Azure Batch uses parallel actions to process the workloads. To fully leverage this capability, applications need to be split up into single and stateless tasks, called multi-instance tasks. By default, a Batch task is executed on a single compute node. By enabling multi-instance tasks, the task is executed on multiple compute nodes.
Those multi-instance tasks are submitted to the Batch job and then distributed over the available nodes inside the Batch pool. Azure Batch automatically creates one primary task and several subtasks based on the multi-instance settings. The tasks that run in parallel then use Azure Storage to save and retrieve the data that's used by the different tasks:
Azure Batch jobs and tasks
Those ...
