Azure High Availability

Azure High Availability

High Availability in Microsoft Azure is the ability of a cloud service to remain available by making sure that there is no single point of failure and to ensure a specified level of performance. Microsoft Azure handles the requirements of High Availability really well since the entire fabric of the system is designed to provide very high levels of resiliency and redundancy. The infrastructure of Azure is made up of Availability Zones, Regions, and Geographies that limit the range of potential failures. This unique solution was created via software and networking initiatives to make sure failures of any single data center are protected against.

azure fundamentals high availability

Microsoft Availability Zones

Microsoft Availability ZonesThe technical diagram here describes the concept of high availability in Microsoft Azure. The idea behind this is that if you have a server that runs a mission-critical web application, it makes sense to have redundant versions of that server so that if anything happens to single server traffic would be routed to other working servers allowing the service to remain available. So having multiple servers is a great concept, but what is even better is having multiple servers in multiple data centers. Things can and do go wrong even at the data center level. What happens if the main network feed to a data center goes down? You’re in trouble.

3 Servers In 3 Data Centers

Availability Zones in Microsoft Azure are essentially Data Centers. Running the business workload across multiple Availability Zones ensures that if 1 or even 2 Availability Zones go down, the cloud services the business relies on will remain available. It is a common practice to run 3 servers across 3 Data Centers or Availability zones to ensure the highest availability.

Azure Load Balancer

microsoft azure load balancerWith multiple servers in multiple availability zones, the question then becomes how do you distribute and manage the traffic to all three servers and data centers? This is where the concept of a load balancer comes into play. We can see the graphical representation of an azure load balancer just to the left here. A load balancer allows you to evenly distribute traffic to multiple servers in one or more Availability Zones. If an Availability Zone becomes unavailable or unhealthy, the load balancer will then route traffic to only available data centers with available servers. Azure Load Balancer can only route traffic inside an Azure region, as it only works with Virtual Machines in the same region. For global Multiregional environments, one can use Azure Traffic Manager. Note that Azure Load Balancer works with traffic at Layer 4, while Azure App Gateway works with Layer 7 traffic, and specifically with HTTP(S) traffic.

  • Azure App Gateway is good for load balancing HTTPS traffic to a web farm in a single Azure region
  • Azure Traffic Manager is good for non-HTTP(S) mobile apps available around the world and to ensure users can access the application from any country in the world with minimal latency.
  • Reliability In Microsoft Azure

    In order to build solutions that can continue to function despite any failure is a key aspect of ensuring high availability in a given cloud service. With this model, being able to quickly and efficiently survive a failure is a shared responsibility between the cloud service provider and the business. It is a three-level hierarchy that consists of a resilient foundation, resilient services, and resilient applications. The foundation is the azure capabilities that are built into the platform itself. Availability Zones come to mind here. The services are Azure capabilities that you enable as needed. The application layer consists of the application and workload architecture. Features like Zone Redundant Storage facilitate replicating data across zones, a key aspect of high availability.

    Learn More About Azure High Availability