Microsoft Azure is generally available in over 60+ regions around the world. Each Azure Region is home to a vast array of services hosted within multiple datacenters in the region. Separate regions is very apparent within Microsoft Azure, as you need to choose a specific Azure Region to host your services within. Although, something that’s not very apparent is the concept of Azure Region Pairs. Specific Azure regions are paired together for disaster recovery purposes. This article discusses what Azure Region Pairs are, and the benefits that come with having them be a part of the Microsoft Azure cloud platform.
What are Azure Region Pairs?
An Azure Region Pair is a relationship between 2 Azure Regions within the same geographic region for disaster recovery purposes. If one of the regions were to experience a disaster or failure, then the services in that region will automatically failover to that regions secondary region in the pair. For example, North Central US region’s pair is South Central US.

Almost all the Azure Regions are located within the same geographic region as at least 1 other Azure Region; it’s Region Pair. The only exception to this rule is the Brazil South region, which is the only Azure Region in Brazil, so it’s Region Pair is South Central US in a one-way pairing connection, as South Central US’s pair is North Central US. I know this sounds confusing, but this is the only Azure Region who’s pair is outside it’s geographic region.
The Azure Region Pairs are more than just a visual concept to think about with Azure Regions. The Azure Region pairs are connected directly together and offer multiple benefits when utilized together in the same distributed or redundant system. Before we get into these benefits of Azure Region Pairs, let’s first discuss the Azure Regions and Datacenters in further details.
Azure Regions in a Pair have direct connections which bring additional benefits to use them together.
The geographic disparity of Azure Regions is important. Global distribution of Azure Regions is strategic according to geography. There are many factors involved in their placement; from geo-political to internet latency for large population centers. However, Azure Region pairs have another big consideration that defines where they are placed.
Each Azure Region in a pair are always located greater than 300 miles apart when possible. This is to isolate each region in the pair from being affected by the same regional disasters that could take one of the regions down. These disasters could be earthquakes, hurricanes, tornados, fires, or some other natural or man-made disaster.
Each Azure Region in a pair are always located greater than 300 miles apart when possible.
Since each Azure Region in an Azure Region Pair are directly connected to each other and the are far enough apart to be isolated from regional disasters, it is recommended by Microsoft that when replicating data or interacting with services across regions that you use Region Pairs.
While you could use more geographically dispersed Azure Regions for replication, redundancy, and more, there are a number of benefits Azure Region Pairs provide. Next, let’s discuss the benefits of Azure Region Pairs.
What are the Azure Region pairs?
Before we get into the benefits of Azure Region Pairs, let’s cover what Azure Regions are paired. Here is a list of a few of the Microsoft Azure Region Pairs:
Geography | Paired regions |
---|---|
Asia | East Asia / Southeast Asia |
Australia | Australia East / Australia Southeast |
Australia | Australia Central / Australia Central 2 |
Brazil | Brazil South (1) / South Central US |
Canada | Canada Central / Canada East |
China | China North / China East |
China | China North 2 / China East 2 |
Europe | North Europe (Ireland) / West Europe (Netherlands) |
France | France Central / France South |
Germany | Germany Central / Germany Northeast |
India | Central India / South India |
India | West India / South India |
Japan | Japan East / Japan West |
Korea | Korea Central / Korea South |
North America | East US / West US |
North America | East US 2 / Central US |
North America | North Central US / South Central US |
North America | West US 2 / West Central US |
South Africa | South Africa North / South Africa West |
UK | UK West / UK South |
United Arab Emirates | UAE North / UAE Central |
US Department of Defense | US DoD East / US DoD Central |
US Government | US Gov Arizona / US Gov Texas |
US Government | US Gov Iowa / US Gov Virginia |
US Government | US Gov Virginia / US Gov Texas |
Her’s a couple notes about the region pairs listed in the table:
- West India is paired to South India in one direction only. South India’s region pair is Central India.
- Brazil South is paired (in one direction) with the South Central US region outside its geographic region. This is a unique region, as it’s the only one without a pair in the same geographic region.
- The US Government regions have similarly unusual paring. The US Gov Iowa’s secondary region is US Gov Virginia; US Gov Virginia’s secondary region is US Gov Texas; and US Gov Texas’ secondary region is US Gov Arizona
When setting up services like SQL Database Geo-Replication and other services, the Azure Portal will guide you by telling you the “recommended” Azure Region to use as a secondary. This will always be the Azure Region Pair, so you don’t need to memorize the entire list Azure Region Pairs.
What are the benefits to Azure Region Pairs?
Some of the benefits to Azure Region Pairs were already discussed. There are a number of benefits to utilizing Azure Region Pairs that add to the basic benefit of geographic isolation. Let’s dig in to a few more of the benefits.
How is Data Residency handled?
The location of data stored within a geography can be very important. There are geo-political, tax, law enforcement jurisdiction, compliance, or other reason that an enterprise may require all their data to be located within the bounds of a certain country. This is the main reason why both regions in Azure Region Pairs are located within the same geographic region; most often within the same country.
