- [Alana] We've already spoken a bit about how AWS defines their pricing model with pay-as-you-go pricing and an on-demand, utility-like structure with no upfront costs or long-term financial requirements. But the best way to understand AWS pricing options is to use an example. Let's say your infrastructure team wants to move a small website over to AWS. This website is currently hosted on servers on premises. The team mentions they want to use the Amazon Elastic Compute Cloud, or Amazon EC2, service to host this website in AWS. This service, Amazon EC2, provides scalable virtual servers, called EC2 instances, that you can essentially rent. Now that we know what the team wants, we have to think about the best pricing options for this scenario. The first consideration is that if the team wants to spin up those four EC2 instances on demand, meaning start them up whenever they'd like shut them down whenever they'd like, and pay only for the length of time they're running, they can do that. The nice thing about running these instances on demand is that it gives you flexibility. You can shut them down, spin up a larger EC2 instance, or change the type of instance you want to run, all without having to worry about commitment. The downside is that if you're running these instances continuously, 24/7, you're less likely to benefit from the On-Demand pricing model because there are a few situations where you'll need to turn the instances off. And if the infrastructure team is planning to run these instances continuously for a 1-year or 3-year term, then there's a more cost-effective pricing model to consider, called Reserved Instances, or RIs. When people hear the name Reserved Instance, most get confused about what it actually means. RIs are not actually physical instances. They are a billing discount that could be applied to instances you have running in your account. So, if you had four instances running on demand in your account, you can purchase RIs that match the attributes of those four instances, such as the instance type and Region, and that billing discount will be immediately applied to them. The upside of RIs is that you have flexibility in how to pay for your instances. With RIs, you have three main payment options. The first is All Upfront, meaning the full payment for the instance and the term you selected is paid upfront, and you incur no other costs for the remainder of the term. The second, which is Partial Upfront, which means that you pay a portion upfront, and the remaining costs are billed at a discounted hourly rate. And then last, No Upfront, which means that you pay nothing upfront and are billed at a discounted hourly rate for the rest of your term. Generally speaking, the more you're willing to pay upfront, the more money you will save in the long run. Now, the downside of RIs is that you pay for them regardless of if you're actually using them. This means that it's your responsibility to monitor your RIs to make sure they're applied to the instances within your account. It's possible that someone accidentally changes an instance's characteristics, and then the RI will no longer be applied, and you're missing out on savings. Good news is, there are ways with governance where you can ensure that the discount is always applied to your instances. We'll talk more about this in Week 2, in the Governance section. If you don't want to perform this monitoring or create automation to do it, there is another discount called Savings Plans that will reduce your overall bill. Savings Plans will provide you a discount in exchange for a commitment of usage, measured in dollars per hour over a 1-year or 3-year period. For example, if I commit to $20 an hour for my usage, then I'm charged at the Savings Plans rate until that $20. If I go over $20, then I'm charged at the On-Demand rate. The upside of using Savings Plans is that the coverage is flexible. Meaning, you can change your instance attributes, such as the instance type or the Region the instance is in, and the discount will still apply. And, they can also be applied to other types of compute options that AWS offers, such as AWS Lambda, a serverless compute solution. So now, we have to make a suggestion for the infrastructure team. Do we suggest On-Demand Instances, Reserved Instances, or Savings Plans? Well, if the infrastructure team is willing to commit to a 1-year or 3-year term period, has a vague idea of their usage patterns, but wants the flexibility of changing Regions and compute type, I would recommend using Savings Plans. In this video, we've mainly focused on pricing through Amazon EC2, as that can often be the most confusing, but other services have unique pricing options that you may need information for as well. Fortunately, each service has a pricing page, and you can find it by appending the service name and then the keyword pricing to the AWS URL, or just searching for the service name, plus pricing. It's important that you get comfortable with finding this information easily, as pricing can really affect your procurement. There's a lot of different ways to pay. AWS pricing is very different than traditional IT vendor pricing, and, overall, an understanding of what you're requesting of AWS or your AWS Partner can only help you. It's worth mentioning that AWS offers other discounts outside of the individual services as well. Check out the readings for more information on pricing details, discounts, and more.
