AWS：按需运行计算繁重的进程（Windows 可执行文件）

Question

I need to run a Windows executable based on an HTTP request. The process is computational heavy hence required large processing powers. I am using AWS ec2 instances so far to run these executions. C5.2xlarge, windows EC2 instance.

These requests are not frequent but have to be served without delay.

Issue here is

• We are wasting quite a lot of money just to keep that heavy instance running for those abrupt requests.
• There is no Linux alternative for this executable.

Is it possible to make it more efficient? Paying only for the execution and not the idle time. Any suggestions for architectural changes? Even on other cloud platforms will work for us. Note: I could not figure out if AWS Lambda or Batch processing is possible for this scenario.

Answer 1

I would suggest changing the architecture a bit to facilitate your case and save some money:

AWS Fargate.

This is how their solution architect describes it:

AWS Fargate is an easy way to deploy your containers on AWS. To put it simply, Fargate is like EC2 but instead of giving you a virtual machine you get a container. It's a compute engine that allows you to use containers as a fundamental compute primitive without having to manage the underlying instances.

Fargate service allows you to run containers on demand. Kind of a serverless for containers.

Assuming you have knowledge of containers (or can learn it) - all you have to do is to create a container with windows based image, that will contain all the executables you need.

Once created, you upload it to AWS ECR (Elastic Container Registry) where your image build will be stored and then, set up a new service in ECS Fargate (ECS = Elastic Container Service) specifying the CPU and memory requirements.

Once all was set up correctly and all is working - you'll be able to launch the containers similarly to how you'd do it with Lambda. Fargate will receive your request, bring the container up, do the computational job, and will shut down once completed.