Microsoft Azure
Core Compute — Virtual Machines & Scale Sets
Run Windows/Linux VMs and scale them elastically with VM Scale Sets behind a load balancer.
Azure Virtual Machines provide on-demand Windows and Linux servers. You choose a VM series by workload — B-series (burstable), D-series (general purpose), E-series (memory), F-series (compute), N-series (GPU) — and a size within it, deployed into a Virtual Network. Availability is built using Availability Zones (separate datacentres in a region) and Availability Sets (fault/update domains within a datacentre).
Leasing identical machines from a depot that automatically adds or removes machines as the workload changes and instantly swaps any that fail — and gives a loyalty discount (hybrid benefit) if you already own the software licences.
Key Concepts
1
For elastic, fault-tolerant fleets, Virtual Machine Scale Sets (VMSS) manage an identical group of VMs that scale automatically on metrics or schedule, spread across zones, and sit behind an Azure Load Balancer or Application Gateway. Cost is optimised with Reserved Instances and Savings Plans for steady workloads, and Spot VMs for interruptible batch at deep discounts — the same baseline/burst pattern as other clouds, with Azure's hybrid-benefit licensing as a distinctive lever for Windows/SQL workloads.