Which EC2 option is good for our organisation?
Amazon Elastic Compute Cloud (Amazon EC2) is a web service interface that allows the user to utilize computing capacity at a cost. It reduces internal friction, the economic load on the computing environment and is easily resizable. The costing of EC2 is based on four tier system. But before identifying which one is right for your organisation, it is important to know what factors determine the pricing.
The number of hours your instance occupies on a server is a direct determinant of cost. From the minute it’s launched to the minute it is terminated, EC2 is costing you something. Its occupancy on Elastic IP is also included in clock hours. In case of On-Demand instances, they will cost you all of that at a fixed price, irrespective of turn downs. Amazon will not turn down anything unless you do it voluntarily.
Total number of Instances to be occupied relies on Machine Configuration and load balancing. If the user has to implement heavy machinery, means, with greater physical capacity, the user will have to consider this total capacity along with the region you chose with the plan. Reserved instances provide reserved capacity as per the availability zones. So if one is not sure about the exact requirement of the instance run time one can think of going for reserved instance billing.
Apart from the machine configuration, machine purchase type is also very important. Are other AWS products clubbed with EC2? Is S3 (Simple Storage Service) being used as a secondary storage place? Are other options such as Amazon Elastic Beanstalk, Cloud Watch, Virtual Private Cloud and AWS Identity and Access Management integrated with EC2? All these questions have to be thought about before clicking on to a single plan. Spot instances are helpful if any other service is added to EC2 as it can increase compute capacity on request.
Operating systems and software packages that can function for this service are Windows Server, SQL Server, and SUSE Linux Enterprise Server. Dedicated hosts allow you to pick ‘n’ choose the server and work on the same. For such type of dedicated working, dedicated hosts pricing is the only option. Some reservations can even be purchased at a very heavy discount over On-Demand. Other things to be considered are which options of computing you would be using — General Purpose, Compute Optimised, Memory Optimised, GPU optimized, Storage optimized, IO Optimised, or T2 performance. Each of these cases requires a different kind of run time and down time.
Summary
In general, if you are not sure of your requirement, then you may choose On-demand as a beginner. It doesn’t require any particular commitment and there is a fixed price. It keeps things straight.
Reserved instances are more useful for those who may require additional capacity in future, but, for now, they have lower needs and do not wish to pay for the capacity that will not be used presently. This is a very cost-effective approach.
In case, your instances have a high chance of flexibility or have urgent need of capacity, Spot pricing is rather more cost-effective.
Dedicated Hosts are for those who want to integrate their existing servers and license terms with additional capacity. They are for those who are looking to expand their present compute capacity.
Still confused over your chosen option. Feel free to contact us for a free analysis of your environment and requirement.