Microsoft plugs Hadoop into SQL Server and Azure Cloud
Microsoft says it plans to port the Hadoop MapReduce platform to both its Azure cloud and Windows Server. Microsoft does not have its own comparable project in that space and this has lead Microsoft's vice president Ted Kummert to make the announcement at the PASS Conference in Seattle, Washington.
Apache's Hadoop project originated at Yahoo as a way to address big data and search challenges. It has since become one of the leading platforms for taking on that class of problem. The system allows problems to be distributed (map) over a large number of actual or virtual systems and then gather the results (reduce) for presentation to the user. The ecosystem around Hadoop includes database connectivity, scheduling and management.
By the end of 2011, a Community Technology Preview (CTP) is to be available in Microsoft's Azure cloud platform, with a Hadoop CTP for Windows Server following next year. Kummert says the firm will work closely with the Hadoop community and return its own contributions to the software back to the ASF. The Yahoo spinoff, Hortonworks, which is currently working to make Hadoop more robust and easier to install, is working with Microsoft on the port. It is providing support, consultancy and training on Hadoop's Java-based platform to Microsoft whilst at the same time helping it maintain compatibility and interoperability.
Microsoft also announced that it had made production versions of Hadoop connectors for its SQL Server and Parallel Data Warehouse available. The CTPs were released two months ago, and the technology will allow SQL Server to be a source of data for Hadoop platforms evaluating "big data" problems. Kummert also announced that "Denali", the code name for the next version of SQL Server, is to be released in the first half of 2012.