1 1 00:00:00,140 --> 00:00:02,150 High availability. 2 2 00:00:02,150 --> 00:00:02,983 In this lesson, 3 3 00:00:02,983 --> 00:00:05,510 we're going to talk all about high availability. 4 4 00:00:05,510 --> 00:00:07,490 Now, when we're talking about high availability, 5 5 00:00:07,490 --> 00:00:09,860 we're really talking about making sure our systems are up 6 6 00:00:09,860 --> 00:00:11,020 and available. 7 7 00:00:11,020 --> 00:00:13,720 Availability is going to be measured in what we call uptime 8 8 00:00:13,720 --> 00:00:16,640 or how many minutes or hours you're up and available 9 9 00:00:16,640 --> 00:00:18,330 as shown as a percentage. 10 10 00:00:18,330 --> 00:00:19,840 Usually, you're going to take the amount of minutes 11 11 00:00:19,840 --> 00:00:20,673 you were up, 12 12 00:00:20,673 --> 00:00:22,770 divided by the total amount of minutes in a period, 13 13 00:00:22,770 --> 00:00:25,470 and that gives you a percentage known as uptime. 14 14 00:00:25,470 --> 00:00:28,270 Now, we try to maintain what is known as the five nines 15 15 00:00:28,270 --> 00:00:30,530 of availability in most commercial networks. 16 16 00:00:30,530 --> 00:00:35,010 This is actually really hard because that's 99.999%. 17 17 00:00:35,010 --> 00:00:37,560 That means I get a maximum of about five minutes 18 18 00:00:37,560 --> 00:00:38,920 of downtime per year, 19 19 00:00:38,920 --> 00:00:40,860 which is not a whole lot of downtime. 20 20 00:00:40,860 --> 00:00:42,270 In some cloud based networks, 21 21 00:00:42,270 --> 00:00:47,270 they aim for six nines of availability or 99.99999%. 22 22 00:00:47,400 --> 00:00:50,270 This equates to just 31 seconds of downtime 23 23 00:00:50,270 --> 00:00:51,463 each and every year. 24 24 00:00:51,463 --> 00:00:52,930 Now, as you imagine, 25 25 00:00:52,930 --> 00:00:55,220 I need more than 31 seconds of downtime 26 26 00:00:55,220 --> 00:00:57,030 or even five minutes of downtime 27 27 00:00:57,030 --> 00:00:59,790 fueled up patch my servers and install a new hard drive 28 28 00:00:59,790 --> 00:01:02,180 or put in a new router or switch when one fails. 29 29 00:01:02,180 --> 00:01:05,570 So, how do I maintain that high level of availability? 30 30 00:01:05,570 --> 00:01:06,910 Well, I'm going to do that, 31 31 00:01:06,910 --> 00:01:09,760 by designing my networks to be highly available. 32 32 00:01:09,760 --> 00:01:11,670 Now, there are two terms you need to understand 33 33 00:01:11,670 --> 00:01:12,503 and be familiar with, 34 34 00:01:12,503 --> 00:01:14,120 when we talk about high availability. 35 35 00:01:14,120 --> 00:01:16,470 There is availability and reliability, 36 36 00:01:16,470 --> 00:01:18,030 and these are different things. 37 37 00:01:18,030 --> 00:01:19,760 When I'm talking about availability, 38 38 00:01:19,760 --> 00:01:22,270 this is concerned with being up and operational. 39 39 00:01:22,270 --> 00:01:23,890 When I talk about reliability, 40 40 00:01:23,890 --> 00:01:25,750 I'm concerned with not dropping packets 41 41 00:01:25,750 --> 00:01:27,300 inside of my network. 42 42 00:01:27,300 --> 00:01:29,040 If your network is highly available, 43 43 00:01:29,040 --> 00:01:30,200 but it's not reliable, 44 44 00:01:30,200 --> 00:01:31,440 it's not a very good network 45 45 00:01:31,440 --> 00:01:32,920 because it's dropping things all the time 46 46 00:01:32,920 --> 00:01:34,490 and isn't doing what it's supposed to. 47 47 00:01:34,490 --> 00:01:35,610 But conversely, 48 48 00:01:35,610 --> 00:01:37,860 you can have a really highly reliable network, 49 49 00:01:37,860 --> 00:01:39,570 but if it's not a highly available one, 50 50 00:01:39,570 --> 00:01:41,870 nobody can use it either because it's down all the time. 51 51 00:01:41,870 --> 00:01:43,400 So that wouldn't be good either. 52 52 00:01:43,400 --> 00:01:45,670 So, let's say I had the most highly reliable network 53 53 00:01:45,670 --> 00:01:46,840 in the entire world, 54 54 00:01:46,840 --> 00:01:48,920 but it's only up 20 minutes a year. 55 55 00:01:48,920 --> 00:01:50,380 That's not going to be any good, right? 