1 00:00:00,000 --> 00:00:02,340 ‫The fourth pillar in the Well-Architecture 2 00:00:02,340 --> 00:00:04,800 ‫Framework is the performance efficiency. 3 00:00:04,800 --> 00:00:05,633 ‫So what is it? 4 00:00:05,633 --> 00:00:07,840 ‫It includes the ability to use computing resources 5 00:00:07,840 --> 00:00:10,890 ‫efficiently to meet system requirements, and to maintain 6 00:00:10,890 --> 00:00:14,620 ‫that efficiency as demand changes and technologies evolve. 7 00:00:14,620 --> 00:00:15,580 ‫So it's all about adapting 8 00:00:15,580 --> 00:00:17,250 ‫and providing the best performance. 9 00:00:17,250 --> 00:00:18,960 ‫The design principles are simple. 10 00:00:18,960 --> 00:00:22,030 ‫First, you need to use advanced technologies, okay? 11 00:00:22,030 --> 00:00:24,480 ‫You need to democratize them, and basically, 12 00:00:24,480 --> 00:00:27,320 ‫as the services become available, maybe they can be helpful 13 00:00:27,320 --> 00:00:29,740 ‫for your product developments so track them. 14 00:00:29,740 --> 00:00:31,910 ‫You need to be able to go global in minutes 15 00:00:31,910 --> 00:00:33,880 ‫so if you need to deploy in multiple regions 16 00:00:33,880 --> 00:00:36,180 ‫it shouldn't last days, it should last minutes. 17 00:00:36,180 --> 00:00:37,780 ‫Maybe using cloud formation. 18 00:00:37,780 --> 00:00:39,170 ‫Use serverless infrastructure. 19 00:00:39,170 --> 00:00:41,100 ‫So that's the golden state. 20 00:00:41,100 --> 00:00:43,210 ‫So that means you don't manage any servers, 21 00:00:43,210 --> 00:00:46,130 ‫and everything scales for you, which is really awesome. 22 00:00:46,130 --> 00:00:47,900 ‫Experiment more often. 23 00:00:47,900 --> 00:00:50,830 ‫Maybe you have something working really well today, 24 00:00:50,830 --> 00:00:53,840 ‫but you think it won't scale to 10 times the load. 25 00:00:53,840 --> 00:00:56,750 ‫Experiment maybe try serverless architectures. 26 00:00:56,750 --> 00:00:57,950 ‫See if that works for you. 27 00:00:57,950 --> 00:00:59,430 ‫Basically, give it a go. 28 00:00:59,430 --> 00:01:00,900 ‫And mechanical sympathy. 29 00:01:00,900 --> 00:01:03,440 ‫So be aware of all the AWS services, 30 00:01:03,440 --> 00:01:05,290 ‫and that's really, really hard. 31 00:01:05,290 --> 00:01:07,720 ‫I mean, doing this course is a right way of doing it. 32 00:01:07,720 --> 00:01:10,110 ‫Reading some blogs is also the right way of doing it. 33 00:01:10,110 --> 00:01:11,920 ‫Even for me, it's really hard to track 34 00:01:11,920 --> 00:01:13,920 ‫of everything of AWS services, 35 00:01:13,920 --> 00:01:16,000 ‫but you still need to be on top of the game. 36 00:01:16,000 --> 00:01:18,020 ‫Because when new changes have to happen, 37 00:01:18,020 --> 00:01:19,500 ‫they can really change dramatically, 38 00:01:19,500 --> 00:01:21,320 ‫your solution architecture. 39 00:01:21,320 --> 00:01:24,120 ‫So for AWS services for performance efficiency, 40 00:01:24,120 --> 00:01:29,010 ‫well, selections, so Auto Scaling, Lamda, EBS, S3, RDS 41 00:01:29,010 --> 00:01:30,550 ‫so you have so many choices 42 00:01:30,550 --> 00:01:33,240 ‫of technology that scale a little different patterns. 43 00:01:33,240 --> 00:01:35,030 ‫So choose the right one for you. 44 00:01:35,030 --> 00:01:37,427 ‫Definitely, for example, Lamda needs to serverless, 45 00:01:37,427 --> 00:01:40,400 ‫Auto Scaling is going to be for more EC2, 46 00:01:40,400 --> 00:01:42,250 ‫EBS is when you know you need to have a disc, 47 00:01:42,250 --> 00:01:44,190 ‫but you can sort of manage performance 48 00:01:44,190 --> 00:01:46,310 ‫over time using GB2 or I01. 49 00:01:46,310 --> 00:01:48,510 ‫S3 if you want to scale globally, 50 00:01:48,510 --> 00:01:51,290 ‫RDS, maybe you wanna use it to provision a database 51 00:01:51,290 --> 00:01:52,680 ‫and maybe you want to migrate to Aurora. 