1 1 00:00:00,400 --> 00:00:02,600 In the last lesson I showed you a couple of diagrams 2 2 00:00:02,600 --> 00:00:04,150 of redundant networks, 3 3 00:00:04,150 --> 00:00:06,670 but one of the things we had to think about in this lesson 4 4 00:00:06,670 --> 00:00:08,160 are the considerations we have 5 5 00:00:08,160 --> 00:00:10,640 when we start designing these redundant networks. 6 6 00:00:10,640 --> 00:00:11,970 First you need to ask yourself, 7 7 00:00:11,970 --> 00:00:13,970 are you going to use redundancy in the network, 8 8 00:00:13,970 --> 00:00:16,670 and if so, where, and how? 9 9 00:00:16,670 --> 00:00:19,590 So are you going to do it from a module or a parts perspective? 10 10 00:00:19,590 --> 00:00:21,750 For instance, are you going to have multiple power supplies, 11 11 00:00:21,750 --> 00:00:25,450 multiple network interface devices, multiple hard drives, 12 12 00:00:25,450 --> 00:00:27,840 or are you going to look at it more from a chassis redundancy 13 13 00:00:27,840 --> 00:00:31,040 and have two sets of routers or two sets of switches? 14 14 00:00:31,040 --> 00:00:32,555 These are things you have to think about. 15 15 00:00:32,555 --> 00:00:34,620 Which one of these are you going to use, 16 16 00:00:34,620 --> 00:00:36,299 because each one is going to affect the cost 17 17 00:00:36,299 --> 00:00:38,960 of your network, based on the decisions you make. 18 18 00:00:38,960 --> 00:00:40,710 You have to be able to make a good business case 19 19 00:00:40,710 --> 00:00:42,870 for which one you're going to use, and why. 20 20 00:00:42,870 --> 00:00:43,960 For instance, if you could just have 21 21 00:00:43,960 --> 00:00:46,582 a second network interface card or a second power supply, 22 22 00:00:46,582 --> 00:00:48,430 that's going to be a lot cheaper 23 23 00:00:48,430 --> 00:00:49,810 than having to have an entire switch 24 24 00:00:49,810 --> 00:00:51,680 or an entire extra router there. 25 25 00:00:51,680 --> 00:00:53,420 Now, each of those switches and routers, 26 26 00:00:53,420 --> 00:00:56,240 some of these can cost 3 or 4 or $5,000, 27 27 00:00:56,240 --> 00:00:57,660 and so it might be a lot cheaper 28 28 00:00:57,660 --> 00:00:59,157 to have a redundant power supply, right, 29 29 00:00:59,157 --> 00:01:00,820 and so these are the things you have to think about 30 30 00:01:00,820 --> 00:01:03,040 and weigh as you're building your networks. 31 31 00:01:03,040 --> 00:01:04,101 Another thing you have to think about 32 32 00:01:04,101 --> 00:01:05,730 is software redundancy, 33 33 00:01:05,730 --> 00:01:07,960 and which features of those are going to be appropriate. 34 34 00:01:07,960 --> 00:01:10,590 Sometimes you can solve a lot of these redundancy problems 35 35 00:01:10,590 --> 00:01:13,090 by using software as opposed to hardware. 36 36 00:01:13,090 --> 00:01:15,660 For example, if you have a virtual network setup, 37 37 00:01:15,660 --> 00:01:17,120 you could just put in a virtual switch 38 38 00:01:17,120 --> 00:01:18,510 or a virtual router in there, 39 39 00:01:18,510 --> 00:01:19,480 and that way you don't have to bring 40 40 00:01:19,480 --> 00:01:21,600 another real router or real switch in, 41 41 00:01:21,600 --> 00:01:23,810 that can save you a lot of money. 42 42 00:01:23,810 --> 00:01:26,180 There's also a lot of other software solutions out there, 43 43 00:01:26,180 --> 00:01:27,500 like a software RAID, 44 44 00:01:27,500 --> 00:01:28,700 that will give you additional redundancy 45 45 00:01:28,700 --> 00:01:29,990 for your storage devices, 46 46 00:01:29,990 --> 00:01:32,700 as opposed to putting in an extra hard drive chassis, 47 47 00:01:32,700 --> 00:01:35,440 or another RAID array or storage area network. 