Data Farming - Getting the most out of Moores Law and Cluster Computing - Video Portal
Data Farming - Getting the most out of Moores Law and Cluster Computing
Dr. Paul Sanchez, NPS
Dr. Paul Sanchez from the SEED center at NPS, talks about Data Farming at the 2010 CENIC conference.
Data farming combines statistical design of experiments with computer simulations on high-speed/ high-bandwidth computing, to "grow" data in such a way that it will maximize our understanding of the system being studied. The SEED Center for Data Farming was created at the Naval Postgraduate School in 2006 to foster the development and use of data farming techniques, and supports an active international community of developers, analysts, and decision makers through research, training, and biannual workshops. Data farming leverages cluster computing to explore a broad range of carefully chosen scenarios in short order - the international data farming community currently shares resources in Germany and Singapore, as well as at NPS, to support the computationally intensive nature of these explorations. High bandwidth networking enhances our ability to generate and analyze the large volumes of data produced in typical data farming studies. This talk will introduce the fundamentals of data farming and illustrate its benefits. The SEED Center has been at the forefront in developing innovative experimental designs to leverage modern hardware capabilities and enable large scale studies. Recent applications include the analysis of pirate attack scenarios off the coast of Africa, modeling social networks and population dynamics, peace support operations, and organizational effectiveness. We will draw on these in our presentation.