46,620
but it's going to cost you a lot more time

241
00:07:46,620 --> 00:07:48,180
and a lot more money.

242
00:07:48,180 --> 00:07:50,070
It is much, much cheaper

243
00:07:50,070 --> 00:07:52,140
to design this stuff early in the process

244
00:07:52,140 --> 00:07:54,210
when you start building a network from scratch.

245
00:07:54,210 --> 00:07:55,980
So, if you're designing a network

246
00:07:55,980 --> 00:07:58,410
and you're asked early on what kind of things you need,

247
00:07:58,410 --> 00:08:00,180
I want you to think about all these things

248
00:08:00,180 --> 00:08:02,490
a redundancy in your initial design.

249
00:08:02,490 --> 00:08:05,700
Adding them in early is going to save you a lot of money.

250
00:08:05,700 --> 00:08:07,950
Every project has three main factors,

251
00:08:07,950 --> 00:08:10,140
time, cost, and quality.

252
00:08:10,140 --> 00:08:11,820
And usually one of these things

253
00:08:11,820 --> 00:08:14,790
is going to suffer at the expense of the other two.

254
00:08:14,790 --> 00:08:17,040
For example, if I asked you to build me a network

255
00:08:17,040 --> 00:08:18,390
and I want it to be fully redundant

256
00:08:18,390 --> 00:08:21,060
and available by tomorrow, could you do it?

257
00:08:21,060 --> 00:08:24,450
Well, maybe, but it's probably going to cost me a lot of money,

258
00:08:24,450 --> 00:08:26,250
and because I give you very little time,

259
00:08:26,250 --> 00:08:27,780
it's going to cost me even more,

260
00:08:27,780 --> 00:08:29,910
or your quality is going to suffer.

261
00:08:29,910 --> 00:08:32,309
So you could do it good, you could do it quick,

262
00:08:32,309 --> 00:08:35,070
or you could do it cheap, but you can't do all three.

263
00:08:35,070 --> 00:08:37,770
It's always going to be a trade off between these three things,

264
00:08:37,770 --> 00:08:39,510
and I want you to remember as you're out there

265
00:08:39,510 --> 00:08:41,070
and you're designing networks,

266
00:08:41,070 --> 00:08:43,350
you need to make sure you're thinking about your redundancy

267
00:08:43,350 --> 00:08:45,660
and your availability and your reliability,

268
00:08:45,660 --> 00:08:48,240
because often that quality is going to suffer

269
00:08:48,240 --> 00:08:50,070
in favor of getting things out quicker

270
00:08:50,070 --> 00:08:52,570
or getting things out cheaper.