1 00:00:00,060 --> 00:00:02,250 Okay, so in this lecture, we are going to talk about 2 00:00:02,250 --> 00:00:04,019 a newer kind of service for AWS called 3 00:00:04,019 --> 00:00:05,939 the AWS Global Accelerator. 4 00:00:05,939 --> 00:00:06,790 But before we do so, 5 00:00:06,790 --> 00:00:08,230 I want to introduce the problem to you 6 00:00:08,230 --> 00:00:10,750 that we're trying to solve and how we're going to solve it. 7 00:00:10,750 --> 00:00:13,180 So, say you have deployed and application, 8 00:00:13,180 --> 00:00:15,290 and it's global and you've global users 9 00:00:15,290 --> 00:00:17,190 who want to access it directly. 10 00:00:17,190 --> 00:00:19,880 But our application is only deployed in one region. 11 00:00:19,880 --> 00:00:21,740 So, for example here in India, 12 00:00:21,740 --> 00:00:24,250 I have deployed a public application advancer. 13 00:00:24,250 --> 00:00:26,020 But my users are all over the world. 14 00:00:26,020 --> 00:00:28,900 They're in America, in Europe, in Australia. 15 00:00:28,900 --> 00:00:30,710 And so as they access the application, 16 00:00:30,710 --> 00:00:32,619 they want to go over the public internet. 17 00:00:32,619 --> 00:00:34,520 And that can add a lot of latency 18 00:00:34,520 --> 00:00:36,790 due to many hops through the routers. Okay? 19 00:00:36,790 --> 00:00:38,510 So, I'm obviously overplaying this, 20 00:00:38,510 --> 00:00:40,030 so you can see that in America, 21 00:00:40,030 --> 00:00:42,270 there's five hops of routers or servers 22 00:00:42,270 --> 00:00:45,020 before we get to the public ALB in India. 23 00:00:45,020 --> 00:00:47,090 And that's because it goes over the public internet. 24 00:00:47,090 --> 00:00:48,650 Same for Australia, many hops. 25 00:00:48,650 --> 00:00:50,290 And same for Europe, many hops. 26 00:00:50,290 --> 00:00:53,920 So you can see these hops I introduce a bit of risk 27 00:00:53,920 --> 00:00:55,270 connection can get lost, 28 00:00:55,270 --> 00:00:56,605 they also add a little bit of latency, 29 00:00:56,605 --> 00:00:58,610 and they're not as direct as possible 30 00:00:58,610 --> 00:01:01,180 into our amazon infrastructure. 31 00:01:01,180 --> 00:01:04,280 So what we want to do is to wish to go as fast as possible 32 00:01:04,280 --> 00:01:07,770 through as soon as possible the AWS US network 33 00:01:07,770 --> 00:01:10,480 to minimize latency. And so before we do so 34 00:01:10,480 --> 00:01:12,630 and use the global accelerator for this, 35 00:01:12,630 --> 00:01:14,320 I want to introduce to you another concept 36 00:01:14,320 --> 00:01:17,250 which is Unicast and Anycast IP. 37 00:01:17,250 --> 00:01:20,010 So first, Unicast IP is what we know. 38 00:01:20,010 --> 00:01:22,050 One server holds one IP address. 39 00:01:22,050 --> 00:01:24,250 So our clients when they talk to two server 40 00:01:24,250 --> 00:01:26,210 one has starting IP address is the left hand side 41 00:01:26,210 --> 00:01:28,190 number 12 and the other one is 98. 42 00:01:28,190 --> 00:01:30,590 Then if you refer to the IP address that begin with 12 43 00:01:30,590 --> 00:01:33,434 we will be sent to the server on the left hand side. 44 00:01:33,434 --> 00:01:35,663 And for the other one if you use the other IP 45 00:01:35,663 --> 00:01:38,050 then we will go to the server on the right hand side. 46 00:01:38,050 --> 00:01:40,220 This we know it makes a lot sense. 47 00:01:40,220 --> 00:01:43,070 But for Anycast IP, is a little different. 48 00:01:43,070 --> 00:01:45,420 All servers will hold the same IP address 49 00:01:45,420 --> 00:01:49,350 and the client will be routed to the nearest one. 50 00:01:49,350 --> 00:01:51,470 So this is contra-intuitive but this is how it works. 51 00:01:51,470 --> 00:01:53,540 So our client has two servers 52 00:01:53,540 --> 00:01:55,440 and these two servers as you can see in the bottom, 53 00:01:55,440 --> 00:01:57,160 they have the same IP. 54 00:01:57,160 --> 00:02:00,319 But our client when it tries to connect to this Anycast IP 55 00:02:00,319 --> 00:02:04,510 will be sent to the server that is the closest to itself. 56 00:02:04,510 --> 00:02:07,350 And as surprising as this is, this is how it works. 