auremar - Fotolia

Navigate through AWS regions and availability zones

Looking for basics on AWS regions and availability zones -- and how that translates into uptime and availability for cloud workloads? We've got answers.

As the world's largest public cloud provider, AWS needs a global focus. It's not enough for AWS to take care of its U.S. customers; it has to make itself a viable infrastructure as a service option for a diverse global market.

AWS regions and availability zones (AZs) as well as data center locations are a big selling point, allowing businesses to provision resources regionally and replicate data and workloads across locations. AWS regions exist across separate geographic areas in five continents and consist of multiple isolated AZs. Each AZ has at least one data center, though exact locations of data centers are not disclosed.

All this amounts to a vast array of resources designed to give AWS customers the highest possible availability. But how do customers put that massive infrastructure to use? We answer several questions about AWS data centers to shed some light on what they mean for your enterprise.

Does reliability vary between AWS regions and availability zones?

Different regions have different reputations. Last year, Amazon EC2 instances had an overall uptime of 99.985%, according to CloudHarmony. Some regions even boasted perfect, 100% uptime records. So, when a region has a poor reputation, it's relative.

Still, outages and disruption trends are noteworthy. In the last five years, the US-East-1 region -- the oldest and largest AWS region -- has seen several service issues, prompting some businesses to move workloads to the US-West-1 or US-West-2 region. Some believed the region had grown too unwieldy, its size pushing it beyond Amazon's realm of control.

CloudHarmony monitoring, however, shows that the US-East-1 region is not the cloud provider's least reliable, proving comparable or better than other state-side regions. The least-reliable AWS region proved to be SA-East-1, in Sao Paulo, Brazil, with eight outages in the past year and more than 35 minutes of downtime. Service disruptions are fairly minimal in any region, but harrowing, nonetheless, for affected businesses. Evaluate a region's track record -- not necessarily its reputation -- before spinning up workloads.

What role do AZs play in disaster recovery?

Disaster recovery is all about planning; often businesses most affected by cloud disruptions are the most unprepared. There are several tools and infrastructure available to minimalize downtime within AWS regions and availability zones.

AWS regions and availability zones as well as data center locations are a big selling point, allowing businesses to provision resources regionally and replicate data and workloads across locations.

Amazon Route 53 allows developers to redirect IP traffic from a failing or overworked region into an operational one. Developers can use Elastic Load Balancer to automatically distribute traffic to applications, which can run with Auto Scaling, ensuring a particular workload always has the necessary computing power to meet demand.

Route 53 Traffic Flow introduced new features, enabling more complex setups using traffic policies. Customers can now build a routing system that directs traffic to different cloud or on-premises endpoints based on geographic location, latency and availability. Third parties previously offered similar technology, but Route 53 Traffic Flow offers cost savings over those options.

Where should AWS build data centers?

AWS is expanding its data centers each year, but it hasn't yet stretched as far as it could. It can be argued, that a competitor has an edge on AWS when it comes to geographic data center diversity. More regions means less delay for a connection, putting a premium on proximity, as delays in the milliseconds can add up.

Microsoft Azure claims to have more data centers than AWS in 2015, and it plans to expand further into Europe in 2016. Though Azure's regions are not an apples-to-apples comparison to those of AWS, which claims to have a greater global reach. Still, data privacy issues have left at least one AWS customer clamoring for a geographic expansion into Canada -- a wish-list item that has since been answered along with further expansion into China, where AWS operates a region (Beijing) in limited preview. Russia is another location on customer wish lists for expansion, but bureaucratic red tape makes expansion difficult at the moment. AWS does not own any data centers in Africa, but that too could change in the coming years.

Where does AWS plan to expand?

AWS recently laid out data center expansion plans for 2016. It should please some customers to learn AWS will open data centers in Montreal, China, India, Ohio and the United Kingdom in 2016.

The data center expansion is of particular interest as governments around the world consider regulations on data and encryption, leaving a somewhat muddy playing field in the future. The end of the Safe Harbor agreement was the first major change to global data sovereignty, but more could be on the horizon. Customers will still be able to choose where their data resides, allowing them to stay a step ahead of changing regulations.

No matter what happens on the legislative end of the equation, AWS offers plenty of tools to protect and encrypt data. Ultimately, it's in the best interests of the cloud provider and AWS to keep workloads and data safe from intrusion.

Next Steps

Spin up regional AWS resources

Improve availability with Amazon Route 53

Set up disaster recovery to avoid performance loss

Dig Deeper on AWS compliance, governance, privacy and regulations