Amazon's Redshift cloudy data warehousing service is now available for general consumption after a trial among blessed customers.
The pay-as-you-go technology, which scales from around 200GB into the petabyte range, represents a direct threat to the warehousing divisions of IBM, Oracle, Teradata, and EMC (Greenplum).
Redshift does "10 times the performance at 1/10th the cost of the on-premises data warehouses that are commonly used today," the company wrote in a blog post announcing the service's availability on Friday.
Its architecture is based around columnar data storage, advanced compression, and fast disk and network I/O.
This, combined with its roots in the open source PostgreSQL database, has prompted The Reg to speculate that Amazon has managed to parallelize the PostgreSQL technology.
The warehouse dovetails into other major AWS components, like the S3 storage cloud and DynamoDB. It can suck in data from Amazon RDS, Elastic MapReduce and data sources in EC2 instances via the the recently-launched AWS Data Pipeline.
Pricing for the service can go to as low as $999 per year per terabyte, if using three-year reserved instances. As is typical, the service will initially be run out of Amazon's main data center – US East in Northern Virginia – with expansion occurring around the world "in the coming months."
The service was announced back in November at Amazon's inaugural cloud conference, AWS re:invent.
Redshift is designed to appeal to companies that either don't have the cash to pay for pricey hardware and software from typical warehousing stalwarts like IBM, Oracle, Teradata and Greenplum, or that want to trim their staffing and equipment budget.
This goes to the heart of the AWS cloud proposition: take on a bit of risk in terms of availability, and in exchange get rid of expensive capital items and save on employee headcount, especially DBAs. A good thing for companies, certainly, but a somewhat frightening prospect for IT workers. ®