Microsoft Azure organizes its global infrastructure into regions, which are geographic areas containing one or more datacenters connected through a low-latency regional network. Within each region, Azure provides Availability Zones, which are physically separate datacenters with independent power, cooling, and networking. Using Availability Zones allows workloads to remain available even when an entire datacenter experiences an outage. Most regions contain three zones, and deploying resources across zones raises the service-level agreement for a single virtual machine from 99.9% to 99.99%. Azure also pairs regions within the same geography to support platform-level disaster recovery, and services such as geo-redundant storage replicate asynchronously between paired regions hundreds of miles apart.
At the core of Azure's management model sits Azure Resource Manager (ARM), which provides a consistent layer for creating, updating, and deleting resources through the portal, command-line tools, PowerShell, or declarative templates. Resources are grouped into Resource Groups, which are logical containers that share a common lifecycle, permissions, and policies. ARM templates and the simplified Bicep language let teams define infrastructure as code using JSON or a domain-specific syntax. For larger organizations, Management Groups organize multiple subscriptions into a hierarchy up to six levels deep, enabling inheritance of policy and role assignments across the enterprise.
Understanding the shared-responsibility and service models is essential. Azure offers Infrastructure as a Service (IaaS) such as virtual machines that customers manage, Platform as a Service (PaaS) such as Azure SQL Database where the platform handles the underlying infrastructure, and Software as a Service (SaaS) offerings. Cloud computing shifts spending from capital expenditure on physical hardware to operational expenditure based on consumption, giving organizations scalable capacity without the maintenance burden of owning infrastructure.