What’s behind Amazon, Microsoft and Google’s aggressive cloud expansions

An arms race is on in the IaaS cloud as leading vendors build out data center capacity around the world

What’s behind Amazon, Microsoft and Google’s aggressive cloud expansions

It wasn’t long ago that the big spectator sport in IaaS cloud computing was to watch a leading provider such as Microsoft or Amazon Web Services announce price cuts and then ready for its rivals to follow suit.

The new game in town plays out in a similar way, except now the vendors are matching or one-upping each other with new data centers and cloud computing regions.

+MORE AT NETWORK WORLD: Winners and losers from the AWS-VMware deal | Mapping the cloud: Where does the public cloud actually live? +

Throughout the month of October, AWS, Microsoft Azure and Google Cloud Platform have all announced plans to build out new regions for their Infrastructure-as-a-Service cloud operations. AWS brought a new region online in Ohio and announced plans to open one in France; Microsoft also announced plans for two new regions in France and new government regions in Arizona and Texas. Google aired big plans to add, on average, a new region per month to its cloud in 2017.

The regions will add to an already impressive roster around the globe for each vendor. Amazon Web Services boasts 14 regions that include a total of 38 Availability Zones, and AWS has plans for four more regions. (Each AWS region is made up of at least two Availability Zones (AZ) for failover protection; each AZ has at least one data center). Microsoft doesn't use an AZ approach but instead has 30 regions for its cloud, with eight more planned. Google has five regions with announced plans for nine more.

Check out our interactive map to get an overview of where these cloud providers’ regions are located. (Click on the interactive map's pins and circles to view information about each region. Zoom in and out for more detail. Note that placement of pins are meant to denote general geographic area of a region not exact location of data centers.)

Keeping up with demand

Experts who track the IaaS market say the slew of regions being added to these already massive cloud platforms is driven by a handful of factors.

First of all, there’s a seemingly insatiable demand for services as users get more comfortable using the public cloud.

AWS says it has more than 1 million active customers using its cloud platform every month; the company is on an $11 billion annual revenue run rate and it’s growing 58% year over year. Microsoft this summer announced that its Azure IaaS cloud usage more than doubled year over year, and IDC estimates that IaaS cloud revenue will more than triple from $12.6 billion in 2015 to $43.6 billion in 2020. Vendors are rapidly onboarding data center capacity to keep up with this demand.

“More and more workloads are moving to the cloud, and increasingly they’re analytical, machine learning, IoT and cognitive,” says IDC Research’s Rick Villars, vice president of data center and cloud. “Latency is critical for these applications.”

Cloud vendors want to be able to service customers no matter where they are, hence the international expansion.

“Big Fortune 500 companies need that global reach not only for scale, but for data sovereignty,” says Jared Wray, former CTO of CenturyLink’s cloud computing platform and now an entrepreneur.

Data sovereignty

The enactment of increasingly stringent data sovereignty laws hasn’t been lost on cloud vendors. Last year, the European Union’s Court of Justice struck down the Safe Harbor agreement, which allowed for trans-Atlantic transfer of data, and that decision served as a catalyst for other countries to implement their own more restrictive regulations. While the EU Privacy Shield is the Safe Harbor’s replacement, the governments of Germany and France are leading the international community in considering even more strict rules related to its citizens’ data remaining in the borders of their countries. Amazon and Microsoft each already have regions in Frankfurt, Germany and both announced new regions in France within days of each other this month.

“There’s an anticipation, certainly in Europe and Asia, of more intensity and focus on data sovereignty,” Villars says.

As data sovereignty laws only become more constricting, vendors will have to build up regions in those countries if they want to service clients there.

Latency matters

Vendors’ expansion efforts also enable them to shave data latency by putting data centers closer to customers.

For example, AWS has two major region hubs on either coast of the U.S. and as of mid-October it built a new region in the center of the country in Ohio. Microsoft and Google each already have regional locations in Iowa in addition to their bi-coastal setup. Locating data centers in the middle of the country allows each of the providers to off-load capacity from their coastal regions.


A recent survey by 451 Research notes that 41% of all enterprise workloads are already running in some type of public or private cloud, and share is expected to increase to 60% by 2018. As mega-scale IaaS providers build data centers around the world, Villars says IDC research shows enterprises are building fewer data centers themselves and outsourcing workloads to cloud, managed service and co-location vendors. Data center construction by enterprises has fallen by about 10% each year since 2014, with no sign of that trend abating anytime soon, Villars says.

Copyright © 2016 IDG Communications, Inc.

The 10 most powerful companies in enterprise networking 2022