In the Article About Amazon Elastic Compute Cloud(EC2), we talked about the AWS Elastic Compute Cloud and its major benefits and usages. But there are different types of deployments of EC2 based on our business needs understanding the right method might help in cost-saving and also improving the performance so it's important to know the different types of AWS Pricing Models.
AWS Elastic Compute Cloud Pricing Models
The Classification depends on the types of usage, Types of storage attached to the EC2 instances, and also the periods for which you want your instance to be available for the customers. Apart from this classification, there are also some minute things which are needed to be considered like the operating system which you want to run on your instance or the memory required. This job is done by AMI or Amazon Machine Images, this is the images that are used in the time of making the instance the image consists of all the requirements like OS, Dependencies for a particular application, we can find the best AMI for our business needs.
So the Basic Classification of pricing models are done in 6 types they are:
1. On-Demand Instances:
In this instance type, the pricing is done on a per-hour basis or per-second basis. There is no need of doing long-term agreements in this mode, It is also eligible for AWS free tier. If you are a developer or Tester you want to test your application on the instances and are not aware of the workloads then this model is perfect for you because this is highly scalable and flexible according to the needs. As we are only using it for short lifespans and not long-term agreements the cost of this model is also less.
2. Reserved Instances:
When the workloads are known and to hold the instances which are needed for development we use reserved instances. This is mostly done in long-terms with a period of 1-3 years. The major benefits of using this instance are, as we are agreeing on a long-term we get a lot of discount on the instance around 50-70% off and also when we know the number of workloads we can properly reserve the number of instances and start working on it.
3. Spot Instances:
This is the instances that are not used in AWS they are available on highly less price compared to other instance almost up to 90% off, the only problem with this instance type is if someone wants to use this instance and has bided higher money or AWS wants to use the instance for another purpose then the instance will be no longer available to you. There will be a notification before 2 minutes of the termination. This is used by the developers who are in urgent need of instances to test their applications. This costing can also be done on an Hourly or second basis.
4. Dedicated Instances:
These are the instances usually run on a Virtual private network or in a network which are dedicated to a single customer or user. These instances are isolated from the other hardware instances. You can launch a dedicated instance with some web application that will be only available to you and you can see how the application run and will be configured for testing purposes.
5. Dedicated Hosts:
A Dedicated Host is Fully operational, a physical server dedicated only for our use. It can be used to run Enterprise servers like Linux, Microsoft Windows, or SQL servers. This is mostly required to have Strict regulatory requirements of software usages. As the instances are only dedicated for our usage we have more control over the instance and look at how actually it works and change according to our needs.
6. Scheduled Reserved Instances:
If you want to use instances in particular times or particles days or months of a year, then this instance type best suits you. We can Reserve an instance that we have scheduled and we get the instance at the right time, for example, we want 3 instances every day at 7:00 am then we can schedule it and can set the range or timing in which we want to use the instance. This mode is mostly in a 1year term pricing. This is most useful for the business needs which occur on particular hour, weeks or months of a year.
Tags:
Cloud_Computing