This article is an overview of the various AWS instance types.
AWS has three different categories of ECMs: General Purpose, Compute Optimized, and Memory Optimized instances. This post looks at general purpose type ecms in detail to see how they compare with each other on price/performance scales for your long-term or short-term compute needs.
What is AWS instance?
An ECM is a virtual machine hosted on the AWS cloud, or in other words, an instance.
A user can create one with just three clicks of their mouse and pay for it by the hour when they need to access it.
With this service you get complete control over your computing environment without any upfront capital expenses. You also have some flexibility to scale up and down as needed too!
ECMs are available across four major groups: General Purpose (GP), Compute Optimized (CO), Memory Optimized (MO) and GPU instances types, each serving different needs for performance versus cost tradeoffs.
AWS instance types and functions
Memory Optimized ECMs are ideal for tasks with high memory requirements, such as data analytics and modeling.
GPU instances provide a significant performance boost over the other three types of ECMS if your workload can leverage it effectively.
Compute-optimized ECM is best used when you need to do a lot of processing without having too many storage needs or accessing frequently accessed files in disk cache all that often.
General purpose ecms suit most use cases because they offer good price/performance balances and have been tuned to handle various computing loads well at any scale level from individual startups up through large enterprises.
The following list will help you understand what each type has to offer in summary:
- GP Instances – These are designed for general purpose workloads and offer a wide range of capabilities to handle many types of computing loads.
- CO Instances – These are designed for compute-intensive applications that require high processing power but not necessarily high memory needs, such as scientific computation or modeling tasks with big data sets.
- MO Instances – This is the most affordable type with good price/performance tradeoffs for everyday use cases where your workload has heavy access to disks (such as working on local files). It also works well if you have workflows requiring frequent disk reads and writes.
- GPU Instances – If you need the best performance in terms of throughput when running tasks at scale without concern for storage costs, then this instance type provides significant speed boosts over ECMs in general.
How to choose the right AWS instances
Depending on the type of workload, you need to choose an ECM that suits your purpose.
For example:
- If you are running a data analytics or modeling program and it requires high memory needs then it is best suited for Memory Optimized ECMs;
- If you have workflows requiring frequent disk reads and writes such as when working with files from Amazon S-Byte store, then MO Instances will suit your purposes well;
- CO Instances can be used if you require higher processing power than General Purpose but not necessarily higher memory needs or storage costs like in GPU instances. It also works well if performing scientific computation tasks using big data sets. Depending on how much CPU utilization your application uses, this could be the best ECM used;
- GP Instances are designed for general purpose workloads and offer a wide range of capabilities to handle many types of computing loads.
If you need more memory than what is offered in General Purpose instances, then this type is not appropriate because it has limitations on how much RAM can be allocated to the instance.
Conclusion
We hope this article helps you make a decision about which ECM would be best for your project. It’s really important that as technology evolves we have access to new ways to process our workloads without spending so much money!–that way my workload doesn’t cause me too much stress about having chosen the wrong one! And this review helped make things more clear!