56 56 00:01:50,380 --> 00:01:52,450 So, we want to make sure we balance these two things. 57 57 00:01:52,450 --> 00:01:55,100 We have to aim for good enough in both areas 58 58 00:01:55,100 --> 00:01:57,602 to meet our business needs based on the available resources 59 59 00:01:57,602 --> 00:02:00,470 and the amount of money we have to build our networks. 60 60 00:02:00,470 --> 00:02:02,980 So, when we measure our different network components, 61 61 00:02:02,980 --> 00:02:05,320 we have to determine how highly available they are. 62 62 00:02:05,320 --> 00:02:09,490 And we do that through measurement of MTTR and MTBF. 63 63 00:02:09,490 --> 00:02:12,440 Now, MTTR is the mean time to repair. 64 64 00:02:12,440 --> 00:02:14,770 This measure is the average time it takes to repair 65 65 00:02:14,770 --> 00:02:16,550 a network device when it breaks. 66 66 00:02:16,550 --> 00:02:17,383 After all, 67 67 00:02:17,383 --> 00:02:19,560 everything is going to break eventually. 68 68 00:02:19,560 --> 00:02:21,220 So, when a device breaks, 69 69 00:02:21,220 --> 00:02:22,920 how long does it take for you to fix it? 70 70 00:02:22,920 --> 00:02:25,460 And how much downtime are you going to experience? 71 71 00:02:25,460 --> 00:02:26,780 That is what we're trying to measure 72 72 00:02:26,780 --> 00:02:29,020 when we deal with the mean time to repair. 73 73 00:02:29,020 --> 00:02:32,100 Now, the mean time between failures or MTBF, 74 74 00:02:32,100 --> 00:02:33,500 is going to measure the average time 75 75 00:02:33,500 --> 00:02:35,438 between when a failure happens on a device 76 76 00:02:35,438 --> 00:02:37,820 and the next failure happens. 77 77 00:02:37,820 --> 00:02:39,910 Now, these two terms can often be confusing. 78 78 00:02:39,910 --> 00:02:42,780 So, let me display it on a timeline and explain a little bit 79 79 00:02:42,780 --> 00:02:44,860 about what they look like in the real world. 80 80 00:02:44,860 --> 00:02:46,470 Now, let's say I had a system failure 81 81 00:02:46,470 --> 00:02:49,050 at this first stop sign here on the left side. 82 82 00:02:49,050 --> 00:02:52,150 Then we resume normal operations because we fix things. 83 83 00:02:52,150 --> 00:02:54,850 That amount of time, was the time to repair. 84 84 00:02:54,850 --> 00:02:56,750 Now, if I average all the times to repair 85 85 00:02:56,750 --> 00:02:59,390 over the entire year for that type of device, 86 86 00:02:59,390 --> 00:03:01,060 that's going to give me my MTTR, 87 87 00:03:01,060 --> 00:03:04,380 my mean time to repair, the average time to repair. 88 88 00:03:04,380 --> 00:03:06,040 Now, on the failure side of things, 89 89 00:03:06,040 --> 00:03:08,442 we want to measure the failure time from one failure 90 90 00:03:08,442 --> 00:03:10,270 thus using it, fixing it, 91 91 00:03:10,270 --> 00:03:12,160 and then another failure happens. 92 92 00:03:12,160 --> 00:03:14,670 This becomes the time between the failures. 93 93 00:03:14,670 --> 00:03:16,230 If I average all those together, 94 94 00:03:16,230 --> 00:03:18,070 I get the average time between failures 95 95 00:03:18,070 --> 00:03:21,620 or the mean time between failures, MTBF. 96 96 00:03:21,620 --> 00:03:23,600 Hopefully, you can see the difference here. 97 97 00:03:23,600 --> 00:03:26,320 Remember, when we're dealing with mean time to repair, 98 98 00:03:26,320 --> 00:03:28,440 we want this to be a very small number. 99 99 00:03:28,440 --> 00:03:30,410 When we deal with the mean time between failures, 100 100 00:03:30,410 --> 00:03:32,620 we want this to be a very large number. 101 101 00:03:32,620 --> 00:03:33,453 This means, 102 102 00:03:33,453 --> 00:03:35,490 and for the very small number for mean time to repair, 103 103 00:03:35,490 --> 00:03:37,230 we can fix things really quickly 104 104 00:03:37,230 --> 00:03:38,780 and get ourselves back online. 