Google has made its IaaS platform, Google Compute Engine, generally available to any developer or business. The announcement came at its sixth annual I/O conference where the company highlighted other Google cloud developments such as updates to App Engine and the introduction of Cloud Datastore, a fully managed, scalable database service.
Google Compute Engine (GCE) was launched by Google at its I/O conference in 2012 but it was made available only to selected enterprises for beta testing and just last month Google extended it to customers that have opted for its “Gold” support.
The platform enables enterprise users to run their large-scale computing workloads on Linux virtual machines hosted on Google's infrastructure. As the platform uses KVM as the hypervisor, it currently only supports guest images running Linux.
Google Compute Engine gives developers access to Google's computing infrastructure, Barak Regev, head of Google's cloud platform in EMEA told Computer Weekly last month.
At that time, Regev added that the general availability of GCE will be the “next milestone for Google enterprise cloud division”.
“One of our main goals in building Google Compute Engine is to enable a new generation of applications with direct access to the capabilities of Google’s vast computing infrastructure,” he said.
One European customer, Grandcentrix which develops mobile applications for big customers, has deployed a major mobile backend to Google Compute Engine.
“Performance is so much better compared with EC2,” said Ralf Rottman, Grandcentrix’s founder on Google+. When asked if he has tried using Microsoft’s Azure cloud services, Rottman said: “I briefly looked into it, but GCE's performance just was too convincing.”
Announced at I/O, New Compute Engine features by-the-minute pricing for short-lived workloads and includes micro and small instance types for low-intensity workloads. Building on its software-defined networking (SDN) strategy, Google also introduced support for software-defined routing capabilities to Compute Engine.
The sub-hour billing feature “charges for instances in one-minute increments with a 10-minute minimum, so you don’t pay for compute minutes that you don’t use,” said Amit Singh, president, Google Enterprise on the company blog.
The updates also include disk support up to 10 terabytes per volume, “which translates to 10X the industry standard,” according to Google’s enterprise blog.
Among other significant development to Google’s enterprise cloud strategy, it has completed ISO 27001:2005 international security certification for Compute Engine, Google App Engine, and Google Cloud Storage – a move that will make enterprises more confident to use Google’s public cloud services.
“What differentiates Google cloud services from other leading players is that our worldwide datacentres are managed by ourselves and we run all services on our own data infrastructure,” Regev said.
“This means when customers sign up to Google cloud, we are unlocking access to Google datacentres to them. Enterprises get to use the same infrastructure that Google uses for its own web services.”
According to Regev, multinational enterprises need to store data in different datacentres for compliance purposes and communication between two datacentres can be affected by high latency.
“But Google owns the data connections and the fibre between our datacentre locations so a user’s latency does not depend on an outside network or router,” he said.
Google is the third major cloud provider in the public cloud space along with AWS and Microsoft. It offers enterprise cloud services such as Google Cloud Storage, Google App Engine and now the infrastructure platform - Google Compute Engine.
At I/O conference, Singh also said that it is releasing the new Google+ API for businesses as a pilot for developers who want to build apps that handle tasks like auto-provisioning “Circles” and reading or writing posts.