Related topics

Yahoo! invites world of boffins into 4,000-node Hadoop cluster

Hello! Yahoo! is! a! tech! company!

DVD it in many colours

Yahoo! has opened up its Hadoop research cluster to computer science boffins at four additional US universities, including Stanford, the University of Washington, the University of Michigan, and Purdue.

The company's M45 cluster — a Hadoop setup spanning 4,000 processors and 1.5-petabyte of disk space inside a data center at Yahoo!'s Sunnyvale headquarters —– was originally launched in 2007, and is now available for Big Data research at eight universities across the country. Other participants include founding member Carnegie Mellon, plus the The University of California at Berkeley, Cornell University, and the University of Massachusetts at Amherst.

Speaking with The Reg, Thomas Kwan, director of research operations for Yahoo! Research, reaffirmed the company's commitment to Hadoop research, saying that the M45 continues not only to drive development of the open source distributed number-crunching platform, but to serve as a means of educating future Yahoo! employees.

Though the M45 is a godsend for Big Data boffins, it's a rather large inconvenience for the mainstream media, which insists on telling the world that Yahoo! is not a tech company. It can be difficult to explain why a non-tech company is funding a 4,000-processor and 1.5-petabyte data cluster that crunches epic amounts of data using an open source distributed-computing platform that now underpins everything from Facebook to Twitter to a portion of Microsoft Bing.

Hadoop — whose development was largely driven by Yahoo! — is based on the GFS and MapReduce distributed-computing platforms developed at Google. In 2004, Mountain View published a pair of papers on these technologies, and soon, Doug Cutting — known previously for developing Lucene, the open-source retrieval library — used them to start an open source shadow project. He called it Hadoop after his son's stuffed elephant.

Today, the M45 drives such projects as the Carnegie Mellon natural language–processing system known as NELL (Never Ending Language Learning System) and the University of Washington's sweeping knowledge database project KnowItAll. The cluster is only available to universities. ®

Sponsored: Driving business with continuous operational intelligence