105 105 00:03:38,780 --> 00:03:40,674 So, the lower the mean time to repair is, 106 106 00:03:40,674 --> 00:03:43,090 the better the network availability. 107 107 00:03:43,090 --> 00:03:44,090 Now, on the other hand, 108 108 00:03:44,090 --> 00:03:46,240 we start talking about mean time between failures, 109 109 00:03:46,240 --> 00:03:48,200 we want a really long time 110 110 00:03:48,200 --> 00:03:49,930 because this means that the device has stayed up 111 111 00:03:49,930 --> 00:03:53,060 and operational for a very long time before they fail. 112 112 00:03:53,060 --> 00:03:55,370 This is going to give us better network availability, 113 113 00:03:55,370 --> 00:03:58,500 and overall, it should give us better reliability too. 114 114 00:03:58,500 --> 00:04:00,090 Now, we don't want a lot of failures here. 115 115 00:04:00,090 --> 00:04:02,330 And so the more time in between failures, 116 116 00:04:02,330 --> 00:04:04,360 the better that is for our network. 117 117 00:04:04,360 --> 00:04:06,000 So, how do we design these networks 118 118 00:04:06,000 --> 00:04:08,560 to be highly reliable and highly available? 119 119 00:04:08,560 --> 00:04:10,830 Well, we're going to add redundancy to our networks 120 120 00:04:10,830 --> 00:04:12,230 and their devices. 121 121 00:04:12,230 --> 00:04:14,560 Now, redundancy can be achieved through a single device 122 122 00:04:14,560 --> 00:04:16,740 or by using multiple devices. 123 123 00:04:16,740 --> 00:04:18,390 If you're using a single device, 124 124 00:04:18,390 --> 00:04:20,110 you're still going to have single points of failure 125 125 00:04:20,110 --> 00:04:21,070 in your network, 126 126 00:04:21,070 --> 00:04:24,230 but it is cheaper than being fully hardware redundant. 127 127 00:04:24,230 --> 00:04:26,400 Let's take a look at this concept for a moment. 128 128 00:04:26,400 --> 00:04:29,060 Here you could see a single point of failure in my network. 129 129 00:04:29,060 --> 00:04:31,300 Even though I have two switches and multiple connections 130 130 00:04:31,300 --> 00:04:32,390 between those switches, 131 131 00:04:32,390 --> 00:04:34,430 which gives me additional redundancy, 132 132 00:04:34,430 --> 00:04:37,120 that router is not giving me additional redundancy. 133 133 00:04:37,120 --> 00:04:38,760 It's a single point of failure 134 134 00:04:38,760 --> 00:04:40,710 because it's the only router I have. 135 135 00:04:40,710 --> 00:04:43,390 So, even if the router has internal hardware redundancy, 136 136 00:04:43,390 --> 00:04:45,750 like two power supplies and two network cards, 137 137 00:04:45,750 --> 00:04:49,010 I still only have one router chassis and one circuit board 138 138 00:04:49,010 --> 00:04:50,300 running in that router. 139 139 00:04:50,300 --> 00:04:51,800 So, if that router goes down, 140 140 00:04:51,800 --> 00:04:53,920 this entire network is going to stop. 141 141 00:04:53,920 --> 00:04:56,820 Therefore, this is considered a single point of failure. 142 142 00:04:56,820 --> 00:04:57,653 Now instead, 143 143 00:04:57,653 --> 00:04:58,700 I could redesign the network 144 144 00:04:58,700 --> 00:05:01,400 and I can increase its redundancy by doing this. 145 145 00:05:01,400 --> 00:05:04,300 Notice, I now have two PCs that want to talk to each other. 146 146 00:05:04,300 --> 00:05:06,790 And each of them has dual network interface cards 147 147 00:05:06,790 --> 00:05:08,370 talking to two different switches. 148 148 00:05:08,370 --> 00:05:11,070 And each of those switches talks to two different routers. 149 149 00:05:11,070 --> 00:05:12,610 Everything is connected to everything else 150 150 00:05:12,610 --> 00:05:14,950 in a mesh topology for these network devices. 151 151 00:05:14,950 --> 00:05:17,330 This gives me multiple connections between each device 152 152 00:05:17,330 --> 00:05:20,110 and provides me with link redundancy, component redundancy, 153 153 00:05:20,110 --> 00:05:21,870 and even inside those devices, 154 154 00:05:21,870 --> 00:05:24,340 I may have two network cards, two power supplies, 155 155 00:05:24,340 --> 00:05:27,240 and to every other internal network component there is, 156 156 00:05:27,240 --> 00:05:28,750 so that I have a very redundant 157 157 00:05:28,750 --> 00:05:30,480 and highly available network. 