How does Microsoft Azure manage System Update Isolation when pushing out updates to the Azure Regions?
Microsoft maintains a huge amount of software and infrastructure that runs and automates the Azure datacenters. When rolling out updates to this “backbone” system, Microsoft will update a single region in the pair first. Then they will validate the update before moving on to the next region in the pair. This sequential updating of the Azure Regions in a pair ensures minimal downtime in the event of bugs or logical failures caused by an update. By using Azure Region Pairs for replication and redundancy will ensure your applications and services are not adversely affected by a rare bad update event.
What about the Region Recovery Order in the event of a disaster?
In the event of a massive Azure outage, each Azure Region Pair has a single region that is prioritized over the other for recovery. Systems deployed across multiple Azure Regions within a pair are guaranteed to have an Azure Region that will be recovered with high priority.
What about the Platform-provided Replication between regions?
There are some services within Microsoft Azure that provide automatic geo-redundant storage. These services will do this by utilizing the paired region automatically.
For example, when you provision an Azure Storage account and configure it for geo-replication then it will automatically use the region the Storage Account is provisioned in as the primary with replication to that regions secondary paired region without any additional configuration.
What about the Physical Isolation between Azure Regions?
Each Azure Region within a pair is physically / geographically located at least 300 miles apart where possible. This helps isolate the pair so that a regional disaster (earthquake, hurricane, fire, riot, power outage, etc.) affects only a single region within the pair.
Don’t believe that statement around ingress/egress bandwidth at no charge between region pairs is correct.
https://azure.microsoft.com/en-us/pricing/details/bandwidth/
Is data transfer between Azure services located in two regions charged?
Yes. Outbound data transfer is charged at the normal rate and inbound data transfer is free.
Hi Chris – there currently isn’t any guidance in place to that effect that I am aware of; the only current option to eliminate bandwidth costs between regions is with ExpressRoute deployed to a region within that geo;
https://azure.microsoft.com/en-us/pricing/details/expressroute/
“Once connected to an ExpressRoute location, users can connect to other regions in the same geo without the need for the premium add-on, and at no additional cost over existing plan charges. For example, in the US, customers can send or retrieve data to/from any Azure region in the US without the need to pay an additional fee on top of their existing plan charges. Similarly, once connected to an ExpressRoute location in Europe, customers can send or retrieve data to/from any Azure region in Europe without the need to pay an additional fee on top of their existing plan charges.”
If you are aware of any publicly accessible statements or advice to this effect, please let me know and I can re-check internally – please use andrew.allen [AT] microsoft.com.
Thanks Chris – keep up the good work on the site!
I’m working on getting further clarity on Egress and will be updating the article shortly to better reflect charges. Thanks!
Hi Chris, I am just wondering If one of my Databricks uses storage redundancy in different regions then Is it possible to change that? I don’t want a DR Databricks procedure but I need to understand if some resource created automatically in databricks is in another region and I don’t have the option to change that. Could not find a clear answer.
I’m not familiar if there’s a way to tell Azure Databricks to create managed resources in a specific region. You should try asking this question on StackOverflow.com or other forums. Sorry I couldn’t be of more help.
Alright, I updated the article to remove the Egress part about Azure Region Pairs since I couldn’t find references to back it up. It seems that the pricing of Egress between Azure Regions and Zones is a little confusing. I may write up an article to explain it. Thanks for the fact checking!
Pairs being > 300 miles apart seems very unlikely for the UK.
This statement: “Each Azure Region in a pair are always located greater than 300 miles apart” is not true. As Microsoft’s own page on paired regions states, “When possible, Azure prefers at least 300 miles of separation between datacenters in a regional pair, although this isn’t practical or possible in all geographies.”
Here’s a fairly prominent counterexample: the UK South region is in London, and is paired with the UK West region in Cardiff, and these are only 150 miles apart.
The UK is small enough that with a carefully chosen centre, a 300 mile radius is very nearly enough to encompass the whole of the mainland. (Put the centre on the coast somewhere near the Lake District. and you can get very nearly everything in aside from the various islands which require a rather larger net.) That said, you could in principle fit two data centres into the UK with a 300 mile separation, but the place is small enough that this constraint would mean that neither data centre would be particularly near the areas that are likely to use Azure most. In practice, distribution of population and (probably of more relevant to Azure data centre location) high-tech industry means that you’re going to want one data centre in or near London, and that’s where the UK South one is. To locate a second one 300 miles away would put it either in Scotland, or at the very North-Eastern-most corner of England (I’m ignoring the possibility of underwater data centres). Berwick upon Tweed isn’t a major tech centre, so that rules out the latter. There is considerable high-tech industry in Scotland, but evidently Microsoft has determined that it’s not yet time for an Azure data centre in that region.
Thanks for the clarification! It seems UK is an exception to this rule. I updated the wording of the article. 🙂
Is it possible to get a dedicated bandwidth between these paired regions ?
What are you trying to do that you want “dedicated bandwidth between” region pairs?