1 1 00:00:00,480 --> 00:00:02,270 Facilities support. 2 2 00:00:02,270 --> 00:00:04,530 In this lesson, we're going to discuss the concept 3 3 00:00:04,530 --> 00:00:06,830 of facilities and infrastructure support 4 4 00:00:06,830 --> 00:00:09,410 for our data centers and our recovery sites. 5 5 00:00:09,410 --> 00:00:11,310 To provide proper facility support, 6 6 00:00:11,310 --> 00:00:13,730 it's important to consider power, cooling, 7 7 00:00:13,730 --> 00:00:15,070 and fire suppression. 8 8 00:00:15,070 --> 00:00:17,820 So we're going to cover uninterrupted power supplies, 9 9 00:00:17,820 --> 00:00:20,810 power distribution units, generators, HVAC, 10 10 00:00:20,810 --> 00:00:23,190 and fire suppression systems in this lesson. 11 11 00:00:23,190 --> 00:00:27,140 First, we have a UPS, or uninterruptible power supply. 12 12 00:00:27,140 --> 00:00:28,850 Now an uninterruptible power supply, 13 13 00:00:28,850 --> 00:00:30,680 or uninterruptible power source, 14 14 00:00:30,680 --> 00:00:31,950 is an electrical apparatus 15 15 00:00:31,950 --> 00:00:34,350 that provides emergency power to a load 16 16 00:00:34,350 --> 00:00:36,590 whenever the input power source or main power 17 17 00:00:36,590 --> 00:00:37,940 is going to fail. 18 18 00:00:37,940 --> 00:00:40,390 Most people think of these as battery backups, 19 19 00:00:40,390 --> 00:00:42,940 but in our data centers and telecommunication closets, 20 20 00:00:42,940 --> 00:00:44,170 we usually see devices 21 21 00:00:44,170 --> 00:00:46,490 that contain more than just a battery backup. 22 22 00:00:46,490 --> 00:00:48,680 For our purposes, we're going to use an UPS 23 23 00:00:48,680 --> 00:00:50,220 that is going to provide line conditioning 24 24 00:00:50,220 --> 00:00:52,840 and protect us from surges and spikes in power. 25 25 00:00:52,840 --> 00:00:54,300 Our goal in using an UPS 26 26 00:00:54,300 --> 00:00:57,210 is to make sure that we have clean reliable power. 27 27 00:00:57,210 --> 00:00:59,950 Now an UPS is great for short duration power outages, 28 28 00:00:59,950 --> 00:01:03,450 but they usually don't last more than about 15 to 30 minutes 29 29 00:01:03,450 --> 00:01:05,840 because they have a relatively short battery life. 30 30 00:01:05,840 --> 00:01:07,250 The good news is the batteries 31 31 00:01:07,250 --> 00:01:08,840 are getting better and better every day 32 32 00:01:08,840 --> 00:01:10,880 and their lives are getting longer and longer 33 33 00:01:10,880 --> 00:01:12,160 in newer units. 34 34 00:01:12,160 --> 00:01:16,050 Second, we have power distribution units or PDUs. 35 35 00:01:16,050 --> 00:01:17,560 Now a power distribution unit 36 36 00:01:17,560 --> 00:01:19,600 is a device fitted with multiple outputs 37 37 00:01:19,600 --> 00:01:21,630 designed to distribute electrical power, 38 38 00:01:21,630 --> 00:01:23,440 especially to racks of computers 39 39 00:01:23,440 --> 00:01:26,550 and networking equipment located within our data centers. 40 40 00:01:26,550 --> 00:01:28,130 PDUs can be rack-mounted 41 41 00:01:28,130 --> 00:01:31,120 or they can take the form of a large cabinet. 