48 48 00:01:35,440 --> 00:01:36,670 Also, these are the types of things 49 49 00:01:36,670 --> 00:01:37,503 you have to be thinking about 50 50 00:01:37,503 --> 00:01:39,570 as you're building out your network, right? 51 51 00:01:39,570 --> 00:01:41,010 When you think about your protocols, 52 52 00:01:41,010 --> 00:01:42,400 what protocol characteristics 53 53 00:01:42,400 --> 00:01:44,540 are going to affect your design requirements? 54 54 00:01:44,540 --> 00:01:46,650 This is really important if you're designing things, 55 55 00:01:46,650 --> 00:01:47,520 and you're using something like 56 56 00:01:47,520 --> 00:01:49,780 TCP versus UDP in your designs, 57 57 00:01:49,780 --> 00:01:52,720 because TCP has that additional redundancy 58 58 00:01:52,720 --> 00:01:55,360 by resending packets, where UDP doesn't, 59 59 00:01:55,360 --> 00:01:57,720 this is something you have to consider as well. 60 60 00:01:57,720 --> 00:01:59,580 As you design all these different things, 61 61 00:01:59,580 --> 00:02:01,792 all of these different factors are going to work together, 62 62 00:02:01,792 --> 00:02:04,710 just like gears, and each one turns another, 63 63 00:02:04,710 --> 00:02:06,750 and each one is going to feed another one, 64 64 00:02:06,750 --> 00:02:08,650 and the more reliability and availability 65 65 00:02:08,650 --> 00:02:09,530 you get in your networks 66 66 00:02:09,530 --> 00:02:11,910 by adding all these components together. 67 67 00:02:11,910 --> 00:02:12,880 In addition to all this, 68 68 00:02:12,880 --> 00:02:13,968 there are other design considerations 69 69 00:02:13,968 --> 00:02:16,090 that we have to think about as well, 70 70 00:02:16,090 --> 00:02:18,110 like what redundancy features should we use 71 71 00:02:18,110 --> 00:02:21,000 in terms of powering the infrastructure devices? 72 72 00:02:21,000 --> 00:02:22,550 Are we going to have internal power supplies 73 73 00:02:22,550 --> 00:02:24,581 and have two of those, and have them redundant? 74 74 00:02:24,581 --> 00:02:27,370 Or, are we going to have battery backups, or UPSs, 75 75 00:02:27,370 --> 00:02:28,850 are we going to have generators? 76 76 00:02:28,850 --> 00:02:30,970 All of these things are things you have to think about, 77 77 00:02:30,970 --> 00:02:33,080 and I don't have necessarily the right answers for you, 78 78 00:02:33,080 --> 00:02:36,030 because it all comes down to a case-by-case basis. 79 79 00:02:36,030 --> 00:02:37,750 Every network is going to be different, 80 80 00:02:37,750 --> 00:02:39,270 and every one has its own needs 81 81 00:02:39,270 --> 00:02:41,730 and its own business case associated with it. 82 82 00:02:41,730 --> 00:02:43,255 The network that I had at former employers 83 83 00:02:43,255 --> 00:02:45,650 were serving hundreds of thousands of clients, 84 84 00:02:45,650 --> 00:02:47,710 and those were vastly different than the ones 85 85 00:02:47,710 --> 00:02:49,850 that are servicing my training company right now, 86 86 00:02:49,850 --> 00:02:51,650 with just a handful of employees. 87 87 00:02:51,650 --> 00:02:53,500 Because when you're dealing with your network design 88 88 00:02:53,500 --> 00:02:54,600 and your redundancies, 89 89 00:02:54,600 --> 00:02:57,320 you have to think about the business case first. 90 90 00:02:57,320 --> 00:02:58,540 Each one is going to be different 91 91 00:02:58,540 --> 00:03:01,160 based on your needs and your considerations. 