AWS Batch

Concept

With AWS Batch you can run very large or very many computing tasks (batch jobs) in a queue, without worrying about servers. You submit a job, and it waits in a job queue. Batch itself decides which job runs when, and how much computing it needs. The jobs run in a compute environment, which can be made of EC2, Spot Instances, Fargate or EKS. How each job runs is written in a job definition.

Simple example

Dropping off clothes at a laundry shop. You hand over the clothes and leave; the shop splits them across its machines and washes them when it suits it.

Why it’s needed

Working out how many servers you need to run thousands of jobs at once, and which job runs where, is hard. Batch does it for you.

Where it’s used

Processing millions of images every night, scientific simulations, and generating large reports.

Azure equivalent

Azure Batch

Remember

Batch = a manager for queues of big jobs. When the work is done, the servers stop too.

Learn more

For details and the latest information, see the official documentation:

• AWS Batch — AWS official documentation ↗