42 42 00:01:31,120 --> 00:01:32,470 In large data center, 43 43 00:01:32,470 --> 00:01:34,510 you're usually going to see these large cabinets, 44 44 00:01:34,510 --> 00:01:35,343 and in general, 45 45 00:01:35,343 --> 00:01:38,080 there's going to be one PDU for each row of servers 46 46 00:01:38,080 --> 00:01:40,230 and it maintains the high current circuits, 47 47 00:01:40,230 --> 00:01:41,140 circuit breakers, 48 48 00:01:41,140 --> 00:01:43,670 and power monitoring panels inside of them. 49 49 00:01:43,670 --> 00:01:46,440 These PDUs can provide power protection from surges, 50 50 00:01:46,440 --> 00:01:47,740 spikes, and brownouts, 51 51 00:01:47,740 --> 00:01:48,940 but they are not designed 52 52 00:01:48,940 --> 00:01:51,670 to provide full blackout protection like an UPS would 53 53 00:01:51,670 --> 00:01:53,760 because they don't have battery backups. 54 54 00:01:53,760 --> 00:01:57,250 Generally, a PDU be combined with an UPS or a generator 55 55 00:01:57,250 --> 00:02:00,600 to provide that power that is needed during a blackout. 56 56 00:02:00,600 --> 00:02:02,300 Third, we have generators. 57 57 00:02:02,300 --> 00:02:04,330 Now large generators are usually going to be installed 58 58 00:02:04,330 --> 00:02:05,740 outside of a data center 59 59 00:02:05,740 --> 00:02:07,660 in order to provide us with longterm power 60 60 00:02:07,660 --> 00:02:10,070 during a power outage inside your region. 61 61 00:02:10,070 --> 00:02:12,070 These generators can be powered by diesel, 62 62 00:02:12,070 --> 00:02:13,740 gasoline, or propane. 63 63 00:02:13,740 --> 00:02:15,330 For example, at my office, 64 64 00:02:15,330 --> 00:02:17,760 I have a 20,000 kilowatt diesel generator 65 65 00:02:17,760 --> 00:02:20,750 that's used to provide power in case we have a power outage. 66 66 00:02:20,750 --> 00:02:22,640 Now the big challenge with a generator though, 67 67 00:02:22,640 --> 00:02:24,750 is that they take time to get up to speed 68 68 00:02:24,750 --> 00:02:26,590 until they're ready to start providing power 69 69 00:02:26,590 --> 00:02:27,630 to your devices. 70 70 00:02:27,630 --> 00:02:30,750 They can take usually between 45 to 90 seconds. 71 71 00:02:30,750 --> 00:02:32,470 So you usually need to pair them up 72 72 00:02:32,470 --> 00:02:34,280 with a battery backup or UPS 73 73 00:02:34,280 --> 00:02:37,020 as you're designing your power redundancy solution. 74 74 00:02:37,020 --> 00:02:39,910 For example, at my office, if the power goes out, 75 75 00:02:39,910 --> 00:02:42,480 the UPS will carry the load for up to 15 minutes. 76 76 00:02:42,480 --> 00:02:43,570 During that time, 77 77 00:02:43,570 --> 00:02:45,780 the generator will automatically be brought online, 78 78 00:02:45,780 --> 00:02:48,420 usually taking 45 to 90 seconds. 79 79 00:02:48,420 --> 00:02:50,100 Once that generator is fully online, 80 80 00:02:50,100 --> 00:02:52,030 and providing the right stable power, 81 81 00:02:52,030 --> 00:02:53,350 and it's ready to take the load, 82 82 00:02:53,350 --> 00:02:54,250 the power gets shifted 83 83 00:02:54,250 --> 00:02:56,380 from the UPS batteries to the generator, 84 84 00:02:56,380 --> 00:02:59,770 using an automatic transfer switch or ATS. 85 85 00:02:59,770 --> 00:03:01,550 Now once the power is restored in our area 86 86 00:03:01,550 --> 00:03:03,830 for at least five minutes being steady, 87 87 00:03:03,830 --> 00:03:06,910 then our ATS will actually shift power back to the grid 88 88 00:03:06,910 --> 00:03:08,970 through our UPS unit, that battery backup, 89 89 00:03:08,970 --> 00:03:10,660 and then shut down our generator. 