92 92 00:03:01,160 --> 00:03:02,850 What redundancy features should be used 93 93 00:03:02,850 --> 00:03:05,310 to maintain the environmental conditions of your space? 94 94 00:03:05,310 --> 00:03:07,580 If you have good power and space and cooling, 95 95 00:03:07,580 --> 00:03:08,413 you need to make sure 96 96 00:03:08,413 --> 00:03:09,820 that you're thinking about air conditioning, 97 97 00:03:09,820 --> 00:03:11,650 and do you have one unit or two? 98 98 00:03:11,650 --> 00:03:13,210 Do you have generators onsite? 99 99 00:03:13,210 --> 00:03:16,080 Do you have additional thermal heating or thermal cooling? 100 100 00:03:16,080 --> 00:03:18,200 All of these things are things you have to think about. 101 101 00:03:18,200 --> 00:03:20,050 What do you do when power goes down? 102 102 00:03:20,050 --> 00:03:21,020 What are some of those things 103 103 00:03:21,020 --> 00:03:21,860 that you're going to have to deal with 104 104 00:03:21,860 --> 00:03:23,290 if you're running a server farm 105 105 00:03:23,290 --> 00:03:25,760 that has to have units running all the time, 106 106 00:03:25,760 --> 00:03:27,160 because it can't afford to go down 107 107 00:03:27,160 --> 00:03:28,220 because it's going to affect 108 108 00:03:28,220 --> 00:03:29,930 thousands and thousands of people, 109 109 00:03:29,930 --> 00:03:32,540 instead of just your one office with 20 people? 110 110 00:03:32,540 --> 00:03:34,140 All of these are things you have to consider 111 111 00:03:34,140 --> 00:03:35,370 as you think about it. 112 112 00:03:35,370 --> 00:03:37,310 In my office, we made the decision 113 113 00:03:37,310 --> 00:03:39,380 that one air conditioning unit was enough, 114 114 00:03:39,380 --> 00:03:41,930 because if it goes down, we might just not work today 115 115 00:03:41,930 --> 00:03:44,510 and we'll come to work tomorrow, we can get over that. 116 116 00:03:44,510 --> 00:03:45,940 But in a server farm, 117 117 00:03:45,940 --> 00:03:48,310 we need to make sure we have multiple air conditioners, 118 118 00:03:48,310 --> 00:03:49,143 because if that goes down 119 119 00:03:49,143 --> 00:03:51,240 it can actually burn up all the components, right? 120 120 00:03:51,240 --> 00:03:53,220 So we have to have additional power and space and cooling 121 121 00:03:53,220 --> 00:03:54,840 that are fully redundant, 122 122 00:03:54,840 --> 00:03:56,226 because of that server infrastructure 123 123 00:03:56,226 --> 00:03:58,040 that we're supporting there. 124 124 00:03:58,040 --> 00:04:00,930 These are the things you have to balance in your practices. 125 125 00:04:00,930 --> 00:04:02,807 And so when you start looking at the best practices, 126 126 00:04:02,807 --> 00:04:05,120 I want you to examine your technical goals 127 127 00:04:05,120 --> 00:04:06,780 and your operational goals. 128 128 00:04:06,780 --> 00:04:08,150 Now what I mean by that is, 129 129 00:04:08,150 --> 00:04:10,230 what is the function of this network? 130 130 00:04:10,230 --> 00:04:12,200 What are you actually trying to accomplish? 131 131 00:04:12,200 --> 00:04:16,350 Are you trying to get to 90% uptime, or 95%, or 99%, 132 132 00:04:16,350 --> 00:04:17,677 or are you going for that gold standard 133 133 00:04:17,677 --> 00:04:20,100 of five nines of availability? 134 134 00:04:20,100 --> 00:04:22,200 Every company has a different technical goal, 135 135 00:04:22,200 --> 00:04:23,960 and that technical goal is going to determine 136 136 00:04:23,960 --> 00:04:25,470 the design of your network. 