A momo shop moves to the cloud
Sita runs a momo shop in Pokhara with a busy ordering website. Her one server sits under the counter: when the power cuts, orders stop. Microsoft Azure rents her the same things — servers, disks, databases, networks — by the minute, in datacenters all over the world, and keeps them powered, cooled and patched.
Azure has hundreds of services, but they all sit inside one simple structure. Learn the structure first and every service after it is easier: where it runs (a region), who pays for it (a subscription), what it belongs with (a resource group), and the one front door every request walks through (Azure Resource Manager).
Regions and availability zones
A region is a set of datacenters in one area — centralindia in Pune, southindia in Chennai, eastus in Virginia. Pick the region closest to your users for low latency, and check that the services you need are offered there: new features roll out region by region. Microsoft groups regions into geographies for data-residency rules, and keeps separate sovereign clouds for some governments.
Many regions have availability zones: at least three physically separate locations, each with its own power, cooling and network. Put two copies of a VM in different zones behind a load balancer and a fire in one building doesn't take you offline. Most regions also have a paired region in the same geography (Central India pairs with South India); Microsoft links pairs with extra network capacity and avoids updating both at once, which makes the pair the natural place for a disaster-recovery copy.
In Azure's own vocabulary, region and location mean the same thing — templates say location: "centralindia".