52 00:01:52,680 --> 00:01:55,520 ‫So you have a lot of things to select from. 53 00:01:55,520 --> 00:01:57,570 ‫Review, so how do we review our performance? 54 00:01:57,570 --> 00:02:00,160 ‫Well, the cloud formation, making sure we get 55 00:02:00,160 --> 00:02:02,910 ‫exactly what we need before we create it. 56 00:02:02,910 --> 00:02:04,480 ‫And how do we keep updated 57 00:02:04,480 --> 00:02:06,140 ‫on all this performance improvements? 58 00:02:06,140 --> 00:02:08,750 ‫Then the AWS News Blog is also a really good one, 59 00:02:08,750 --> 00:02:11,860 ‫and that's a tricky one, yes, but do read the AWS News. 60 00:02:11,860 --> 00:02:13,860 ‫I do read them once in a while, once a week, 61 00:02:13,860 --> 00:02:16,570 ‫and really keep myself updated on what's coming up. 62 00:02:16,570 --> 00:02:17,440 ‫Monitoring. 63 00:02:17,440 --> 00:02:19,270 ‫So how do we know we are performing 64 00:02:19,270 --> 00:02:21,210 ‫really well and as expected? 65 00:02:21,210 --> 00:02:24,030 ‫Then CloudWatch with CloudWatch Alarms, CloudWatch metrics. 66 00:02:24,030 --> 00:02:26,350 ‫All these things CloudWatch dashboards can help you 67 00:02:26,350 --> 00:02:28,200 ‫understand better how things work. 68 00:02:28,200 --> 00:02:30,390 ‫AWS Lamda, as well. 69 00:02:30,390 --> 00:02:32,170 ‫Making sure that you don't throttle, 70 00:02:32,170 --> 00:02:34,320 ‫that your application Lambda function runs 71 00:02:34,320 --> 00:02:36,660 ‫in a minimal time, all that kind of stuff, 72 00:02:36,660 --> 00:02:37,720 ‫And tradeoffs. 73 00:02:37,720 --> 00:02:39,760 ‫So how do we make sure that we are doing 74 00:02:39,760 --> 00:02:41,100 ‫the right performance decision? 75 00:02:41,100 --> 00:02:43,730 ‫So RDS maybe versus Aurora. 76 00:02:43,730 --> 00:02:46,470 ‫ElastiCache if you want to improve real performance, 77 00:02:46,470 --> 00:02:47,670 ‫maybe using Snowball. 78 00:02:47,670 --> 00:02:51,190 ‫So Snowball, for example, will give us a lot of data 79 00:02:51,190 --> 00:02:53,550 ‫moving very fast, but it will take maybe a week 80 00:02:53,550 --> 00:02:54,610 ‫for the data to arrive. 81 00:02:54,610 --> 00:02:56,690 ‫So the tradeoff is, do we want the data right away 82 00:02:56,690 --> 00:02:59,230 ‫in the cloud and use all our network capacity 83 00:02:59,230 --> 00:03:02,150 ‫or do we wanna move that data through a track 84 00:03:02,150 --> 00:03:03,520 ‫and get this in a week from now. 85 00:03:03,520 --> 00:03:04,670 ‫So always Tradeoffs, right? 86 00:03:04,670 --> 00:03:06,830 ‫With ElastiCache, always a Tradeoff as well. 87 00:03:06,830 --> 00:03:09,480 ‫Do I want to have possibly outdated, 88 00:03:09,480 --> 00:03:12,160 ‫stale data in a cache but really improve performance? 89 00:03:12,160 --> 00:03:14,643 ‫Or do I wanna get the latest and not use ElastiCache? 90 00:03:14,643 --> 00:03:16,220 ‫CloudFront same thing. 91 00:03:16,220 --> 00:03:18,500 ‫It does cache stuff around the edges. 92 00:03:18,500 --> 00:03:21,650 ‫So if you use CloudFront, yes, you go global in minutes, 93 00:03:21,650 --> 00:03:24,130 ‫but you have the possibility of everything being cached 94 00:03:24,130 --> 00:03:26,490 ‫for one day on people's laptops. 95 00:03:26,490 --> 00:03:28,380 ‫So when you release an update to your websites, 96 00:03:28,380 --> 00:03:31,320 ‫maybe it will take time for people to get the new stuff. 97 00:03:31,320 --> 00:03:33,100 ‫So think about all these things. 98 00:03:33,100 --> 00:03:35,370 ‫Really, it's important, but overall performance 99 00:03:35,370 --> 00:03:38,660 ‫should be really in the middle of your thought process 100 00:03:38,660 --> 00:03:40,370 ‫as you do Solution Architecture. 101 00:03:40,370 --> 00:03:41,203 ‫Okay, that's it. 102 00:03:41,203 --> 00:03:42,730 ‫I will see you in the next lecture.