137 137 00:04:25,470 --> 00:04:26,519 And you need to identify that 138 138 00:04:26,519 --> 00:04:28,390 inside of your budgeting as well, 139 139 00:04:28,390 --> 00:04:30,471 because funding these high-availability features 140 140 00:04:30,471 --> 00:04:32,440 is really expensive. 141 141 00:04:32,440 --> 00:04:34,630 As I said, if I want to put a second router in there, 142 142 00:04:34,630 --> 00:04:37,820 that might cost me another 3,000 or $5,000. 143 143 00:04:37,820 --> 00:04:39,160 In my own personal network, 144 144 00:04:39,160 --> 00:04:42,050 we have a file server, and it's a small NAS device. 145 145 00:04:42,050 --> 00:04:45,150 We're not comfortable just having all of our devices there, 146 146 00:04:45,150 --> 00:04:46,480 so we decided we weren't comfortable 147 147 00:04:46,480 --> 00:04:48,980 having all of our file storage on a single hard drive, 148 148 00:04:48,980 --> 00:04:51,070 and we built this NAS array instead, 149 149 00:04:51,070 --> 00:04:52,760 so if one of those drives goes out, 150 150 00:04:52,760 --> 00:04:55,220 we have three others that are carrying the load. 151 151 00:04:55,220 --> 00:04:56,590 This is the idea here. 152 152 00:04:56,590 --> 00:04:59,300 Now, eventually we decided we didn't need that NAS anymore, 153 153 00:04:59,300 --> 00:05:02,550 and so we replaced that NAS enclosure with a full RAID 5. 154 154 00:05:02,550 --> 00:05:04,320 Later on we took that full RAID 5 155 155 00:05:04,320 --> 00:05:06,180 and we switched it over to a cloud server 156 156 00:05:06,180 --> 00:05:07,310 that has redundant backups 157 157 00:05:07,310 --> 00:05:09,040 in two different cloud environments. 158 158 00:05:09,040 --> 00:05:10,740 And so all of these things work together 159 159 00:05:10,740 --> 00:05:12,100 based on our decisions, 160 160 00:05:12,100 --> 00:05:13,560 but as we moved up that scale 161 161 00:05:13,560 --> 00:05:15,180 and got more and more redundancy, 162 162 00:05:15,180 --> 00:05:17,150 we have more and more costs associated. 163 163 00:05:17,150 --> 00:05:19,440 It was a lot cheaper just to have an 8-terabyte hard drive 164 164 00:05:19,440 --> 00:05:20,780 with all of our files on it, 165 165 00:05:20,780 --> 00:05:21,960 then we went to a NAS array 166 166 00:05:21,960 --> 00:05:23,670 and that cost two or three times that money, 167 167 00:05:23,670 --> 00:05:24,586 then we went to a full RAID 5 168 168 00:05:24,586 --> 00:05:26,640 and that cost a couple more times that, 169 169 00:05:26,640 --> 00:05:29,010 then we went to the cloud and we have to pay more for that. 170 170 00:05:29,010 --> 00:05:30,740 Remember, all your decisions here 171 171 00:05:30,740 --> 00:05:32,350 are going to cost you more money, 172 172 00:05:32,350 --> 00:05:35,440 but if it's worth it to you, that would be important, right, 173 173 00:05:35,440 --> 00:05:37,071 and so these are the things you have to balance 174 174 00:05:37,071 --> 00:05:39,590 as you're designing these fully redundant networks, 175 175 00:05:39,590 --> 00:05:41,800 based on those technical goals. 176 176 00:05:41,800 --> 00:05:42,970 You also need to categorize 177 177 00:05:42,970 --> 00:05:45,630 all of your business applications into profiles, 178 178 00:05:45,630 --> 00:05:47,180 to help with this redundancy mission 179 179 00:05:47,180 --> 00:05:49,070 that you're trying to go and accomplish here. 180 180 00:05:49,070 --> 00:05:50,280 This will really help you as you start 181 181 00:05:50,280 --> 00:05:52,630 going into the quality of service as well. 182 182 00:05:52,630 --> 00:05:53,900 Now if I said, for instance, 183 183 00:05:53,900 --> 00:05:55,940 that web is considered category one 184 184 00:05:55,940 --> 00:05:57,550 and email is category two 185 185 00:05:57,550 --> 00:05:59,580 and streaming video's going to be category three, 186 186 00:05:59,580 --> 00:06:00,732 then we can apply profiles 187 187 00:06:00,732 --> 00:06:02,291 and give certain levels of service 188 188 00:06:02,291 --> 00:06:04,290 to each of those categories. 