One of the reasons Amazon has found success in the cloud space is its really simplified pricing. The cost of AWS Elastic Cloud Compute, is kept simple and straightforward by allowing users to choose a pricing model that makes sense to them.
This allows them to spin up the exact kind of instances they want on EC2, without worrying about contracts and conventions.
AWS offers a pay-as-you-go approach for more than 70 of its cloud services. This means that you only pay for these as you need them, for as long as you need them, without requiring any long-term contract lockdowns or complex licensing. On top of that, the company also offers you deep savings of up to 75% for certain services like EC2 and RDS if you reserve things beforehand.
Before you get down to the business of calculating the cost, you should know that there are four pricing models that are prevalent across AWS. All four are differentiated by how you create your instances.
These are currently allocated by the hour, though the company also finally unveiled per second pricing for EC2 earlier this week that should result in further savings for its users.
Let’s have a quick look at the four EC2 instance pricing options.
This is pretty much the most commonly used instance deployment method. By far, On-Demand Instances are created only when you require them, hence the term. AWS makes available your chosen computational resources whenever you request them, and you have the option to shut them down anytime during the tenure. You only pay for what you use here, making these instances ideal for web hosting and file storage.
Deploying instances on-demand has one slight drawback, and that is that AWS does not guarantee the deployment. This is because Amazon has to ensure that the adequate capacity is present in its datacenters at all times. The dynamics involved here mean that in rare cases, AWS will fail to power on your on-demand instance. In which case, you are better off using what the company calls Reserved Instances, where AWS actually guarantees you resource capacity. Three pricing options are available here, you can pay all upfront, partial upfront or no upfront to reserve your instance. This can be done for a minimum period of a year, and maximum of up to three years.
Spot Instances are specifically created to address the problem of excess capacity in the Elastic Cloud Compute. They allow you to bid for this unused EC2 compute capacity, and it works just like any other bidding system. AWS sets the hourly price for a particular spot instance that changes as the demand for the spot instance either grows or shrinks. And when the bid you placed exceeds that of the current spot price, your instances are then made to run. Note that these instances will stop the moment someone outbids you. For this reason, they are better suited for applications that are not critical in nature, and don’t require large processing time. For example, image resizing or document format conversion.
And finally, the fourth and final type of instance that AWS launched not so long ago, are Dedicated Hosts. This is basically a physical EC2 server dedicated for your use, as opposed to a virtual one. These are great for reducing costs, as you can use your existing server-bound software licenses, Windows for example. They are also perfect for meeting compliance requirements. Dedicated Hosts can be purchased both on-demand or reserved beforehand.
You can also find out the complete details of what sorts of costs you are looking at for your AWS Elastic Cloud Compute adventure here.