1 00:00:00,110 --> 00:00:01,160 ‫So now let's talk about 2 00:00:01,160 --> 00:00:03,440 ‫the AWS Global Accelerator. 3 00:00:03,440 --> 00:00:07,410 ‫It is used to improve the global application availability 4 00:00:07,410 --> 00:00:10,800 ‫and performance using the AWS global network. 5 00:00:10,800 --> 00:00:13,270 ‫The idea is that your requests are going to be routed 6 00:00:13,270 --> 00:00:16,570 ‫through the internal network we saw from before from AWS 7 00:00:16,570 --> 00:00:19,260 ‫and this is going to allow you to optimize the route 8 00:00:19,260 --> 00:00:22,520 ‫to your application for about 60%. 9 00:00:22,520 --> 00:00:23,400 ‫So lets take an example, 10 00:00:23,400 --> 00:00:26,500 ‫We have deployed an application browser in India 11 00:00:26,500 --> 00:00:28,310 ‫and users all around the world 12 00:00:28,310 --> 00:00:30,630 ‫want to access our application. 13 00:00:30,630 --> 00:00:32,790 ‫What they will do using the Google Accelerator 14 00:00:32,790 --> 00:00:34,880 ‫is actually connect to an edge location 15 00:00:34,880 --> 00:00:37,930 ‫and the edge location will be routing the traffic 16 00:00:37,930 --> 00:00:40,110 ‫directly into India. 17 00:00:40,110 --> 00:00:41,798 ‫Well, the benefit we get out of it is that 18 00:00:41,798 --> 00:00:45,570 ‫the traffic on the public internet only happens 19 00:00:45,570 --> 00:00:48,700 ‫between America and the closest edge location 20 00:00:48,700 --> 00:00:51,890 ‫and then it leverages the private AWS network 21 00:00:51,890 --> 00:00:54,810 ‫to speed up the connection from the Edge location 22 00:00:54,810 --> 00:00:56,690 ‫to your application browser. 23 00:00:56,690 --> 00:00:57,910 ‫So this is the same for Europe 24 00:00:57,910 --> 00:00:59,713 ‫and this is the same for Australia. 25 00:01:01,216 --> 00:01:03,180 ‫On top of it, you only access your application 26 00:01:03,180 --> 00:01:05,838 ‫through two static IPs called Anycast IPs 27 00:01:05,838 --> 00:01:07,902 ‫and then using the static Anycast IP, 28 00:01:07,902 --> 00:01:10,640 ‫you will be redacted automatically 29 00:01:10,640 --> 00:01:12,200 ‫to the correct edge location. 30 00:01:12,200 --> 00:01:14,618 ‫And then the edge location, as I said, 31 00:01:14,618 --> 00:01:16,420 ‫will be sending the traffic to your application. 32 00:01:16,420 --> 00:01:19,510 ‫So there's a diagram I really liked from the website of AWS 33 00:01:19,510 --> 00:01:22,440 ‫which is showing what happens without a Global Accelerator. 34 00:01:22,440 --> 00:01:26,290 ‫So your clients to get to your application in your region 35 00:01:26,290 --> 00:01:28,610 ‫can go through a lot of hops an a lot of network 36 00:01:28,610 --> 00:01:29,690 ‫and there could be problems 37 00:01:29,690 --> 00:01:32,100 ‫and there could be latencies added to it. 38 00:01:32,100 --> 00:01:34,170 ‫But if you use the Global Accelerator 39 00:01:34,170 --> 00:01:36,770 ‫then you just connect to an Edge location of AWS 40 00:01:37,640 --> 00:01:39,830 ‫and then very quickly from the Edge location 41 00:01:39,830 --> 00:01:41,970 ‫to the region you're connecting to, 42 00:01:41,970 --> 00:01:44,510 ‫it goes through the private AWS network, 43 00:01:44,510 --> 00:01:46,790 ‫which is much faster. 44 00:01:46,790 --> 00:01:49,080 ‫Finally, I also get questions around what is the difference 45 00:01:49,080 --> 00:01:51,230 ‫between Global Accelerator and CloudFront. 46 00:01:51,230 --> 00:01:54,230 ‫So they both use the global network of AWS 47 00:01:54,230 --> 00:01:56,030 ‫and the edge locations around the world. 48 00:01:56,030 --> 00:02:00,040 ‫And they will both integrate with Sheild for DDoS protection 49 00:02:00,040 --> 00:02:02,780 ‫but CloudFront is a content delivery network 50 00:02:02,780 --> 00:02:06,180 ‫which is used to cage content at the edge such as images, 51 00:02:06,180 --> 00:02:07,460 ‫videos, and websites 52 00:02:07,460 --> 00:02:10,240 ‫and the content is served at the edge with CloudFront. 53 00:02:10,240 --> 00:02:11,507 ‫Okay. 54 00:02:11,507 --> 00:02:12,340 ‫Because it caged at the edge location. 55 00:02:12,340 --> 00:02:14,870 ‫For Global Accelerator, there is no caching. 56 00:02:14,870 --> 00:02:16,900 ‫All the requests are going to be passed on 57 00:02:16,900 --> 00:02:17,980 ‫from the edge locations 58 00:02:17,980 --> 00:02:20,450 ‫back to your application in your regions. 