Azure Compute and Storage: VMs, App Services, and Storage Options

Why This Topic Matters

Compute and storage are the building blocks of applications in Azure. For AZ-900, recognize the major service types, when to use them, and the operational implications (scalability, durability, cost).

Key considerations include:

• Workload profile (stateless vs stateful)
• Scale and performance requirements
• Durability and backup needs
• Operational model (PaaS vs IaaS)

Compute Options

Virtual Machines (IaaS)

• Full control over OS and runtime.
• Best for lift-and-shift, legacy apps, or when you need custom OS configuration.
• You manage OS updates, scaling, and availability.

App Service (PaaS)

• Managed platform for web apps, APIs and mobile backends.
• Built-in scaling, deployment slots, and platform patching.
• Lower operational overhead vs VMs.

Azure Functions (Serverless)

• Event-driven, pay-per-execution model.
• Good for small, bursty workloads and integrations.

Storage Options

Azure Storage (Blob, File, Queue, Table)

• Blob Storage: Object storage for unstructured data (hot/cool/archival tiers).
• File Storage: SMB/NFS-like file shares for lift-and-shift or shared files.
• Queue/Table: Messaging and simple NoSQL key-value data.

Managed Disks and Snapshots

• Disks are used by VMs; choose Premium/Standard based on IOPS and latency needs.
• Use snapshots/backups for point-in-time recovery.

Azure SQL / Cosmos DB (Managed Data)

• Platform-managed databases with built-in replication and SLA guarantees.

Practical Guidance for AZ-900 Candidates

1. Choose PaaS when you want to reduce operational overhead and focus on application logic.
2. Use VMs when you need full OS control or specific third-party software.
3. Match storage tier (hot/cool/archive) to access patterns to control cost.
4. Plan backups and replication for stateful services; verify SLAs for the chosen services.

Simple Decision Checklist

1. Is full OS control required? -> VM
2. Is rapid deployment and minimal ops desired? -> App Service or Functions
3. Is object storage or file semantics needed? -> Blob vs File
4. What are the performance and durability requirements? -> Disk type / redundancy

Common Mistakes to Avoid

• Using VMs by default without considering PaaS alternatives
• Storing frequently accessed data in archive tiers
• Neglecting backup and replication for stateful workloads

References (Microsoft Learn)

• Azure Compute documentation
• Azure Storage documentation

Final Thoughts

Understanding compute and storage trade-offs helps you choose the right services for cost, performance, and operational model—core knowledge for AZ-900 and practical cloud work.