90 90 00:03:10,660 --> 00:03:12,890 Fourth, we have HVAC units. 91 91 00:03:12,890 --> 00:03:15,980 HVAC stands for heating, ventilation, and air conditioning. 92 92 00:03:15,980 --> 00:03:16,813 Our data centers 93 93 00:03:16,813 --> 00:03:19,160 are going to generate a ton of heat inside of them 94 94 00:03:19,160 --> 00:03:20,830 because of all these servers, and switches, 95 95 00:03:20,830 --> 00:03:22,030 and routers, and firewalls, 96 96 00:03:22,030 --> 00:03:24,090 that are doing processing inside of them. 97 97 00:03:24,090 --> 00:03:25,580 To cool down these devices, 98 98 00:03:25,580 --> 00:03:27,780 we need to have a good HVAC system. 99 99 00:03:27,780 --> 00:03:29,070 Now to help with this cooling, 100 100 00:03:29,070 --> 00:03:30,850 most data centers are going to utilize 101 101 00:03:30,850 --> 00:03:32,890 a hot and cold aisle concept. 102 102 00:03:32,890 --> 00:03:34,200 Now in the simplest form, 103 103 00:03:34,200 --> 00:03:37,270 each row of servers is going to face another row of servers. 104 104 00:03:37,270 --> 00:03:38,380 These two rows 105 105 00:03:38,380 --> 00:03:40,350 will have the front of the servers facing each other 106 106 00:03:40,350 --> 00:03:43,320 and the rear of the servers facing away from the aisle. 107 107 00:03:43,320 --> 00:03:44,960 This is because the servers are designed 108 108 00:03:44,960 --> 00:03:47,130 to push air out the rear of the device. 109 109 00:03:47,130 --> 00:03:49,540 So the front of the servers is in the cold aisle 110 110 00:03:49,540 --> 00:03:52,330 and the rear of the servers is in the hot aisle. 111 111 00:03:52,330 --> 00:03:55,560 This lets us focus our HVAC systems into the hot aisles 112 112 00:03:55,560 --> 00:03:57,170 to suck that hot air out, 113 113 00:03:57,170 --> 00:03:59,720 cool it down, and return it back to the cold aisle, 114 114 00:03:59,720 --> 00:04:02,470 where it can then be circulated over the servers once again. 115 115 00:04:02,470 --> 00:04:04,910 Remember, proper cooling is important to the health 116 116 00:04:04,910 --> 00:04:07,400 and security of our networks and our devices. 117 117 00:04:07,400 --> 00:04:09,370 If the network devices start to overheat, 118 118 00:04:09,370 --> 00:04:10,630 they will shut themselves down 119 119 00:04:10,630 --> 00:04:12,140 to protect their critical components, 120 120 00:04:12,140 --> 00:04:14,850 and if those components get overheated for too long, 121 121 00:04:14,850 --> 00:04:16,500 permanent damage can occur 122 122 00:04:16,500 --> 00:04:19,750 or it can decrease the life expectancy of those devices. 123 123 00:04:19,750 --> 00:04:21,870 Now our fifth and final thing we need to discuss 124 124 00:04:21,870 --> 00:04:23,350 is fire suppression. 125 125 00:04:23,350 --> 00:04:24,400 In a data center, 126 126 00:04:24,400 --> 00:04:26,900 we usually have built-in fire suppression systems. 127 127 00:04:26,900 --> 00:04:28,670 These can include wet pipe sprinklers, 128 128 00:04:28,670 --> 00:04:31,630 pre-action sprinklers, and special suppression systems. 129 129 00:04:31,630 --> 00:04:33,980 Now a wet pipe system is the most basic type 130 130 00:04:33,980 --> 00:04:35,350 of fire suppression system, 131 131 00:04:35,350 --> 00:04:37,650 and it involves a sprinkler system and pipes 132 132 00:04:37,650 --> 00:04:40,080 that always contain water in those pipes. 