59 00:02:20,450 --> 00:02:21,840 ‫And so therefore what it's going to do 60 00:02:21,840 --> 00:02:23,440 ‫is that it's going to improve the performance 61 00:02:23,440 --> 00:02:27,480 ‫of these queries over a wide range of TCP or UDP 62 00:02:27,480 --> 00:02:29,730 ‫which is amazing if you have an HTTP use case 63 00:02:29,730 --> 00:02:31,640 ‫if you require a static IP address 64 00:02:31,640 --> 00:02:34,150 ‫or if you require fast deterministic 65 00:02:34,150 --> 00:02:36,760 ‫regional failover and good performance. 66 00:02:36,760 --> 00:02:37,960 ‫So remember they're very different. 67 00:02:37,960 --> 00:02:39,950 ‫CloudFront is a content delivery network, 68 00:02:39,950 --> 00:02:41,530 ‫you cage content at the edge 69 00:02:41,530 --> 00:02:44,650 ‫and Global Accelerator is to make your request go faster 70 00:02:44,650 --> 00:02:47,670 ‫and go through the internal AWS network globally. 71 00:02:47,670 --> 00:02:49,540 ‫So I'd like to show you a small tool 72 00:02:49,540 --> 00:02:51,640 ‫which allows you to do a speed comparison 73 00:02:51,640 --> 00:02:53,853 ‫for AWS Global Accelerator. 74 00:02:54,750 --> 00:02:55,967 ‫Okay. 75 00:02:55,967 --> 00:02:58,230 ‫So this is a speed comparison tool which allows you to see 76 00:02:58,230 --> 00:03:00,690 ‫if Global Accelerator makes a difference 77 00:03:00,690 --> 00:03:01,830 ‫compared to the public internet. 78 00:03:01,830 --> 00:03:04,530 ‫So let's do a test of five megabytes. 79 00:03:04,530 --> 00:03:08,570 ‫And what this is going to do is to try to upload a file 80 00:03:08,570 --> 00:03:12,090 ‫of five megabytes or download it as well, sorry, 81 00:03:12,090 --> 00:03:14,450 ‫in different regions around the world 82 00:03:14,450 --> 00:03:18,110 ‫regardless of if it's going over direct internet 83 00:03:18,110 --> 00:03:20,030 ‫or global accelerator. 84 00:03:20,030 --> 00:03:21,370 ‫And so it does it region by region 85 00:03:21,370 --> 00:03:25,160 ‫as we can see in a US-based one, Northern Virginia, 86 00:03:25,160 --> 00:03:27,460 ‫if I wanted to download a file over the internet 87 00:03:27,460 --> 00:03:31,109 ‫it was done in over 1,293 millisecond. 88 00:03:31,109 --> 00:03:33,740 ‫Whereas using Global Accelerator, 89 00:03:33,740 --> 00:03:38,740 ‫it was 23% faster because it was only 1002 milliseconds. 90 00:03:39,490 --> 00:03:41,864 ‫So this is going to make a huge difference for regions 91 00:03:41,864 --> 00:03:43,850 ‫that are far away from me. 92 00:03:43,850 --> 00:03:46,770 ‫So I'm in Europe, so Northern Virginia is far away from me, 93 00:03:46,770 --> 00:03:47,972 ‫it's across the ocean. 94 00:03:47,972 --> 00:03:51,390 ‫Oregon as well, it's on the West Coast of the US 95 00:03:51,390 --> 00:03:53,940 ‫so it's across the ocean and across the US. 96 00:03:53,940 --> 00:03:57,030 ‫So I'm getting a 31% improvement of speed, 97 00:03:57,030 --> 00:03:59,429 ‫but Ireland, which is a region that's literally next to me 98 00:03:59,429 --> 00:04:02,770 ‫as we can see, we have the same performance 99 00:04:02,770 --> 00:04:04,140 ‫over the internet or Global Accelerator 100 00:04:04,140 --> 00:04:07,433 ‫because it says next door, same for Frankfurt, Tokyo, 101 00:04:07,433 --> 00:04:09,760 ‫I'm getting better performance. 102 00:04:09,760 --> 00:04:11,670 ‫Singapore I expect also a better performance 103 00:04:11,670 --> 00:04:14,270 ‫and Sydney I expect a huge performance differential. 104 00:04:14,270 --> 00:04:17,150 ‫So let's have a look at the results in a second, 105 00:04:17,150 --> 00:04:20,230 ‫but what this shows you is that using Global Accelerator 106 00:04:20,230 --> 00:04:22,550 ‫because it goes over the private network of AWS 107 00:04:22,550 --> 00:04:24,980 ‫as fast as early on as possible, 108 00:04:24,980 --> 00:04:26,930 ‫you are going to get better performance 109 00:04:28,033 --> 00:04:30,460 ‫and this is the whole idea behind Global Accelerator. 110 00:04:30,460 --> 00:04:34,620 ‫So Singapore, yes, was 34% faster, which was great. 111 00:04:34,620 --> 00:04:38,423 ‫So 5,000 milliseconds versus 3.4 seconds. 112 00:04:39,850 --> 00:04:44,850 ‫And to finish with Sydney, I'm seeing a 53% improvement 113 00:04:45,500 --> 00:04:47,100 ‫in speed with Global Accelerator. 114 00:04:47,100 --> 00:04:50,080 ‫So this really shows you the power of Global Accelerator 115 00:04:50,080 --> 00:04:51,750 ‫when you have global applications. 116 00:04:51,750 --> 00:04:52,980 ‫So that's it for this lecture. 117 00:04:52,980 --> 00:04:55,903 ‫I hope you liked it and I will see you in the next lecture.