189 189 00:06:04,290 --> 00:06:06,450 Now we'll talk specifically of how that works 190 190 00:06:06,450 --> 00:06:09,200 when we talk about quality of service in a future lesson. 191 191 00:06:09,200 --> 00:06:10,240 Another thing we want to do 192 192 00:06:10,240 --> 00:06:11,820 is establish performance standards 193 193 00:06:11,820 --> 00:06:14,010 for our high-availability networks. 194 194 00:06:14,010 --> 00:06:16,370 What are the standards that we're going to have to have? 195 195 00:06:16,370 --> 00:06:17,670 These standards are going to drive 196 196 00:06:17,670 --> 00:06:19,550 how success is measured for us, 197 197 00:06:19,550 --> 00:06:21,960 and in the case of my file server, for instance, 198 198 00:06:21,960 --> 00:06:24,540 we measure success as it being up and available 199 199 00:06:24,540 --> 00:06:26,660 when my video editors need to access it, 200 200 00:06:26,660 --> 00:06:28,020 and that they don't lose data, 201 201 00:06:28,020 --> 00:06:29,520 because if we lost all of our files, 202 202 00:06:29,520 --> 00:06:31,110 that'd be bad for us, right? 203 203 00:06:31,110 --> 00:06:32,570 Those are two metrics that we have, 204 204 00:06:32,570 --> 00:06:35,035 and we have numbers associated with each of those things. 205 205 00:06:35,035 --> 00:06:38,020 In other organizations, we measure it based on the uptime 206 206 00:06:38,020 --> 00:06:40,030 of the entire end-to-end service, 207 207 00:06:40,030 --> 00:06:42,670 so if a client can't get out to the internet for an ISP, 208 208 00:06:42,670 --> 00:06:45,530 that would be a bad thing, that's one of their measurements. 209 209 00:06:45,530 --> 00:06:47,890 Now the other one might be, what is their uptime? 210 210 00:06:47,890 --> 00:06:49,420 All of these performance standards are developed 211 211 00:06:49,420 --> 00:06:51,760 through metrics and key performance indicators. 212 212 00:06:51,760 --> 00:06:52,950 If you're using something like ITIL 213 213 00:06:52,950 --> 00:06:54,850 as your IT service management standards, 214 214 00:06:54,850 --> 00:06:56,520 this is what you're going to be doing as you're trying 215 215 00:06:56,520 --> 00:06:59,310 to run those inside your organization as well. 216 216 00:06:59,310 --> 00:07:02,490 Finally, here we wanted to find how we manage and measure 217 217 00:07:02,490 --> 00:07:04,940 the high-availability solutions for ourselves. 218 218 00:07:04,940 --> 00:07:07,710 Metrics are going to be really useful to quantify success, 219 219 00:07:07,710 --> 00:07:09,860 if you develop those metrics correctly. 220 220 00:07:09,860 --> 00:07:12,920 Decision-makers and leaders love seeing metrics. 221 221 00:07:12,920 --> 00:07:15,180 They love seeing charts and seeing the performance, 222 222 00:07:15,180 --> 00:07:16,640 and how it's going up over time, 223 223 00:07:16,640 --> 00:07:18,200 and how our availability is going up, 224 224 00:07:18,200 --> 00:07:19,970 and how our costs are going down. 225 225 00:07:19,970 --> 00:07:21,330 Those are all good things, 226 226 00:07:21,330 --> 00:07:22,780 but if you don't know what you're measuring 227 227 00:07:22,780 --> 00:07:24,110 or why you're measuring it, 228 228 00:07:24,110 --> 00:07:26,190 it really goes back to your performance standards. 