AZ-900 Describe Azure architecture and services Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Azure Functions

  Why it's wrong here

  Azure Functions is a serverless, event-driven compute service that runs short-duration code in response to triggers like HTTP requests or queue messages. The default timeout is 5 minutes and the maximum is 10 minutes for the App Service plan, which is far too short for most HPC workloads that can run for hours or days. HPC also requires stateful execution, coordination across processes, and dedicated high-bandwidth connections between nodes, but Functions is stateless and scales out by spinning up isolated function instances. It is designed for rapid, ephemeral tasks, not for sustained parallel computation.

• ✓

  Azure Batch

  Why this is correct

  Azure Batch is the correct service because it is purpose-built for HPC-grade job scheduling and large-scale parallel and scientific computing. It creates and manages a pool of virtual machines, automatically scales them based on the number of queued tasks, and schedules work across hundreds or thousands of cores. Batch supports MPI workloads, GPU-enabled VMs, and both Linux and Windows operating systems, along with task dependencies, retries, and timeout constraints. It decouples the job submission layer from the underlying compute pool, allowing you to run massive simulations, rendering, or genomic analysis in a managed way.

• ✗

  Azure Container Instances

  Why it's wrong here

  Azure Container Instances (ACI) allows you to run a single container or a small group of containers on-demand with per-second billing, but it is not a cluster scheduler. ACI offers no ability to coordinate tens, hundreds, or thousands of containers working on a parallel computational problem. HPC requires job scheduling, task dependency management, and the ability to allocate a large, interconnected pool of compute nodes; ACI is designed for simple, isolated workloads like one-off scripts or small microservices. It also has strict limits on instance count and lacks specialized high-performance networking features like InfiniBand.

• ✗

  Azure App Service

  Why it's wrong here

  Azure App Service is a fully managed PaaS designed for hosting web applications, REST APIs, and mobile app backends. It provides built-in auto-scaling and load balancing, but its compute instances are optimized for stateless web tiers, not for high-performance computing. HPC workloads require tightly coupled clusters with high-memory, high-CPU, or GPU resources and low-latency inter-node communication, none of which are available in a standard App Service plan. Even if you scale multiple instances, App Service does not support distributed parallel computing or message-passing interfaces (MPI).