158 158 00:05:30,480 --> 00:05:32,570 Now, if one of those routers needs to be upgraded, 159 159 00:05:32,570 --> 00:05:34,910 I can take it offline and update its firmware, 160 160 00:05:34,910 --> 00:05:37,270 and then the entire time that second router 161 161 00:05:37,270 --> 00:05:38,670 is still on the network, 162 162 00:05:38,670 --> 00:05:41,840 maintaining the load and providing service to all the users. 163 163 00:05:41,840 --> 00:05:44,480 Then I can put the first router back on the network, 164 164 00:05:44,480 --> 00:05:47,150 take off the second router and then do its upgrades. 165 165 00:05:47,150 --> 00:05:49,170 By doing this and taking turns, 166 166 00:05:49,170 --> 00:05:51,730 I still am able to have network functions run, 167 167 00:05:51,730 --> 00:05:54,360 and I have no downtime associated with this network. 168 168 00:05:54,360 --> 00:05:57,390 This is how you keep a network highly available. 169 169 00:05:57,390 --> 00:06:00,610 Now, let's talk a little bit more about hardware redundancy. 170 170 00:06:00,610 --> 00:06:02,990 Inside these routers and other network devices, 171 171 00:06:02,990 --> 00:06:06,075 we can have hardware redundancy or the devices themselves 172 172 00:06:06,075 --> 00:06:08,270 could be hardware redundant. 173 173 00:06:08,270 --> 00:06:09,960 Now, if I have two routers and they're both 174 174 00:06:09,960 --> 00:06:11,470 serving the same function, 175 175 00:06:11,470 --> 00:06:13,690 this is considered hardware redundancy. 176 176 00:06:13,690 --> 00:06:16,220 I could also have hardware redundancy in the components 177 177 00:06:16,220 --> 00:06:18,800 by having two network cards or two hard drives 178 178 00:06:18,800 --> 00:06:21,850 or two internal power supplies on a single device. 179 179 00:06:21,850 --> 00:06:25,610 That way, if one of them fails, the second one takes over. 180 180 00:06:25,610 --> 00:06:27,000 Usually, you're going to find this 181 181 00:06:27,000 --> 00:06:28,630 in strategic network devices, 182 182 00:06:28,630 --> 00:06:31,210 things like your switches, your routers, your firewalls, 183 183 00:06:31,210 --> 00:06:32,080 and your servers, 184 184 00:06:32,080 --> 00:06:33,300 because you can't afford a failure 185 185 00:06:33,300 --> 00:06:34,760 in any one of those devices, 186 186 00:06:34,760 --> 00:06:36,330 because they would take down large portions 187 187 00:06:36,330 --> 00:06:38,470 of your network or its services. 188 188 00:06:38,470 --> 00:06:40,740 On the other hand, if I'm considering my laptop, 189 189 00:06:40,740 --> 00:06:42,520 I only have one hard drive in it. 190 190 00:06:42,520 --> 00:06:44,900 If that laptop fails or that hard drive fails, 191 191 00:06:44,900 --> 00:06:46,520 I would just deal with that downtime. 192 192 00:06:46,520 --> 00:06:48,540 I might buy a new laptop or a new hard drive 193 193 00:06:48,540 --> 00:06:50,680 and then restore from an old backup. 194 194 00:06:50,680 --> 00:06:52,450 That would get me back up and running. 195 195 00:06:52,450 --> 00:06:54,480 Now, when we're working with end-user devices 196 196 00:06:54,480 --> 00:06:56,020 like workstations and clients, 197 197 00:06:56,020 --> 00:06:58,180 we often don't deal with redundancy. 198 198 00:06:58,180 --> 00:07:00,620 But when you start getting to servers and routers 199 199 00:07:00,620 --> 00:07:02,200 and switches and firewalls, 200 200 00:07:02,200 --> 00:07:03,750 you need to start having hardware 201 201 00:07:03,750 --> 00:07:05,340 and component level redundancy 202 202 00:07:05,340 --> 00:07:07,950 because these serve lots of end-users. 