For a long time, since the beginning of Microsoft Azure, the most granular control you could have for choosing what servers / hardware to host a workload on has been the Azure Region. Each Azure Region is made up of at least 2 or 3 datacenters, but you didn’t have control over which data center was used or how replication was really controlled amongst the datacenters in a particular region. However, now Microsoft Azure has the ability to use Availability Zones to increase the resiliency and high availability of services hosted in a single Azure Region. These Availability Zones now offer a more granular control over hosting workloads in Azure beyond just choosing the less granular Azure Region to host in.
Microsoft Azure is available in 36 regions globally, with 6 additional regions in the works to be added soon. (At least at the time of writing this article; Microsoft keeps investing and building new regions. It’s growing!) These Azure Regions are basically geographic groupings of 2, 3, or more datacenters. When provisioning any cloud resource within Microsoft Azure it must be placed in, and reside within, an Azure Region.
You can view a “live” map of the Microsoft Azure Regions using Bing Maps, here: http://map.buildazure.com
When deploying out your cloud resources and workloads, it’s generally best to choose the Azure Region that’s closest to your employees, users, or customers. This will keep the latency down when communicating across the Internet. It also helps to utilize multi-region deployments when adding global / geographic redundancy and failover to better implement increased resiliency and high availability. It’s best not to solely rely on a single Azure Region, and instead be prepared to gracefully handle some kind of Regional outage that may occur beyond your control.
Another useful thing to know about Azure Regions is the concept of Azure Region Pairs. Using Azure Region Pairs for multi-region deployments will ensure that in the event that the Primary Region experiences an outage, then you can failover to the Secondary Region. Additionally, by using the Region Pair for the Secondary, then you will also ensure that if both Regions were to experience an outage then Microsoft will prioritize one of the Regions in the pair to restore service to first; thus improving the quickness in service restoration.
You can read the following article for more information and an expanded explanation on Azure Regions Pairs: Azure Region Pairs Explained.
Azure Availability Zones
For a long time the most granular control you had of where to host resources within Microsoft Azure was at the Azure Region level. With Azure Availability Zones you can choose which Zone within that Azure Region to host a resource. This enables a more granular choice of where and how to host resources within an Azure Region.
Each Zone within an Azure Region is essentially a separate datacenter. Each Zone has independent power source, networking, cooling, etc. Each Azure Region with Availability Zones made available will have at minimum 3 separate Zones. This is to ensure maximum resilience and high availability can be enabled within the Azure Region.
The Microsoft Azure Fabric is the underlying technology that manages the datacenter automation of the Microsoft Azure Platform. This “Fabric” has managed distributing resources for resiliency and high availability for running the Azure Platform itself across all the Zones of each Azure Region since Microsoft Azure as originally launched. It’s just now that Microsoft is making these Zones available to customers of the platform to help them better host and manage better resiliency and high availability for their individual workloads.
When configuring and provisioning Azure Resources, such as Virtual Machines or Managed Disks, you first choose the Azure Region to host it in, then you will have the option to pick the specific Availability Zone to use for that resource as well.
At the time of writing this, the Azure Availability Zones features is newly in Public Preview. To signup for the Azure Availability Zones Preview and enable the feature on your Azure Subscription, go here: http://aka.ms/azenroll
There are a few restrictions on Azure Availability Zones at this time, during the initial Preview availability.
Here’s the list of Azure Regions that are supported during this initial Preview:
- East US 2
- West Europe
There are additional restrictions beyond just the Azure Regions. The service is also only available for a few services and VM SKUs as well.
Here’s the list of Azure services that currently support Availability Zones during this initial Preview:
- Windows Virtual Machines
- Linux Virtual Machines
- Zonal Virtual Machine Scale Sets
- Managed Disks
Here’s the initial supported Virtual Machine SKU’s / Pricing tiers as well:
Additionally, as with other Preview services within Microsoft Azure, the Azure Availability Zones features does NOT include an SLA at this time. Once the service is made Generally Available (GA) then it will get a full 99.99% SLA guarantee.