Virtual Machines

Infrastructure as a service (IaaS)
Virtual machines, or VMs, are software emulations of physical computers.
They include a virtual processor, memory, storage, and networking resources.
They host an operating system (OS), and you’re able to install and run software like a physical computer.
You can connect to VM & control it using a remote desktop client.
Good choice when you need:
- Total control over the operating system (OS)
- The ability to run custom software
- To use custom hosting configurations
Azure takes care of the physical hardware
- You take care of configuring, updating, and maintaining the software that runs on the VM.
An image is a template used to create a VM.
- Includes an OS and often other software, like development tools or web hosting environments.

Examples of use-cases

During test + dev as it’s easy to create different OS & application configurations. Easy to delete when not needed.
Minor tasks. E.g. application handles fluctations in demand and shut down VMs when you don’t need them & quickly start them to meet a suddenly increased demand.
Extending your datacenter to the cloud, e.g. running SharePoint.
During disaster recovery. E.g. if primary datacenter fails, create VMs running on Azure to run your critical applications and then shut them down when the primary datacenter becomes operational again.
Lift and shift: Moving from physical datacenter to cloud. You can take image of the server & run within a VM with little to no changes.
Learn more: Typical scenarios for running Azure VMs

📝 99.99% uptime guarantee for all Virtual Machines that have two or more instances deployed across two or more Availability Zones.

Groups of VMs and underlying physical hardware that can be rebooted at the same time.

When the underlying Azure fabric that hosts VMs is updated by Microsoft.
Done to patch security vulnerabilities, improve performance, and add or update features.
Often no impact, sometimes requires reboot.
When the VM is part of an availability set, the Azure fabric updates are sequenced so not all of the associated VMs are rebooted at the same time.
- VMs are put into different update domains

Fault domain = rack of servers.
provides the physical separation of your workload across different power, cooling, and network hardware that support the physical servers in the data center server racks.
In the event the hardware that supports a server rack becomes unavailable, only that rack of servers is affected by the outage.

Involve a hardware failure in the data center e.g. a power outage or disk failure
VMs that are part of an availability set automatically switch to a working physical server so the VM continues to run.
The group of virtual machines that share common hardware are in the same fault domain.

Logical grouping of two or more VMs that help keep your application available during planned or unplanned maintenance.
With an availability set, you get:
- ❗ Up to three fault domains
  - each have a server rack with dedicated power and network resources.
- Five logical update domains
  - ❗ can be increased to a maximum of 20.
There’s no cost for an availability set.
- Only pay for the VMs within the availability set.
💡📝 Recommended for high availability.

Lets you create & manage a group of identical, load balanced VMs.
Allow you to centrally manage, configure, and update a large number of VMs to provide highly available applications.
The number of VM instances can automatically increase or decrease in response to demand or a defined schedule.
💡 Helps you build large-scale services for areas such as compute, big data, and container workloads.
Provides high availability through regional or multiple Availability Zones deployment options.

Large-scale job scheduling and compute management.
When running a job, batch:
1. Starts a pool of compute VMs for you
2. Installs applications and staging data
3. Runs jobs with as many tasks as you have
4. Identifies failures
5. Requeues work
6. Scales down the pool as work completes
💡 Good for cases where you need raw computing power or supercomputer level compute power.