203 203 00:07:07,950 --> 00:07:09,420 We deal with this redundancy, 204 204 00:07:09,420 --> 00:07:12,020 we can then cluster our devices and have them work either 205 205 00:07:12,020 --> 00:07:13,140 an active-active, 206 206 00:07:13,140 --> 00:07:15,220 or active-passive configuration. 207 207 00:07:15,220 --> 00:07:16,053 All right. 208 208 00:07:16,053 --> 00:07:17,100 Let's assume I have this one computer 209 209 00:07:17,100 --> 00:07:18,700 and it has two network interface cards 210 210 00:07:18,700 --> 00:07:20,530 that are connected to the network. 211 211 00:07:20,530 --> 00:07:23,150 Do I want to talk to both routers at the same time? 212 212 00:07:23,150 --> 00:07:24,930 Well, if I'm active-active, 213 213 00:07:24,930 --> 00:07:26,630 then both of those network interface cards 214 214 00:07:26,630 --> 00:07:28,230 are going to be active at the same time, 215 215 00:07:28,230 --> 00:07:30,170 and they each are going to have their own Mac address, 216 216 00:07:30,170 --> 00:07:31,970 and they're going to be talking at the same time 217 217 00:07:31,970 --> 00:07:33,740 to either of these two routers. 218 218 00:07:33,740 --> 00:07:36,040 This can then be done to increase the bandwidth 219 219 00:07:36,040 --> 00:07:37,640 of this computer and load balance 220 220 00:07:37,640 --> 00:07:39,930 across both network interface cards. 221 221 00:07:39,930 --> 00:07:42,340 This is known as Network Interface Card teaming, 222 222 00:07:42,340 --> 00:07:43,460 or NIC teaming, 223 223 00:07:43,460 --> 00:07:45,330 where a group of network interface cards, 224 224 00:07:45,330 --> 00:07:48,360 is used for load balancing and failover for a server 225 225 00:07:48,360 --> 00:07:50,090 or another device like that. 226 226 00:07:50,090 --> 00:07:50,950 Now, on the other hand, 227 227 00:07:50,950 --> 00:07:52,650 we can use active-passive, 228 228 00:07:52,650 --> 00:07:53,830 and this is going to have a primary 229 229 00:07:53,830 --> 00:07:55,870 and a backup network interface card. 230 230 00:07:55,870 --> 00:07:57,690 Now, one of these cards is going to be active 231 231 00:07:57,690 --> 00:07:59,400 and being used at all times. 232 232 00:07:59,400 --> 00:08:00,400 And when it fails, 233 233 00:08:00,400 --> 00:08:03,590 the other card is going to go from standby and take over. 234 234 00:08:03,590 --> 00:08:04,423 In this case, 235 235 00:08:04,423 --> 00:08:06,800 there is no performance increased by having two cards, 236 236 00:08:06,800 --> 00:08:10,140 but I have true redundancy and failover capabilities. 237 237 00:08:10,140 --> 00:08:12,110 In an active-passive configuration, 238 238 00:08:12,110 --> 00:08:14,200 both NICs are going to be working together 239 239 00:08:14,200 --> 00:08:15,800 and they're going to have a single Mac address 240 240 00:08:15,800 --> 00:08:17,300 that they're going to display to the network, 241 241 00:08:17,300 --> 00:08:19,320 so they look like they're a single device. 242 242 00:08:19,320 --> 00:08:21,490 Now, when you start looking at redundancy at layer three, 243 243 00:08:21,490 --> 00:08:23,260 we're going to start talking about our routers. 244 244 00:08:23,260 --> 00:08:24,093 Now here, 245 245 00:08:24,093 --> 00:08:26,270 our clients are getting configured with a default gateway, 246 246 00:08:26,270 --> 00:08:27,950 which is our router by default. 247 247 00:08:27,950 --> 00:08:29,880 But, if the default gateway went down, 248 248 00:08:29,880 --> 00:08:31,500 we wouldn't be able to leave the sub-net, 249 249 00:08:31,500 --> 00:08:33,490 and so it'd be stuck on the internal network. 250 250 00:08:33,490 --> 00:08:34,610 Now, we don't want that. 251 251 00:08:34,610 --> 00:08:35,443 So instead, 252 252 00:08:35,443 --> 00:08:36,780 we want to add some redundancy 253 253 00:08:36,780 --> 00:08:38,256 and we'll use layer three redundancy 254 254 00:08:38,256 --> 00:08:40,420 using a virtual gateway. 