We've covered three of the six advantages of cloud computing that deal with saving money. We're moving on now to the other three benefits to your operations and serving your customers. The first operational benefit is to stop guessing capacity. You either have or don't have the capacity to serve a new group of customers. On the left, the on-premises model gives you two choices. You can either 1. by too much idle capacity and let us sit idle, or 2. be unable to serve customers due to not having enough capacity. Neither of those are good outcomes. Here's a story about guessing about capacity. Target had Amazon build its own website, but then it left Amazon due to competitive concerns and it had other people build a different website. That website went down the 11 times during the holiday season and the Target executive in charge of it was fired. That is what happens when capacity estimations are not correct. On the right side is the cloud's approach to capacity. The AWS Cloud allows you to use elasticity to scale your capacity in or out to accommodate spikes and drops in-demand automatically, in minutes. The exam tests you on two types of scaling, horizontal scaling and vertical scaling. Horizontal scaling means that you automatically scale more of the instances you already have to meet demand, like in our Airbnb diagram. You can configure AWS to do this for you, and it happens automatically, seamlessly, with no negative impact to your consumers. Vertical scaling means that you upgrade your existing model. For example, you could put in a bigger hard drive or, in the cloud, you can switch from a t3.medium to a t3.large instance. This will cause a service interruption to your customers as you shut down a server to install a new hard drive, or as you boot up and configure your new t3.large instance. The exam will ask you at least once whether something is horizontal or vertical scaling. Speaking from the standpoint of not interrupting customers, horizontal scaling is better. The second operational benefit of the cloud is to increase speed and agility. On the exam, speed and agility mean two different things. Speed refers to how fast it takes for you to get a server or other resource into production for your customers to use. Agility refers to how fast you can experiment with something like machine learning and decide whether or not to use it or to stop using it for now. In the first row of this table, we see that the cloud has much greater speed than on-premises. On-premises, to get a new server purchased, delivered, installed, configured, and running takes six months to a year. But you can launch and use servers for yourself on the AWS Cloud in several minutes. In the second row, we see that the cloud also delivers greater agility than on-premises. I was reading an article about this machine learning software, that the publisher said had sold 1,000 licenses. The next sentence of the article said that half of those licenses were live. Think about what that means for a moment. It means that the software company sold half of the licenses to customers who no longer use the software, but are still legally obligated to continue paying for those licenses year after year until their contract expires. This sort of on-premises model should give you pause to buy licenses unless you're confident with the software. That diminishes your ability to simply pick up new software, experiment them for a few weeks, decide whether or not it's worthwhile. The AWS Cloud gives you an alternative. You can spin up a new server instance in minutes and test it. If you don't like it, you can terminate it and never have to pay for it again. Same with any other service, such as the machine learning service Amazon SageMaker. With the pay as you go model, you can use it as much as you want or as little as you want, and when you stop using it, you stop paying for it as well. Taken together, speed and agility allow you to pick up one of over 200 AWS services, figure out if it's right for you and spend very little money if it's not. This promotes the type of experimentation and innovation that can drive your business faster. Our third and final operational benefit is to go global in minutes. In the on-premises model, each time you want to serve customers in another part of the world quickly, you need to open up or lease space in a data center in another country, then ship your servers there and install them and so on. In the AWS Cloud, you can deploy two AWS regions around the world with just a few clicks, reducing latency and creating a better customer experience for your applications running in those regions. On the left is a map of AWS regions. The specific regions aren't important for the purposes of the exam because AWS is building and launching more regions all the time. The exam can't test you about regions because it can't keep up. Just know that AWS has many, many regions. On the right side, you can see AWS' South America infrastructure as an example of a continent with just one region, at least during this time of recording. Even if there's a single region here, it's served by many Amazon CloudFront Edge locations that deliver a cached content at faster speeds. For example, while the AWS region is in Sao Paulo, AWS maintains edge regions spanning the entire continent, from Rio de Janeiro to Bogota, to Buenos Aires, to Santiago. Come to think of it, many unicorn startups, which are defined as those achieving valuations of one billion dollars or more embraced all six benefits of Cloud computing to achieve their success. To learn more about these startups, and in particular, video conferencing unicorns Zoom, let's welcome our guest appearance from Dan Scheinman, former Cisco General Counsel and first investor in Zoom.
