AWS Batch

AWS Batch is an AWS managed service that supports batch computing jobs. It can be run across EC2, Spot, and ECS systems depending on your desired cost profile. See more about the Cost Pillar in the AWS Well-Architected Framework.

See our coverage of other AWS Services, including Stackitecture Diagrams and Patterns.

Stackitecture Diagrams

We’ve included an example diagram for a batch processing architecture:

Patterns

High Performance Computing

Computer simulations are a good example for high performance computing when the results of one simulation do not impact the execution of another simulation run. In this example, you can feed in a number of different inputs, run all of the variations in batch, and then get a result dataset.

Batch Processing

There’s a number of high performance computing patterns enabled by batch processing. For example, bulk image processing can be rum as a batch workload. Assuming you have a source dataset of raw images and you want to have image processing algorithms run across the set, you can easily think about batching the process if the algorithm is independent.

AWS Documentation and Videos

We provide links into AWS Documentation we find useful for each service. Our you can start with your AWS Architecture Hub to guide you on your way.

• AWS Batch User Guide

Three-minute overview from Dr. Matt Wood

AWS re:Invent 2016 - Launch video for AWS Batch