Previous | Session 108 | Next
T. Axelrod (Steward Observatory), A. Connolly (U. Pittsburgh), Z. Ivezic (U. Washington), J. Kantor (LSST Corp.), R. Lupton (Princeton Univ.), R. Plante (NCSA), C. Stubbs (Harvard Univ.), D. Wittman (UC Davis)
In science observing mode, the LSST telescope and camera system will deliver a 2.8Gpixel image every 12 sec, a data rate of about 0.5 GByte/sec. The data processing pipeline must process this incoming data to produce the LSST's primary data products: the calibrated image archive; the detection catalog; the object catalog; and real-time alerts. Additionally, the quality of the incoming data must be rapidly assessed and fed back to the observatory control system. The pipeline must be flexible enough to allow addition of new processing stages and replacement of existing algorithms with improved ones, and must be robust in the face of the failure of hardware components such as disk drives. We present a preliminary design of the primary LSST data products, and of the data processing pipeline. The mapping of the pipeline onto computing hardware is discussed, along with estimates of the computational, I/O, and network bandwidths required.
Previous | Session 108 | Next
Bulletin of the American Astronomical Society, 36 5
© 2004. The American Astronomical Society.