1 00:00:00,350 --> 00:00:02,220 To create redundancy for our servers, 2 00:00:02,220 --> 00:00:04,790 we're going to use a concept known as clustering. 3 00:00:04,790 --> 00:00:06,740 A cluster is when you take two or more servers 4 00:00:06,740 --> 00:00:08,010 and have them work together 5 00:00:08,010 --> 00:00:10,180 to perform a particular job function. 6 00:00:10,180 --> 00:00:12,880 We can cluster our servers as either failover clusters, 7 00:00:12,880 --> 00:00:14,830 or load-balancing clusters. 8 00:00:14,830 --> 00:00:16,660 When we talk about failover clusters, 9 00:00:16,660 --> 00:00:19,400 failover clusters are concerned with high availability. 10 00:00:19,400 --> 00:00:22,130 They're designed so the secondary server can take over 11 00:00:22,130 --> 00:00:23,810 in case the primary one fails 12 00:00:23,810 --> 00:00:25,890 with limited or no down time. 13 00:00:25,890 --> 00:00:28,720 A good example of this is your active domain controllers. 14 00:00:28,720 --> 00:00:31,310 You have a DC1 and a DC2. 15 00:00:31,310 --> 00:00:33,570 DC1 acts as the primary server, 16 00:00:33,570 --> 00:00:37,460 and if DC1 goes down, then DC2 would take over. 17 00:00:37,460 --> 00:00:39,210 Also, you could do this with a mail server, 18 00:00:39,210 --> 00:00:41,580 having a primary and a secondary exchange server. 19 00:00:41,580 --> 00:00:42,960 If the first server goes down, 20 00:00:42,960 --> 00:00:44,530 the second one takes over. 21 00:00:44,530 --> 00:00:45,610 That's the idea we're talking about 22 00:00:45,610 --> 00:00:47,670 when we're dealing with failover clustering. 23 00:00:47,670 --> 00:00:49,280 Now, the second type of clustering is known 24 00:00:49,280 --> 00:00:51,020 as load-balancing clustering. 25 00:00:51,020 --> 00:00:52,770 Load-balancing clustering is focused 26 00:00:52,770 --> 00:00:55,090 on servers sharing their resources, 27 00:00:55,090 --> 00:00:56,870 such as their processing power, 28 00:00:56,870 --> 00:00:58,820 their memory, and their hard disks. 29 00:00:58,820 --> 00:01:00,480 For example, Microsoft has a thing 30 00:01:00,480 --> 00:01:01,940 called the cluster server. 31 00:01:01,940 --> 00:01:04,630 This allows for parallel, high-performance computing 32 00:01:04,630 --> 00:01:06,550 where multiple servers can all be used 33 00:01:06,550 --> 00:01:08,100 as part of a server farm 34 00:01:08,100 --> 00:01:10,590 to conduct processing towards a common need. 35 00:01:10,590 --> 00:01:12,680 For example, let's say I had an application 36 00:01:12,680 --> 00:01:15,540 and I wanted to break some kind of encryption or password. 37 00:01:15,540 --> 00:01:17,510 I might use clustered servers to do that 38 00:01:17,510 --> 00:01:20,140 because it allows me to combine all of that processing power 39 00:01:20,140 --> 00:01:21,500 from all of those machines 40 00:01:21,500 --> 00:01:22,800 to help me crack the encryption 41 00:01:22,800 --> 00:01:25,290 or that password much, much faster. 42 00:01:25,290 --> 00:01:27,410 Load-balancing clusters are also commonly used 43 00:01:27,410 --> 00:01:29,870 to provide web servers for busy websites. 44 00:01:29,870 --> 00:01:32,780 For example, Facebook.com is really busy. 45 00:01:32,780 --> 00:01:35,650 They can't have a single server handling all that load. 46 00:01:35,650 --> 00:01:38,600 So they use a large cluster of load-balancing servers 47 00:01:38,600 --> 00:01:40,930 to do that, because that single server couldn't handle 48 00:01:40,930 --> 00:01:43,250 all those requests on its own. 49 00:01:43,250 --> 00:01:45,397 (techno music)