229 229 00:07:26,190 --> 00:07:27,470 Then, these are the kind of things 230 230 00:07:27,470 --> 00:07:29,220 that are wasting your time with metrics. 231 231 00:07:29,220 --> 00:07:31,130 A lot of people measure a lot of things, 232 232 00:07:31,130 --> 00:07:32,140 and they don't really tell you 233 233 00:07:32,140 --> 00:07:34,240 if you're getting the outcome you're wanting. 234 234 00:07:34,240 --> 00:07:35,960 I want to make sure that you think about 235 235 00:07:35,960 --> 00:07:38,570 how you decide on what metrics you're going to use. 236 236 00:07:38,570 --> 00:07:40,478 Now, we've covered a lot of different design criteria 237 237 00:07:40,478 --> 00:07:43,448 in this lesson, but the real big takeaway here 238 238 00:07:43,448 --> 00:07:45,530 that I want you to think about is this. 239 239 00:07:45,530 --> 00:07:47,230 If you have an existing network, 240 240 00:07:47,230 --> 00:07:48,820 you can add availability to it, 241 241 00:07:48,820 --> 00:07:50,600 and you can add redundancy to it. 242 242 00:07:50,600 --> 00:07:52,320 You can retrofit stuff in, 243 243 00:07:52,320 --> 00:07:53,834 but it's going to cost you a lot more time 244 244 00:07:53,834 --> 00:07:55,720 and a lot more money. 245 245 00:07:55,720 --> 00:07:57,600 It is much, much cheaper 246 246 00:07:57,600 --> 00:07:59,377 to design this stuff early in the process 247 247 00:07:59,377 --> 00:08:01,740 when you start building a network from scratch. 248 248 00:08:01,740 --> 00:08:04,490 So, if you're designing a network and you're asked early on 249 249 00:08:04,490 --> 00:08:05,970 what kind of things you need, 250 250 00:08:05,970 --> 00:08:08,480 I want you to think about all these things of redundancy 251 251 00:08:08,480 --> 00:08:10,020 in your initial design. 252 252 00:08:10,020 --> 00:08:13,210 Adding them in early is going to save you a lot of money. 253 253 00:08:13,210 --> 00:08:15,470 Every project has three main factors, 254 254 00:08:15,470 --> 00:08:17,680 time, cost, and quality, 255 255 00:08:17,680 --> 00:08:20,029 and usually, one of these things is going to suffer 256 256 00:08:20,029 --> 00:08:22,300 at the expense of the other two. 257 257 00:08:22,300 --> 00:08:24,570 For example, if I asked you to build me a network 258 258 00:08:24,570 --> 00:08:25,940 and I want it to be fully redundant 259 259 00:08:25,940 --> 00:08:28,600 and available by tomorrow, could you do it? 260 260 00:08:28,600 --> 00:08:32,000 Well, maybe, but it's probably going to cost me a lot of money, 261 261 00:08:32,000 --> 00:08:33,790 and because they give you very little time, 262 262 00:08:33,790 --> 00:08:35,340 it's going to cost me even more, 263 263 00:08:35,340 --> 00:08:37,420 or your quality is going to suffer. 264 264 00:08:37,420 --> 00:08:39,810 So, you could do it good, you could do it quick, 265 265 00:08:39,810 --> 00:08:42,610 or you could do it cheap, but you can't do all three. 266 266 00:08:42,610 --> 00:08:45,260 It's always going to be a trade-off between these three things, 267 267 00:08:45,260 --> 00:08:46,270 and I want you to remember 268 268 00:08:46,270 --> 00:08:48,610 as you're out there and you're designing networks, 269 269 00:08:48,610 --> 00:08:50,850 you need to make sure you're thinking about your redundancy 270 270 00:08:50,850 --> 00:08:53,190 and your availability and your reliability, 271 271 00:08:53,190 --> 00:08:55,393 because often that quality is going to suffer 272 272 00:08:55,393 --> 00:08:57,630 in favor of getting things out quicker 273 273 00:08:57,630 --> 00:08:59,130 or getting things out cheaper.