Building Scalable
Systems and
Moving Large Datasets
In order to manage ever-increasing computation
needs, Google continues to scale its hardware and software systems to meet the
need to store more data, serve more requests, and at the same time improve
results. Using Google Code's Subversion
server as a case study, this talk will cover Google's hardware philosophy and
several core infrastructure technologies such as GFS, BigTable, and
MapReduce. We'll also review advances in storage as they relate to
Google's project for moving large scientific datasets.