255 255 00:08:40,420 --> 00:08:41,730 To create a virtual gateway, 256 256 00:08:41,730 --> 00:08:42,860 we need to use either 257 257 00:08:42,860 --> 00:08:45,810 the First Hop Redundancy Protocol, FHRP, 258 258 00:08:45,810 --> 00:08:49,280 or the Virtual Router Redundancy Protocol, VRRP. 259 259 00:08:49,280 --> 00:08:52,480 Now, the most commonly used First Hop Redundancy Protocol 260 260 00:08:52,480 --> 00:08:56,550 is known as HSRP or the Hot Standby Router Protocol. 261 261 00:08:56,550 --> 00:08:58,670 This is a layer three redundancy protocol 262 262 00:08:58,670 --> 00:09:01,520 that's used as a proprietary First Hop Redundancy Protocol 263 263 00:09:01,520 --> 00:09:03,020 in Cisco devices. 264 264 00:09:03,020 --> 00:09:06,270 HSRP is going to allow for an active and a standby router 265 265 00:09:06,270 --> 00:09:07,420 to be used together. 266 266 00:09:07,420 --> 00:09:08,253 And instead, 267 267 00:09:08,253 --> 00:09:10,380 we get a virtual router that's defined 268 268 00:09:10,380 --> 00:09:11,900 as our default gateway. 269 269 00:09:11,900 --> 00:09:14,420 The client devices like the workstations and servers 270 270 00:09:14,420 --> 00:09:16,410 are then going to be configured to use the virtual router 271 271 00:09:16,410 --> 00:09:17,600 as its gateway. 272 272 00:09:17,600 --> 00:09:20,350 When the PC communicates to the IP of the virtual router, 273 273 00:09:20,350 --> 00:09:23,035 the router will determine which physical router is active 274 274 00:09:23,035 --> 00:09:24,680 and which one is standby. 275 275 00:09:24,680 --> 00:09:27,650 And then, it forwards the traffic to that active router. 276 276 00:09:27,650 --> 00:09:29,330 If the active router goes down, 277 277 00:09:29,330 --> 00:09:31,730 the standby router will pick up the responsibility 278 278 00:09:31,730 --> 00:09:33,020 for that active router 279 279 00:09:33,020 --> 00:09:35,030 until the other router comes back online 280 280 00:09:35,030 --> 00:09:37,010 and takes over its job again. 281 281 00:09:37,010 --> 00:09:38,470 Now, with VRRP, 282 282 00:09:38,470 --> 00:09:40,550 the Virtual Router Redundancy Protocol, 283 283 00:09:40,550 --> 00:09:41,510 this is one that was created 284 284 00:09:41,510 --> 00:09:43,470 by the internet engineering task force. 285 285 00:09:43,470 --> 00:09:45,000 Is an open standard variant 286 286 00:09:45,000 --> 00:09:48,546 of the Hot Standby Router Protocol or HSRP. 287 287 00:09:48,546 --> 00:09:52,000 VRRP allows for one master or active router, 288 288 00:09:52,000 --> 00:09:55,290 and the rest can then be added in a cluster as backups. 289 289 00:09:55,290 --> 00:09:56,760 Unlike HSRP, 290 290 00:09:56,760 --> 00:09:58,560 where you can only have one router as active, 291 291 00:09:58,560 --> 00:09:59,930 and one as standby. 292 292 00:09:59,930 --> 00:10:03,240 VRRP is going to allow you to have multiple standby routers. 293 293 00:10:03,240 --> 00:10:04,600 Just like HSRP, 294 294 00:10:04,600 --> 00:10:06,590 you're going to configure the VRRP 295 295 00:10:06,590 --> 00:10:08,040 to create a virtual router 296 296 00:10:08,040 --> 00:10:09,530 that's going to be used as a default gateway 297 297 00:10:09,530 --> 00:10:11,510 for all of your client devices. 298 298 00:10:11,510 --> 00:10:13,890 Now, in order to provide load balancing on your networks 299 299 00:10:13,890 --> 00:10:15,410 and to increase both redundancy 300 300 00:10:15,410 --> 00:10:17,030 and performance of your networks, 301 301 00:10:17,030 --> 00:10:18,980 you can use GLBP, 302 302 00:10:18,980 --> 00:10:21,300 which is the Gateway Load Balancing Protocol, 303 303 00:10:21,300 --> 00:10:25,930 or you can use LACP the Link Aggregation Control Protocol. 304 304 00:10:25,930 --> 00:10:29,420 Now, GLBP or the Gateway Load Balancing Protocol 305 305 00:10:29,420 --> 00:10:30,760 is a Cisco protocol, 306 306 00:10:30,760 --> 00:10:34,280 and it's another proprietary First Hop Redundancy Protocol. 307 307 00:10:34,280 --> 00:10:37,570 Now, GLBP will allow us to create a virtual router 308 308 00:10:37,570 --> 00:10:39,800 and that virtual router will have two routers 309 309 00:10:39,800 --> 00:10:40,820 being placed behind it, 310 310 00:10:40,820 --> 00:10:43,130 in active and standby configuration. 