Google will announce a cloud storage service within the next few weeks that competes with Amazon S3. Pundits say the service is the fabled “GDrive,” in development for the past few years.
According to internetnews.com, Mike Repass, Product Manager at Google, let the news slip while on stage at the Interop Enterprise Cloud summit talking about Google’s AppEngine and related cloud initiatives.
“It’s a feature debt issue,” Repass told the audience. “We started out Google AppEngine as an abstract virtualization service. Our static content solution is something we’re shipping soon.”
In March of last year, Google launched Google App Engine. GDrive has been in development for years. It is one of Google’s most anticipated products. In 2006, some information about GDrive leaked after a Google manager inadvertently posted a presentattion with notes about the product:
GDrive seems to be a storage service capable of, according to the speaker’s notes, storing 100 percent of a user’s data on Google’s servers and turning personal computers into temporary data caches.
Amazon S3 and Google Big Table
Google does not have a product like Amazon’s S3, though there are other similarities in their web services.
First off, Amazon S3 stores large, binary objects. They do not index the block of data itself. Storage and retrieval is always at the block level. S3 objects can be public (every object has a unique URL) or private (the default unless an ACL entry is used to change it). It is unknown if Google will index blocks of binary objects as part of their new cloud storage service.
Google’s Big Table is most similar to Amazon’s SimpleDB. Both index data fro fast retrieval with automatic scaling to handle large amounts of data and a high degree of concurrency. SimpleDB automatically indexes all of the keys and attribute values.
Reuven Cohen of ElasticVapor says that GDrive should not be considered as an Amazon S3 killer as some have tried to do. He sees an emerging cloud pool arising where Google and Amazon will co-exist:
Moving forward I believe we are about about to see the emergence of a “Cloud RAID” model where you are able connect multiple remote cloud storage services (S3, Nirvanix, CloudFS, GDrive) in a broader data cloud.
What the effect of the GDrive is anyone’s guess except accelerating the computer’s future as simple, lightweight instrument that is more virtual than anything else.