Google's Compute Engine, announced Thursday morning at the company's annual
developers' conference, gives Google a new way to monetize its huge investment in data centers and networking, offers new options for companies looking to run their businesses in the cloud.
developers' conference, gives Google a new way to monetize its huge investment in data centers and networking, offers new options for companies looking to run their businesses in the cloud.
The service is still in “limited preview” and details were initially sketchy, but the
Google Compute Engine offers companies access to the Linux Virtual Machines and network backbone connections that Google uses to power its own business. Pricing
can be seen here, but the company claims “50% more compute for your money than with other leading cloud providers.”
Google Compute Engine offers companies access to the Linux Virtual Machines and network backbone connections that Google uses to power its own business. Pricing
can be seen here, but the company claims “50% more compute for your money than with other leading cloud providers.”
By entering the IaaS market, Google is following Microsoft, which also recently added Linux IaaS services to its Windows Azure PaaS (platform as a service).
The service will offer the ability to run Linux virtual machines (VMs) based on either the Ubuntu 12.04 or the CentOS 6.2 Linux distributions. The distributions run over the KVM (Kernel-based Virtual Machine) hypervisor. The service also comes with storage and networking capabilities. To offer additional management functionality with the service, Google has also partnered with a number of software vendors, including Puppet Labs, RightScale and Opscode.
The company will initially market the service to those who will need 100 or more VMs. The ideal workload would be more compute-intensive rather than I/O-intensive. Administrators can add more cores to an application while it is still running.
VMs will be available with either one, two, four or eight cores, with each core allotted 3.75GB of memory. Running a standard single core will cost US$0.145 per hour, with additional costs for networking and storage.
The company has achieved pretty good adoption of its Google App Engine PaaS offering thus far. Launched in 2008, the service currently hosts 1 million active applications. Collectively the services get about 7.5 billion Web requests per day. The Cloud SQL executes 50 million queries per day and the datastore conducts 2 trillion operations per month.
Who Is Google Computer Engine For?
According to a Google spokesperson, during the limited preview, customers interested in Compute Engine can let Google know of their interest, and it will choose the users likely to get the most benefit.
So who would fit the profile? The spokesperson said the rule of thumb is the need for 100 cores or more.
The spokesperson added that the limited preview is likely to last months, not years. The technology itself is ready, apparently, but the company wants to make sure it can offer a comparable sales and customer support experience.
It’s still very early, of course, and the Google Compute Engine is far from offering the complete suite of cloud computing products already available from Amazon and smaller competitors like Rackspace. But there’s no question that Google has the infrastructure in place to make Compute Engine a credible alternative.
The big question is how much Google cares about this market, and how much attention and resources it will devote to a brand-new set of very demanding customers who couldn’t be more different than the millions of consumers using Gmail.
The Specs:
According to Craig McLuckie, product manager of Google Compute Engine in
Compute. Launch Linux VMs on-demand. 1, 2, 4 and 8 virtual core VMs are available with 3.75GB RAM per virtual core. Storage. Store data on local disk, on our new persistent block device, or on our Internet-scale object store, Google Cloud Storage. Network. Connect your VMs together using our high-performance network technology to form powerful compute clusters and manage connectivity to the Internet with configurable firewalls. Tooling. Configure and control your VMs via a scriptable command line tool or web UI. Or you can create your own dynamic management system using our API.
Initial partners include RightScale, Puppet Labs, OpsCode, Numerate, Cliqr and MapR.
0 comments:
Post a Comment