WEBVTT 00:00.140 --> 00:00.680 Hey, everyone. 00:00.680 --> 00:01.690 And welcome back. 00:01.700 --> 00:06.940 Now, in today's video, we will be discussing about a very important topic of load balancers. 00:06.950 --> 00:12.740 In fact, in most of the organizations that you will see, Load Balancer is one of the key technology 00:12.740 --> 00:15.590 that is implemented across all of them. 00:15.590 --> 00:19.700 So let's go ahead and understand the basics of load balancers. 00:19.700 --> 00:25.040 Now, before we discuss about Load Balancer, let's look into a typical challenge that organizations 00:25.040 --> 00:27.860 used to face before this technology was introduced. 00:27.860 --> 00:34.100 So let's assume that you have a single server and organization has installed a application on this server, 00:34.100 --> 00:37.970 and this server itself is serving all of the production traffic. 00:38.120 --> 00:45.830 Now due to some reason, if this single server goes down, then the entire production traffic is affected. 00:45.830 --> 00:48.770 Now a server can go down due to any number of reasons. 00:48.770 --> 00:52.340 Maybe application is not working, maybe the kernel is not working. 00:52.340 --> 00:54.560 Maybe there is some issue related to hardware. 00:54.560 --> 00:59.300 Maybe there is some issue related to the internet that is connected to the server and so on. 00:59.300 --> 01:05.730 So the aim is that a single server going down should not really affect the production traffic. 01:05.730 --> 01:10.770 And in order to overcome this you have a recommended way of using a load balancer. 01:10.770 --> 01:15.180 So what happens is this revised approach is you create multiple servers. 01:15.180 --> 01:21.000 So now you have two servers and on top of that you make use of a load balancer. 01:21.000 --> 01:26.940 So load balancer is an additional component and this load balancer will distribute the traffic across 01:26.940 --> 01:27.930 both of the server. 01:27.930 --> 01:33.570 So now what will happen is any client who is visiting the website, the request will first go to the 01:33.570 --> 01:34.530 load balancer. 01:34.560 --> 01:40.740 Load Balancer will go ahead and distribute the request depending upon various algorithm that is been 01:40.740 --> 01:41.700 configured here. 01:41.700 --> 01:47.460 In a simplistic term, the request is distributed 5,050%, so first request will go to server one, 01:47.460 --> 01:49.710 second request will go to server two and so on. 01:49.710 --> 01:52.950 So this is a very high level overview about load Balancer. 01:52.950 --> 01:57.060 Now let's go ahead and quickly look into it at a demo perspective. 01:57.060 --> 02:02.720 So for today's demo, very similar to the diagram that we were discussing, we have two servers, we 02:02.720 --> 02:08.150 have server one, server two, and we have a new component of load balancer that is introduced. 02:08.180 --> 02:15.290 Now this load balancer again is basically a server with a software that can do a load balancing across 02:15.290 --> 02:17.120 both of the servers here. 02:17.120 --> 02:23.420 So now when a client wants to open a website, he will not really open up the server one or a server 02:23.420 --> 02:23.750 to. 02:23.780 --> 02:27.160 He'll basically the request will go to the load balancer IP. 02:27.410 --> 02:33.890 So in this case, all of the server has their individual IPS, but the client will basically open up 02:33.890 --> 02:36.050 the load balancer IP over here. 02:36.320 --> 02:41.960 Now before we go ahead and make a request to the load balancer IP, let's quickly open up the IP of 02:41.960 --> 02:43.040 server one, Server two. 02:43.040 --> 02:45.500 Let's look into what exactly is being loaded here. 02:45.500 --> 02:51.440 So for server one, I'll add the IP here and the response is very straightforward. 02:51.440 --> 02:53.270 It states this is server one. 02:53.270 --> 02:58.400 Similarly, let's open up the server two IP here, I'll paste it here. 02:58.400 --> 03:01.350 And again, the response is very straightforward. 03:01.350 --> 03:04.800 This is server two, so very straightforward response here. 03:04.800 --> 03:08.340 Now let's go ahead and make a request to the load balancer here. 03:08.340 --> 03:12.270 So from a new tab, I'll make a request to the load balancer IP. 03:12.540 --> 03:15.390 This time you see, the request went to the server one. 03:15.390 --> 03:18.090 So first request that we made to the load balancer. 03:18.090 --> 03:19.680 IP went to server one. 03:19.680 --> 03:21.660 Let's make a request again. 03:23.710 --> 03:26.770 And this time you see the request went to server two. 03:27.250 --> 03:28.750 Let's refresh. 03:28.780 --> 03:31.330 You see requests again went to server one. 03:31.930 --> 03:34.750 We'll refresh request again, went to server two. 03:34.780 --> 03:38.080 So this is what the load balancer is all about. 03:38.110 --> 03:44.830 Now one important part to remember is that in this specific demo I have used a specific server where 03:44.830 --> 03:47.140 a load balancer based software is installed. 03:47.170 --> 03:53.430 Now, if this load balancer server goes down, then again your entire application also goes down. 03:53.440 --> 03:59.920 So it is very important that whatever technology that you use at a load balancer level, you have to 03:59.920 --> 04:01.970 ensure that this is highly available. 04:01.990 --> 04:07.630 Now, depending upon various cloud provider that you might use, maybe Azure, et cetera, they have 04:07.630 --> 04:13.210 a highly available load balancing solution that is already available, so you do not really need to 04:13.210 --> 04:16.000 worry that the load balancer itself will go down. 04:16.030 --> 04:18.730 This is something that you do not need to worry. 04:18.760 --> 04:22.130 All you have to worry is about your applications itself. 04:22.150 --> 04:28.100 However, if you are hosting your own load balancing based solution in a virtual machine or in a server, 04:28.130 --> 04:33.260 then the high availability part of load balancer again goes into your responsibility. 04:33.290 --> 04:37.130 So before we conclude, let's quickly look into some of the important pointers here. 04:37.160 --> 04:43.280 First is that load Balancer does many things apart from just distributing the traffic among a set of 04:43.280 --> 04:43.940 servers. 04:43.970 --> 04:48.830 Now, one of the important feature of load balancer is health checking of the back end servers. 04:48.860 --> 04:53.470 For example, let's say that the server one is not working due to some reason. 04:53.480 --> 04:56.690 Load balancer can quickly identify that server. 04:56.690 --> 04:57.710 One is not working. 04:57.710 --> 05:01.200 Then what it will do, it will stop sending traffic to the server one. 05:01.220 --> 05:07.940 All the traffic will now go to the server two and after say after 15 minutes it realizes that server 05:07.940 --> 05:09.450 one is back up again. 05:09.470 --> 05:13.970 Then again, it will go ahead and send the traffic in a 5,050% based scenario. 05:13.970 --> 05:19.370 So health checking is one of the very important feature of a specific load balancer. 05:19.400 --> 05:24.980 Also note that there are multiple algorithms based on which the traffic can be distributed to the back 05:24.980 --> 05:25.660 end servers. 05:25.670 --> 05:28.340 So 5,050% is a very simplistic terms. 05:28.340 --> 05:34.520 You might need say 10% should go to server one because server one has a little less hardware size and 05:34.520 --> 05:38.890 90% of the traffic should go to server two because server two has a great hardware. 05:38.900 --> 05:44.000 So you can have X number of different configurations depending upon your use cases. 05:44.000 --> 05:50.930 And along with that load balancer can also support SSL termination and various other features as well. 05:50.960 --> 05:54.350 So this is a very high level overview about the load balancers. 05:54.350 --> 05:56.600 And with this, we'll conclude today's video.