57 00:02:07,350 --> 00:02:11,890 So Global Accelerator uses that Anycast IP concept to work. 58 00:02:11,890 --> 00:02:12,970 So how does it work? 59 00:02:12,970 --> 00:02:16,520 We're able to leverage the AWS internal global network 60 00:02:16,520 --> 00:02:18,380 to route to our application. 61 00:02:18,380 --> 00:02:19,610 So we have the same idea, 62 00:02:19,610 --> 00:02:21,270 we want to route to India 63 00:02:21,270 --> 00:02:23,530 and we have users all around the globe. 64 00:02:23,530 --> 00:02:26,510 And is going to happen is that instead of sending it 65 00:02:26,510 --> 00:02:28,610 through the public internet in America, 66 00:02:28,610 --> 00:02:31,560 it's going to come to the closest edge location. 67 00:02:31,560 --> 00:02:32,680 And from edge location, 68 00:02:32,680 --> 00:02:35,920 it's going to go all the way straight to our ALB 69 00:02:35,920 --> 00:02:38,380 through the internal AWS network. 70 00:02:38,380 --> 00:02:41,510 Same for Australia, so it goes to closest edge location 71 00:02:41,510 --> 00:02:42,690 near to Australia 72 00:02:42,690 --> 00:02:44,990 and then it goes over the private AWS network 73 00:02:44,990 --> 00:02:48,000 to get to the ALB and same for Europe. 74 00:02:48,000 --> 00:02:49,230 So the idea with this is that, 75 00:02:49,230 --> 00:02:51,140 we're going to use an Anycast IP 76 00:02:51,140 --> 00:02:52,042 and there's actually going to be two of those 77 00:02:52,042 --> 00:02:54,270 that are going to be created for your application 78 00:02:54,270 --> 00:02:55,450 and they're global. 79 00:02:55,450 --> 00:02:58,670 And the Anycast IP will send the traffic directly 80 00:02:58,670 --> 00:03:01,010 to the closest edge location of your users. 81 00:03:01,010 --> 00:03:03,230 And this is the power of Anycast IP. 82 00:03:03,230 --> 00:03:05,060 So the edge location will then send the traffic 83 00:03:05,060 --> 00:03:06,052 to you application, (mumbles) 84 00:03:06,052 --> 00:03:07,010 For example, 85 00:03:07,010 --> 00:03:10,150 through the private AWS network which is much more stable, 86 00:03:10,150 --> 00:03:12,410 has less latency and so on. 87 00:03:12,410 --> 00:03:14,650 So the Global Accelerator is absolutely unique 88 00:03:14,650 --> 00:03:17,960 because it really allows to give two static IP addresses 89 00:03:17,960 --> 00:03:19,660 all around the globe for the users 90 00:03:19,660 --> 00:03:21,750 for whatever application you may have. 91 00:03:21,750 --> 00:03:23,860 And right now I'm showing one ALB in one region 92 00:03:23,860 --> 00:03:25,180 but it could be global as well 93 00:03:25,180 --> 00:03:27,400 it could be multiple ALBs in multiple regions. 94 00:03:27,400 --> 00:03:29,020 So it's a game changer to me. 95 00:03:29,020 --> 00:03:30,240 So, what does it work with? 96 00:03:30,240 --> 00:03:33,200 It works with Elastic IP, EC2 instances, 97 00:03:33,200 --> 00:03:35,480 Application Load Balancer, Network Load Balancer 98 00:03:35,480 --> 00:03:38,360 and they can either be public or private. 99 00:03:38,360 --> 00:03:39,480 There is consistent performance 100 00:03:39,480 --> 00:03:41,070 because we go over the network 101 00:03:41,070 --> 00:03:42,240 so we have intelligent routing 102 00:03:42,240 --> 00:03:44,400 to the lowest latency edge location 103 00:03:44,400 --> 00:03:46,060 and we'll have fast regional failover 104 00:03:46,060 --> 00:03:47,490 in case anything goes wrong. 105 00:03:47,490 --> 00:03:48,890 There's no issue with client cache 106 00:03:48,890 --> 00:03:50,380 because the client doesn't cache anything 107 00:03:50,380 --> 00:03:54,350 the IP, the two Anycast IP we're using don't change. 108 00:03:54,350 --> 00:03:58,170 It's internal AWS network to go after the edge location 109 00:03:58,170 --> 00:03:59,170 so it's perfect. 110 00:03:59,170 --> 00:04:00,550 And then we have health checks. 