311 311 00:10:43,130 --> 00:10:45,690 The virtual router or gateway will then forward traffic 312 312 00:10:45,690 --> 00:10:47,460 to the active or standby router 313 313 00:10:47,460 --> 00:10:49,690 based on which one has the lower current loading, 314 314 00:10:49,690 --> 00:10:51,690 when the gateway receives that traffic. 315 315 00:10:51,690 --> 00:10:53,280 If both can support the loading, 316 316 00:10:53,280 --> 00:10:55,760 then the GLBP will send it to the active 317 317 00:10:55,760 --> 00:10:58,010 since it's considered the primary device. 318 318 00:10:58,010 --> 00:10:58,952 By using GLBP, 319 319 00:10:58,952 --> 00:11:01,250 you can increase the speeds of your network 320 320 00:11:01,250 --> 00:11:04,650 by using load balancing between two routers or gateways. 321 321 00:11:04,650 --> 00:11:05,960 Now, the second thing we can use, 322 322 00:11:05,960 --> 00:11:09,510 is LACP or Link Aggregation Control Protocol. 323 323 00:11:09,510 --> 00:11:12,450 This is a redundancy protocol that's used at layer two. 324 324 00:11:12,450 --> 00:11:14,340 So we're going to be using this with switches. 325 325 00:11:14,340 --> 00:11:17,730 LACP is going to achieve redundancy by having multiple links 326 326 00:11:17,730 --> 00:11:19,090 between the network devices, 327 327 00:11:19,090 --> 00:11:21,910 where load balancing over multiple links can occur. 328 328 00:11:21,910 --> 00:11:23,550 The devices are all going to be considered 329 329 00:11:23,550 --> 00:11:25,620 part of a single combined link 330 330 00:11:25,620 --> 00:11:27,510 even when we have multiple links. 331 331 00:11:27,510 --> 00:11:30,390 This gives us higher speeds and increases our bandwidth. 332 332 00:11:30,390 --> 00:11:31,240 For example, 333 333 00:11:31,240 --> 00:11:33,890 let's pretend I have four Cat5 cables. 334 334 00:11:33,890 --> 00:11:36,030 Each of these are connected to the same switch. 335 335 00:11:36,030 --> 00:11:38,640 Now, each of those cables has 100 megabits per second 336 336 00:11:38,640 --> 00:11:39,473 of bandwidth. 337 337 00:11:39,473 --> 00:11:42,350 Now, if I use the Link Aggregation Control Protocol, 338 338 00:11:42,350 --> 00:11:45,270 I can bind these altogether and aggregate them 339 339 00:11:45,270 --> 00:11:47,550 to give me 400 megabits per second 340 340 00:11:47,550 --> 00:11:49,550 of continuous bandwidth by creating 341 341 00:11:49,550 --> 00:11:51,510 this one single combined bandwidth, 342 342 00:11:51,510 --> 00:11:53,260 from those four connections. 343 343 00:11:53,260 --> 00:11:54,870 Now, let's consider what happens 344 344 00:11:54,870 --> 00:11:57,280 when that is trying to leave our default gateway 345 345 00:11:57,280 --> 00:11:58,810 and get out to the internet. 346 346 00:11:58,810 --> 00:11:59,910 Now in your home, 347 347 00:11:59,910 --> 00:12:02,220 you probably only have one internet connection, 348 348 00:12:02,220 --> 00:12:03,550 but for a company, 349 349 00:12:03,550 --> 00:12:05,730 you may wish to have redundant paths. 350 350 00:12:05,730 --> 00:12:07,260 For example, at my office, 351 351 00:12:07,260 --> 00:12:09,380 we have three different internet connections. 352 352 00:12:09,380 --> 00:12:11,380 The first is a microwave link that operates 353 353 00:12:11,380 --> 00:12:14,750 at 215 megabits per second for uploads and downloads. 354 354 00:12:14,750 --> 00:12:16,690 The second, is a cable modem connection. 355 355 00:12:16,690 --> 00:12:19,310 It operates at 300 megabits per second for downloads 356 356 00:12:19,310 --> 00:12:21,690 and 30 megabits per second for uploads. 357 357 00:12:21,690 --> 00:12:23,350 Now the third is a cellular modem, 358 358 00:12:23,350 --> 00:12:25,410 and that gives me about 100 megabits per second 359 359 00:12:25,410 --> 00:12:28,660 for downloads and about 30 megabits per second for uploads. 