133 133 00:04:40,080 --> 00:04:42,360 Now in a server room or data center environment, 134 134 00:04:42,360 --> 00:04:43,640 this is kind of dangerous 135 135 00:04:43,640 --> 00:04:46,090 because a leak in that pipe could damage your servers 136 136 00:04:46,090 --> 00:04:47,690 that are sitting underneath them. 137 137 00:04:47,690 --> 00:04:50,380 In general, you should avoid using a wet pipe system 138 138 00:04:50,380 --> 00:04:52,320 in and around your data centers. 139 139 00:04:52,320 --> 00:04:54,840 Instead, you should use a pre-action system 140 140 00:04:54,840 --> 00:04:57,060 to minimize the risk of accidental release 141 141 00:04:57,060 --> 00:04:59,430 if you're going to be using a wet pipe system. 142 142 00:04:59,430 --> 00:05:00,840 With a pre-action system, 143 143 00:05:00,840 --> 00:05:02,490 both the detector actuator 144 144 00:05:02,490 --> 00:05:04,270 is going to work like a smoke detector, 145 145 00:05:04,270 --> 00:05:05,660 and then there's going to be a sprinkler 146 146 00:05:05,660 --> 00:05:07,070 that has to be tripped also 147 147 00:05:07,070 --> 00:05:09,010 before the water is going to be released. 148 148 00:05:09,010 --> 00:05:11,310 Again, using water in a data center, 149 149 00:05:11,310 --> 00:05:12,810 even in a pre-action system, 150 150 00:05:12,810 --> 00:05:16,190 is not really a good idea though, so I try to avoid it. 151 151 00:05:16,190 --> 00:05:19,260 Instead, I like to rely on special suppression systems 152 152 00:05:19,260 --> 00:05:20,840 for most of my data centers. 153 153 00:05:20,840 --> 00:05:23,380 This will use something like a clean agent system. 154 154 00:05:23,380 --> 00:05:26,020 Now a clean agent is something like halocarbon agents 155 155 00:05:26,020 --> 00:05:27,340 or inert gases, 156 156 00:05:27,340 --> 00:05:29,830 which released, the agents will displace the oxygen 157 157 00:05:29,830 --> 00:05:31,560 in the room with that inert gas 158 158 00:05:31,560 --> 00:05:33,420 and essentially suffocate the fire. 159 159 00:05:33,420 --> 00:05:34,500 Now, the danger with using 160 160 00:05:34,500 --> 00:05:36,370 a special suppressant system like this 161 161 00:05:36,370 --> 00:05:38,790 is that if there's people working in your data center, 162 162 00:05:38,790 --> 00:05:40,480 those people can suffocate 163 163 00:05:40,480 --> 00:05:42,430 if the clean agent is being released. 164 164 00:05:42,430 --> 00:05:45,550 So your data center needs to be equipped with an alarm 165 165 00:05:45,550 --> 00:05:47,910 that announces when the clean agent is being released, 166 166 00:05:47,910 --> 00:05:49,150 and you also need to make sure 167 167 00:05:49,150 --> 00:05:51,150 there's supplemental oxygen masks available 168 168 00:05:51,150 --> 00:05:52,540 and easily accessible 169 169 00:05:52,540 --> 00:05:54,840 by any person who's working in that data center 170 170 00:05:54,840 --> 00:05:56,490 whenever they hear the alarm go off 171 171 00:05:56,490 --> 00:05:58,390 for that clean agent release. 172 172 00:05:58,390 --> 00:06:01,110 So remember, when you're designing your data centers 173 173 00:06:01,110 --> 00:06:03,930 and your primary work environment or your recovery sites, 174 174 00:06:03,930 --> 00:06:05,460 you need to consider your power, 175 175 00:06:05,460 --> 00:06:07,760 your cooling, and your fire suppression needs.