Tanubhrt.This.Side

Azure batch is used to run large scale applications in cloud. Without needing to modify the infrastructure, you can dynamically adjust resources. Azure batch is best when you don’t require a full control, however, need a cloud optimised solution for a high performing compute workload. It is effective for applications that run independently or need to communicate with each other.

When you are ready to run the job, azure batch performs the following activities –

• Starts a pool of compute virtual machines
• Installs application and staging data
• Runs jobs with as many as tasks as you have
• Identifies failures, requeues task, scales down the pool when work completes.

COMPONENTS OF AZURE BATCH

Pool – They are the compute resources for executing jobs on the batch.

Jobs – A job is a container designed to contain hundreds, thousands, and even millions of tasks.

Tasks – Tasks are individual units of work that comprise a job.

Node – A compute node is an azure virtual machine or cloud service VM that is dedicated to processing a portion of your application workload.

Things to consider when using Azure batch –

Consider Pools – Pools are best used when they are used dynamically. When creating your job if the job uses the same pool for everything, there are chances that when the pool runs into error, the job fails. If the job is for lesser amount of time, it is best advised to use the same pool rather than creating new pools, as creation of pools will be an over head.

Consider Nodes – Creating pools with multiple nodes is a good place to create reliability as if one of the nodes fails, the pool still can work through.

Consider Jobs – When creating jobs, try to create unique jobs so that you can uniquely monitor and log the activity. Further, having efficient distribution of tasks for a job is advisable. For example, creating 1 job with 1000 task is better than creating 10 jobs with 100 tasks.

Foolishly Yours,
AvantikAnmol