VM options and required resources¶
Virtual machines can be deployed individually or in patterns that improve scaling and resilience. These three options appear often on AZ-900 scenario questions.
Resources required for a virtual machine¶
An Azure VM depends on supporting resources and configuration choices. At AZ-900 level, recognize these components:
| Component | Purpose |
|---|---|
| VM image and operating system | Defines the starting OS and software image |
| VM size | Determines vCPUs, memory, storage limits, and network capacity |
| Managed OS disk | Persistent disk containing the operating system |
| Optional data disks | Persistent storage for application or business data |
| Virtual network and subnet | Provide the VM's private network location |
| Network interface (NIC) | Connects the VM to the virtual network |
| Private IP address | Enables private communication inside connected networks |
| Optional public IP address | Enables direct internet communication when required |
| Network security group (NSG) | Allows or denies network traffic by rules such as source, port, and protocol |
The VM, disks, public IP, and other supporting resources can have separate charges. A public IP is optional; many production designs use private connectivity or a load balancer instead of exposing a VM directly.
Exam rule: a VM needs compute, an OS disk, and network connectivity. Data disks and public IP addresses depend on the scenario.
Virtual Machine Scale Sets (VMSS)¶
A scale set creates and manages a group of VM instances with centralized scaling, availability, and lifecycle management.
| Orchestration mode | Exam-level distinction |
|---|---|
| Flexible | Recommended mode; can manage identical or mixed VM types and spread instances across zones or fault domains |
| Uniform | Uses a shared VM profile to create a large set of identical instances |
Use VMSS when the scenario mentions:
- many VM instances managed together;
- automatic scale out and scale in;
- load-balanced web or application tiers;
- updating many VMs with a consistent configuration.
Exam rule: a group of VMs that scales and is managed together → Virtual Machine Scale Sets. If the question specifically says identical instances, it is describing the Uniform pattern.
Availability sets¶
An availability set is a logical grouping that spreads VMs across fault domains and update domains within one region.
| Domain | Purpose |
|---|---|
| Fault domain | Separate groups of hardware that have independent power and networking failure boundaries |
| Update domain | Group VMs so planned maintenance reboots only part of the set at a time |
Availability sets protect against local hardware and planned maintenance events inside a region. They do not protect against an entire region outage.
Exam clue: maximize VM uptime within one region using classic IaaS patterns → availability set.
Availability sets versus availability zones¶
| Feature | Availability set | Availability zone |
|---|---|---|
| Scope | Within one region, across fault/update domains | Physically separated datacenter groups inside a region |
| Protects against | Correlated hardware, power, network, and planned-maintenance events | Zone-level datacenter failure |
| Common pairing | Individual VMs or smaller deployments | Zone-redundant or zonal VM deployments |
For new designs, Microsoft often recommends zones when the region and service support them. Availability sets remain a valid exam answer for spreading VMs within a region.
Azure Virtual Desktop (AVD)¶
Azure Virtual Desktop delivers Windows desktops and applications from Azure to users on many devices.
Use AVD when the scenario mentions:
- virtual desktops for remote workers;
- centralized desktop management;
- multi-session Windows experiences in Azure;
- replacing or supplementing traditional VDI.
Exam rule: virtual desktops hosted in Azure → Azure Virtual Desktop, not a generic VM or App Service.
Scenario¶
A web front end runs on identical VMs behind a load balancer. Traffic grows on weekdays and shrinks at night.
Best fit: Virtual Machine Scale Sets with autoscale rules.
Scenario¶
Two legacy application servers must stay available if one server rack fails, but the app cannot yet use zones.
Best fit: place the VMs in an availability set.
Exam clues¶
- VM instances managed together with autoscale and load balancing → VMSS.
- Spread VMs across fault and update domains in one region → availability set.
- Remote Windows desktops from the cloud → Azure Virtual Desktop.
- VM must communicate inside a VNet → NIC attached to a subnet, normally with a private IP.
- Persistent operating system storage → managed OS disk.