360 360 00:12:28,660 --> 00:12:30,400 Now, the reason I have multiple connections, 361 361 00:12:30,400 --> 00:12:33,140 is to provide us with increased speed and redundancy. 362 362 00:12:33,140 --> 00:12:34,720 So, to achieve this, 363 363 00:12:34,720 --> 00:12:36,203 I take all three of these connections 364 364 00:12:36,203 --> 00:12:38,340 and connect them to a single gateway 365 365 00:12:38,340 --> 00:12:40,060 that's going to act as a load balancer. 366 366 00:12:40,060 --> 00:12:42,070 If all the connections are up and running, 367 367 00:12:42,070 --> 00:12:43,540 they're going to load balance my traffic 368 368 00:12:43,540 --> 00:12:45,220 across all three of those connections 369 369 00:12:45,220 --> 00:12:47,690 to give me the highest speeds at any given time. 370 370 00:12:47,690 --> 00:12:49,990 But, if one of those connections drops, 371 371 00:12:49,990 --> 00:12:52,040 the load balancer will remove it from the pool 372 372 00:12:52,040 --> 00:12:53,810 until it can be returned to service. 373 373 00:12:53,810 --> 00:12:54,790 By doing this, 374 374 00:12:54,790 --> 00:12:58,380 I can get a maximum speed of about 615 megabits per second 375 375 00:12:58,380 --> 00:12:59,680 for a combined download. 376 376 00:12:59,680 --> 00:13:00,850 And on the upload, 377 377 00:13:00,850 --> 00:13:03,160 I can get about 310 megabits per second, 378 378 00:13:03,160 --> 00:13:04,790 when using all three connections 379 379 00:13:04,790 --> 00:13:06,630 and they're all functioning in online. 380 380 00:13:06,630 --> 00:13:07,463 Similarly, 381 381 00:13:07,463 --> 00:13:09,440 you might be in an area where you can get fiber connections 382 382 00:13:09,440 --> 00:13:10,273 to your building. 383 383 00:13:10,273 --> 00:13:11,260 Now, in those cases, 384 384 00:13:11,260 --> 00:13:13,890 you may purchase a primary and a backup connection. 385 385 00:13:13,890 --> 00:13:15,130 And if you do, 386 386 00:13:15,130 --> 00:13:17,460 you should buy them from two different providers. 387 387 00:13:17,460 --> 00:13:19,800 If both of your connections are coming from the same company 388 388 00:13:19,800 --> 00:13:20,730 and they go down, 389 389 00:13:20,730 --> 00:13:21,670 well, guess what? 390 390 00:13:21,670 --> 00:13:23,580 You just lost both of your connections 391 391 00:13:23,580 --> 00:13:26,080 because the upstream ISP might be down. 392 392 00:13:26,080 --> 00:13:26,913 For this reason, 393 393 00:13:26,913 --> 00:13:29,450 it's always important to have diversity in your path 394 394 00:13:29,450 --> 00:13:31,240 when you're creating link redundancy, 395 395 00:13:31,240 --> 00:13:32,620 just like I did in my office. 396 396 00:13:32,620 --> 00:13:34,830 I have a microwave connection through one ISP. 397 397 00:13:34,830 --> 00:13:36,430 I have a cable modem through another ISP, 398 398 00:13:36,430 --> 00:13:38,810 and I have a cellular modem through a third ISP. 399 399 00:13:38,810 --> 00:13:39,643 That way, 400 400 00:13:39,643 --> 00:13:40,490 if anyone goes down, 401 401 00:13:40,490 --> 00:13:42,950 I still have two other paths I can use. 402 402 00:13:42,950 --> 00:13:45,330 Now, the final type of redundancy that we need to discuss, 403 403 00:13:45,330 --> 00:13:47,270 is known as multipathing. 404 404 00:13:47,270 --> 00:13:50,130 Multipathing is used in our storage area networks. 405 405 00:13:50,130 --> 00:13:53,020 Multipathing is used to create more than one physical path 406 406 00:13:53,020 --> 00:13:55,110 between the server and its storage devices. 407 407 00:13:55,110 --> 00:13:56,930 And this allows for better fault tolerance 408 408 00:13:56,930 --> 00:13:58,570 and performance enhancements. 409 409 00:13:58,570 --> 00:14:00,350 Basically, think of multipathing 410 410 00:14:00,350 --> 00:14:02,060 as a form of link aggregation, 411 411 00:14:02,060 --> 00:14:03,730 but instead of using it for switches, 412 412 00:14:03,730 --> 00:14:06,180 we're going to use it for our storage area networks.