Azure Batch Runs Jobs in VM Pools

Azure Batch, a service that is now in preview, allows customers to distribute workloads over groups of virtual machines (VMs) to leverage Azure’s scalability and utility computing model to run tasks that benefit from distributed computing power. The service could be useful for running jobs such as payroll processing, software builds and test passes, media file encoding, weather analysis, and scientific simulations. Batch handles tasks that comprise the traditional programming pattern of batch processing, such as managing and scaling compute instances, ingesting application and data files, and returning output files, but customers must write code to use the service.

Parallel Processing of Data Speeds Results

The main benefit of Batch is its ability to distribute a workload over a collection of VMs called a pool, which could range in scale from dozens to hundreds of thousands of VMs. The service has an auto-scale feature that allows customers to define how the service adds or shuts down VMs in a pool to arrive at results more quickly or avoid unnecessary costs.