Availability Options in Azure | Understanding Fault Domain and Update Domain

In this article, you will be learning about availability Options in Azure, which is essential to provide a seamless experience to the end-users. If you are a beginner in Azure, it is important to get the look and feel of the Azure Portal and understand what Resource Groups to have a better understanding of Availability options are.

Having high availability for critical applications is crucial because of the simple reason that you cannot afford to have downtime as it could have serious consequences. Now critical applications can be anything from a banking application to a Medical facility application. Imagine waiting in the long queues to withdraw money from the ATM or in the hospital to fill a form. It’s inefficient on the part of the support group and, at the same time, stressful for the consumers/end users.

Here, we will look at the available options available and the scenarios in which each of the options suits better in the real world. We will also look at how exactly the options work in the backend or how Microsoft handles it from the backend. Trust me; it is an interesting concept 😊

All of you must know how a data center looks like; for those who do not know or are new to the cloud, I have added an image below to show you how an actual Microsoft data center looks like from the inside. To those who are new to the cloud, this is where all your resources and data are stored. “This is Cloud !”.

Understanding Fault Domain and Update Domain

Before we move on, we need to understand few terms in Azure. As you can see from the image above of the data center is divided into racks, and each rack consists of physical servers.

Now those physical racks are called Fault Domains in Azure. Each rack has an independent network supply and power supply connections. If a power supply is interrupted to any of the racks, all the resources in that particular rack go down with it, or we could say the entire Fault Domain is down.

Microsoft calls this an Unplanned maintenance activity. Obviously, this was unexpected, so was the name. Availability Options in Azure portal.

In the above diagram, I took an example of two Fault Domains (racks) to help you understand what Update Domains (UD) are. During planned maintenance activity like patching or any upgrade activity, Fault Domains are divided into Update Domains (UD) like all the VM’s in the first row of each Fault domain is Update Domain1 (UD1), and the second row is called Update Domain2 (UD2) and so on.

Fault Domain: Each rack in the Datacenter is called Fault Domain in Azure.

Update Domain: Subsets of Fault Domain. Fault Domains are divided into Update domains to avoid downtime during planned maintenance activity.

Planned Maintenance activity: Any resource upgrade or monthly patching etc.

Unplanned Maintenance activity: Sudden interruption of Network or Power supply.

NOTE! The limit for no. of Fault Domains you can have is 3, and for Update Domains, it is 20.

What are the availability options in Azure ?

You will be asked to select one of the available options as an admin while creating a Virtual Machine. Below are some of the available options in Azure. Availability Options in Azure in the portal.

• No Infrastructure Redundancy required : Single VM in the selected region with no copies.
• Availability Set : Copies of VM with in same Datacenter and different Fault Domains.
• Availability Zone : Copies of VM across different datacenters ensuring high availability.

Scenarios of selecting Availability Options

By referring to the above image, consider if your application is running on VM1 and VM2 and there is a power supply interruption to the Fault Domain1. Will you be able to run your application? No ! because both the VM’s are in the same Fault Domain. This is where Availability Set comes into the Picture.

When you select Availability Set as the available option, you will be asked to select the no of Fault Domains and Update Domains, thereby creating a replica of your VM in other Fault domains, ensuring you have high availability even during unplanned maintenance activity even when one of the Fault Domain goes down. Microsoft offers 99.95% VM uptime SLA for Availability Set.

NOTE! The cost for Data Transfer in the Availability set is Zero since the resources are in the same Datacenter.

Imagine it is a very critical application, and you can’t even afford a data center going down. In such a scenario, you can use Availability Zone. Microsoft offers the industry’s best 99.99% VM uptime SLA for Availability Zone. You can select up to 3 availability zones; in Availability Zone, your resources are replicated across different zones in the same region, thereby ensuring high availability and protection from Datacenter failures.

Summary

NOTE! As an Admin, you can only select no. of Fault Domains and Update Domains. You cannot select which Domain you would like to have in the Datacenter. It is taken care of by Microsoft, and they try to ensure your resources are not in the same Update Domain.

Conclusion

• Availability Set is considered over Availability Zone when Cost is factor as data transfer cost is Nil.
• Availability Zone is preferred over Availability Set when high availability is considered as priority because Availability set is vulnerable during calamities as all the VM’s are in single Datacenter.

Resources

• Understanding Availability Sets
• Understanding Availability Zones