Yahoo Inc said on Monday it would launch an open source program for developing software for distributed computing, with Carnegie Mellon University as its first academic partner in the venture.
Yahoo has been a main supporter of Hadoop, an open source distributed file system that lets users process massive amounts of data. The company said it would make Hadoop available to academic research in a supercomputing-class data center.
It is interesting that Hadoop is actually an implementation of Google's Map Reduce. Very shrewd move by Yahoo to support Hadoop rather than copying Map Reduce directly, same result but face is saved :)
It also has to be noted here that this "distributed" thing is not what normally people might think of hearing about distributed computing, ie SETI@home and others - Hadoop is a file system and it won't work well over puny Internet connections, so this is really a LAN (1 Gbit+) solution, and really you would do well with 10 Gbit one.