Google Compute Engine opens up to all
11 months on from its unveiling, Google say theyre ready to tackle Amazon Web Services with their infrastructure project.
In order to distinguish itself from the undoubted kingpin of cloud, Google are offering sub-hour billing – something which rivla Amazon Web Services doesn’t. Rather than buying instances by the hour, Google Compute Engine users can obtain instances in one minute increments with a ten-minute minimum. It’s a bold pay-as-you-go pricing strategy but one that could entice users over with no wasted compute instances.
Shared-core instances have been introduced for lower-intensity workloads, while advanced routing to create gateways and VPN servers has been added to span Google’s plethora of cloud services. Compute Engine is also able to attach large persistent disks of up to 10 terabytes, which Google claim is up to “10x the industry standard”.The other big I/O cloud announcement was Google’s lifting the lid on Cloud Datastore, which is primed to challenge Amazon’s DynamoDB. The schemaless NoSQL database previously sat under Google App Engine as BigTable, but now is a standalone project in its own right. The highly available and scalable Cloud Datastore uses the columnar format and offers ACID transactions, SQL-like queries and indexes. Google’s increasingly fragmented approach, in breaking up larger cloud projects into smaller simplified services, appears to be working, with over 300,000 developers and 3 million applications using Google Cloud Platform services. Whether it can dislodge Amazon Web Services from its perch remains another matter. Splitting up services is certainly the method to put it onto the par with its rival at least. The challenge now is to differentiate them all from one another. Image courtesy of yukop