Google slashes App Engine NoSQL data storage prices by 25 per cent

Amazon doesn't care much about profits and both Google and Microsoft have monopolies that give them deep pockets. And so it is no surprise that the three companies will be engaged in a cloud price war that will very likely leave a lot of smaller cloud providers dead by the side of the road in the coming years.
Google launched the Cloud Datastore NoSQL datastore, which was broken free of its Google App Engine at the I/O event only a week ago. Cloud Datastore is based on the High Replication Datastore (HRD) that underpins App Engine, which is a columnar data store that spreads data across multiple Google data centers for high availability using the Paxos algorithm that offers "eventual consistency" of replicated data for queries and "strong consistency" for reads. The service is based on Google's BigTable technology, which is has an SQL-like interface called GQL.

On the same day that Google launched Cloud Datastore, Amazon Web Services cut the price of big reads on its somewhat similar DynamoDB flash-backed NoSQL datastore by 75 per cent, following a similar price cut back in March that chopped the cost of storing data in the service by 75 per cent to 25 cents per GB per month.

In a blog post, Peter Magnusson, engineering director at Google, said that the HRD service, which launched in 2011, now processes 4.5 trillion transactions per month with 99.95 per cent uptime.
He added that Google was "always evaluating opportunities to create more value" for customers and hence was chopping prices on storing and operating on data that goes into the App Engines HRD service and the free-standing Cloud Datastore.
Storing data on these two services now costs 18 cents per GB per month, down 25 per cent. It now costs 9 cents per 100,000 operations to do writes to these two services, down 10 per cent, and read operations in buckets of 100,000 cost 6 cents, down 14 per cent.