ktsdesign - Fotolia

Google to open one new datacentre region a month in 2017 to boost multi-cloud appeal

Google uses cloud user conference to outline how its commitment to innovation, low prices and new datacentre regions is helping it stake a claim in its competitors’ customer bases

Google claims to be on course to open one new datacentre region each month during 2017, as the internet giant moves to entice more Amazon cloud users to start using its services too.

Speaking at the Google Cloud Platform (GCP) Next conference in London, Brian Stevens, vice-president of products at GCP, said a lot of the early cloud adopters flocked to Amazon Web Services (AWS), but are starting shift some of their workloads to Google’s platform now as well.

“A lot of the first adopters of cloud ran on AWS, yet many of them still come to GCP because of some of the things and innovation we have that AWS doesn’t,” he said.

As examples, Stevens cited the customisation and speed capabilities of Google’s virtual machines, as well as its commitment to building out its global datacentre footprint and the investments in submarine cables the company is making to underpin its cloud.

From a datacentre perspective, Stevens said the company is committed to “standing up” one new region a month in 2017, which it is on course to do with new facilities already planned for Germany, Finland, Singapore, Brazil, North Virginia and London.

“That is really fast and no one has ever done that before,” he said.

The need for speed in this area is being driven by user demand for low latency connections to their applications and workloads, and the cost involved is a secondary matter, he said.

“As we expand our capacity inside GCP for users, it makes the most sense to put those datacentres near where our customers are because it has a huge impact on latency,” he said.

“Because of this, you want to be able to expand greater into all the global metros. It’s not a capital expenditure issue, because the ROI [return on investment] is already there because of the global demand.”

Read more about the Google cloud

On a pricing front, Google also claims to be outperforming its competition in the public cloud, while its compute and storage services are becoming 20% more efficient to use on a year-by-year basis.

“We charge based on what you use, so you pay less when you build exactly what you need and that turns into price advantage, so it’s not just our retail prices per hour are lower,” said Miles Ward, global head of systems at GCP.

Google’s decision to charge users of its cloud platform on a per minute – rather than a per hour – basis is another reason why Ward believes users are shifting from other providers to Google.

“What if you could only put £100 pound on your Oyster Card and every time you get on the tube, it costs you two or three pounds to use the tube, but it rounds up and [charges] the entire 100 pounds? Is that price efficient? Does that sound like a good plan if you use the tube a lot?” he asked.

“On the Google cloud, we don’t round up and you get the change back. You can use it again when you want to turn a machine on, whether for 17 minutes or two hours and 14 minutes.”

Multi-cloud enterprises

It is Google’s view that enterprises will run applications in hybrid cloud environments and use services sourced from a range of suppliers, including Microsoft and Amazon.

How big a pool of suppliers they will have to choose from in the long term is subject to debate, however, with Stevens naming a couple of big IT firms that have backtracked on their public cloud commitments in recent years.

“There has definitely been a lot of cloud consolidation over the past few years. Before I came to Google, HP and VMware was in the cloud business. Rackspace had already fundamentally shifted their model, but it’s still going to be a multi-cloud world,” he said.

“Moreover, enterprises are still building their own private clouds as well because, unlike the startups, they can’t do this shift [to the public cloud] overnight. Our motivation and ambition is to get rid of any of the unneeded burden for enterprises to actually use hybrid cloud.”

Read more on Infrastructure-as-a-Service (IaaS)