The flood of Big Data now being created across the sciences and social sciences has produced extraordinary new intellectual and empirical opportunities with implications throughout society. While high performance computing has long been available for processing such large data sets, researchers and institutions must carefully design network computing environments in order to effect useful analyses in real time. One of the greatest limitations of such analyses at Dartmouth currently is simply moving data from where it is produced and stored to where analyses are done: simply getting the data to the computer for analysis can now often take longer than the analysis will take once the data is there. Dartmouth has a number of superior high performance computing assets, but these currently exist as silos on the network. The work supported by this proposal is consolidating Dartmouth's high performance computing assets onto one high-speed network - the Science DMZ - that is separate from the main campus network. The Science DMZ allows optimal configuration of the Dartmouth computing clusters and file storage systems; alleviates file storage and movement issues from the workflow of researchers; and provides a robust and secure scientific network backbone for future system growth and innovation. The Science DMZ network will tremendously increase the pace of scientific research done at Dartmouth. In particular, the Science DMZ will greatly expand the abilities of the more than 500 undergraduates who perform independent research each year and graduate students across Dartmouth's graduate and professional programs to accomplish their work.
