Microsoft's cloud offering Windows Azure can now deploy and manage Hadoop clusters after the public preview of the HDInsight feature developed with partner HortonWorks officially went live this morning.
HDInsight, Microsoft's Apache-compatible Hadoop distribution, has been in testing mode since last year, the company's general manager for SQL Server product management, Eron Kelly, said in a blog post .
Kelly said HDInsight integrates with Microsoft business analytics tools such as Excel, PowerPivot and Power View.
Customers in the United States and select countries such Australia and NZ are able to access the public preview as of today.
Microsoft is offering a 50 per cent discount for HDInsight for Azure during the preview period. It is priced at US$0.72 (A$0.69) per hour or US$518.40 a month per computing node, and on a pay-as-you-go basis.
Six-monthly and yearly commitments bring down the cost to US$0.558 and US$0.54 per hour respectively, with further cost reductions available for customers who prepay for the sevice.
Up to 25 computing nodes can be specified for HDInsight. A storage account that resides in the East US region is required to create the clusters.
Hadoop is an open-source framework written in Java that is used to run large computing clusters with distributed applications to process large amounts of data. Created by former Yahoo employee Doug Cutting and University of Michigan computer science professor Mike Cafarella, Hadoop is named after Cutting son's toy elephant.
It is currently developed by the Apache Software Foundation and has become popular with many large organisations around the world such as Amazon, Google, IBM, Apple, Twitter and SAP.
Facebook claims to be the largest Hadoop user currently. In November last year it was running a cluster with 100 petabyte of data, growing by half a petabyte a day.