Since 2007, Google has been working with IBM to provide universities with access to compute clusters based on Hadoop, an open-source cluster system based on Google’s distributed file system GFS and using Google’s open-soruce framework for massively parallel data-crunching MapReduce. It’s a move to prep students for work on the types of Massively-parallel algorithms in use by HPC tools and companies like Google and Yahoo. Google and IBM have shoveled between $20m and $25m into the project, and now they’ve got another $5m from the National Science Foundation.
The computational and storage resources provided by this Google-IBM initiative allows us to perform complicated interactive analysis of a pretty-much unprecedentedly large amount of data ” Claudio Silva associate professor at the University of Utah tells The Reg. “It has the ability to completely transform the way we do data analysis and visualization…