Microsoft Azure Architect Design (AZ-304) Practice Test

What type of virtual machines should be used to minimize costs for Azure Batch jobs running a parallel, compute-intensive workload?

1. Basic A-series virtual machines

2. Low-priority virtual machines

3. Burstable virtual machines

4. Azure virtual machine sizes that support the Message Passing Interface (MPI) API

The correct answer is: Low-priority virtual machines

To minimize costs for Azure Batch jobs that involve parallel, compute-intensive workloads, low-priority virtual machines are a suitable choice. These virtual machines are designed to offer significant cost savings compared to standard virtual machines by utilizing unused capacity in the Azure data centers. Since they are considered "interruptible," they are suitable for batch processing workloads where job completion time is flexible and not mission-critical. Low-priority VMs are ideal for scenarios where cost efficiency is paramount, as they can be up to 80% cheaper than regular VM pricing, allowing users to run large-scale batch operations without incurring prohibitive costs. They can be allocated in bulk, which is essential for parallel workloads, allowing a large number of VMs to operate simultaneously while keeping expenditure in check. Using basic A-series or burstable virtual machines may not provide the necessary performance characteristics for compute-intensive workloads, as they are more suited for light workloads. While Azure virtual machine sizes that support the Message Passing Interface (MPI) API could facilitate certain types of parallel processing, they do not inherently address the cost aspect as effectively as low-priority VMs do.