PanDA for the Large Synoptic Survey Telescope
The PanDA distributed production and analysis system was originally developed for the ATLAS Experiment at the LHC to meet that experiment’s extreme demands for large scale data intensive computing. PanDA processes over an exabyte of ATLAS data per year at a rate of close to a million jobs per day at more than 100 computing facilities worldwide, for about 1400 ATLAS physicist users.
In the coming years other science communities will have computing demands at similar scales, and through the BigPanDA project begun in 2012 the PanDA system has been extended to uses beyond ATLAS and a broader range of computing platforms, notably including HPCs (supercomputers).
The LSST and its Dark Energy Science Collaboration (DESC) is one community that will have large scale computing needs. In a collaboration between the PanDA team and SLAC we have drawn on the generalization of PanDA done in the BigPanDA Project to enable LSST scientists with substantial computing demands to access resources via PanDA, including HPC resources.