111 00:04:00,550 --> 00:04:02,680 So the Global Accelerator will perform a health check 112 00:04:02,680 --> 00:04:04,640 on your application and then make sure 113 00:04:04,640 --> 00:04:06,390 that the application is global 114 00:04:06,390 --> 00:04:09,400 and if health check fails for one ALB and one region 115 00:04:09,400 --> 00:04:12,240 then there is automated failover in less than one minute 116 00:04:12,240 --> 00:04:13,710 to a healthy end port. 117 00:04:13,710 --> 00:04:15,170 So this is really, really great. 118 00:04:15,170 --> 00:04:17,040 And it's amazing for disaster recovery, 119 00:04:17,040 --> 00:04:18,450 thanks, to the health checks. 120 00:04:18,450 --> 00:04:19,640 Now for security, 121 00:04:19,640 --> 00:04:20,829 it's appropriately secured 122 00:04:20,829 --> 00:04:22,380 because we only have two external IP 123 00:04:22,380 --> 00:04:24,170 that needs to be whitelisted by your clients. 124 00:04:24,170 --> 00:04:26,600 And you get DDoS protection automatically 125 00:04:26,600 --> 00:04:28,070 through the Global Accelerator, 126 00:04:28,070 --> 00:04:28,948 thanks to something called 127 00:04:28,948 --> 00:04:31,820 AWS Shield that we'll see in the security session. 128 00:04:31,820 --> 00:04:32,653 So this is great. 129 00:04:32,653 --> 00:04:34,150 So you're really asking me what is the difference 130 00:04:34,150 --> 00:04:36,200 between Global Accelerator and CloudFront? 131 00:04:36,200 --> 00:04:37,860 And hopefully you already see the difference, 132 00:04:37,860 --> 00:04:40,440 otherwise I failed but I want to outline these differences 133 00:04:40,440 --> 00:04:41,870 very, very clearly. 134 00:04:41,870 --> 00:04:44,170 So, Global Accelerator and CloudFront 135 00:04:44,170 --> 00:04:47,300 they both use the same global network 136 00:04:47,300 --> 00:04:50,310 and they will both use edge locations all around the globe 137 00:04:50,310 --> 00:04:52,350 that AWS has created. 138 00:04:52,350 --> 00:04:55,220 They both integrate with Shield for DDoS protection 139 00:04:55,220 --> 00:04:56,400 so they both receive the same thing 140 00:04:56,400 --> 00:04:57,680 but now their difference. 141 00:04:57,680 --> 00:05:00,080 CloudFront is going to improve the performance 142 00:05:00,080 --> 00:05:03,150 for both cacheable content such as images and video, 143 00:05:03,150 --> 00:05:05,090 and also dynamic content such as, 144 00:05:05,090 --> 00:05:07,930 API acceleration and dynamic site delivery. 145 00:05:07,930 --> 00:05:09,780 And the content is going to be served 146 00:05:09,780 --> 00:05:11,660 from the edge locations. 147 00:05:11,660 --> 00:05:13,270 So once in a while the edge locations 148 00:05:13,270 --> 00:05:15,283 are going to fetch the content from the origin 149 00:05:15,283 --> 00:05:18,030 but most of the time hopefully CloudFront 150 00:05:18,030 --> 00:05:21,100 will deliver cache content from the edges. 151 00:05:21,100 --> 00:05:24,220 So here the users are getting content from the edges. 152 00:05:24,220 --> 00:05:25,750 Whereas, for Global Accelerator, 153 00:05:25,750 --> 00:05:28,020 it improves the performance of the wide range of application 154 00:05:28,020 --> 00:05:29,810 over TCP or UDP. 155 00:05:29,810 --> 00:05:33,124 But the packets are being proxyeed from the edge locations 156 00:05:33,124 --> 00:05:37,090 to the applications running in two one or more AWS regions. 157 00:05:37,090 --> 00:05:40,010 So in that case, all the request still make it 158 00:05:40,010 --> 00:05:41,120 to our application end. 159 00:05:41,120 --> 00:05:43,610 There is no caching available. 160 00:05:43,610 --> 00:05:46,470 So it's a really good fit if you have a non-HTTP uses cases, 161 00:05:46,470 --> 00:05:49,100 such as gaming, IoT or Voice over IP 162 00:05:49,100 --> 00:05:51,420 or it's also really good if you have and HTTP use case 163 00:05:51,420 --> 00:05:54,800 that require a static IP addresses globally. 164 00:05:54,800 --> 00:05:55,633 Okay? 165 00:05:55,633 --> 00:05:57,520 Or in case you need to have deterministic 166 00:05:57,520 --> 00:05:59,320 and fast regional failover. 167 00:05:59,320 --> 00:06:01,300 So, Global Accelerator is a newer kind of service 168 00:06:01,300 --> 00:06:03,100 and it will appear in your exam. 169 00:06:03,100 --> 00:06:04,060 I hope you like this lecture. 170 00:06:04,060 --> 00:06:06,313 I will see you in the next lecture for some hands on.