|
Chris Smith
Abstract
Astronomical investigation has traditionally been based around analysis of data from several hours to several nights of observations, amounting to several gigabytes of data interactively processed by individual researchers. However, this model of investigation is rapidly evolving towards one involving much larger scale datasets, in which astronomers cull through terabytes and eventually petabytes of data to identify trends in large samples. These datasets often span a variety of wavelengths, including data from radio, infrared, optical, X-ray, and other telescopes, allowing astronomers to tap into the additional information the broad wavelength coverage offers. At the center of this new model of investigation is the concept of the "Virtual Observatory" (VO), from which astronomers can effectively extract and analyze data from a variety of archived observations. Large surveys covering much of the sky are key to the utility of the VO. Among the most promising (and challenging) upcoming survey projects is the Large Synoptic Survey Telescope (LSST), which plans to cover the available sky every four to five nights, obtaining 15 to 20 terabytes per night, and accumulating an archive of almost 5 petabytes per year. These data must be transferred from the mountaintop telescope to the archive center, where they must be rapidly processed and archived to allow access to the astronomical community using the data mining tools the VO will offer.
|
