With its launch at the very end of 2013, ESA's astrometry satellite Gaia began its endeavor to compile astrometric and photometric measurements of at least one billion objects, as well as high resolution optical spectra of hundred million objects. The Gaia catalog therefore results in a wealth of coherently determined astrophysical parameters of these objects. After its extensive commissioning phase, Gaia entered the nominal mission phase in July 2014. The science ground segment, which is formed by the Gaia Data Processing and Analysis Consortium (DPAC), has since then started its operations. DPAC is a large, multi-national, science consortium which has to handle and process the dense and complex Gaia data stream. With its decentralized management and its distributed infrastructure, the Gaia DPAC is a remarkable undertaking. In this paper we will summarize some of the experiences of the DPAC facing the real Gaia data, compare this to the pre-launch expectations, and critically review the development phase.
The Gaia survey mission, operated by the European Space Agency (ESA) and launched on 19 December 2013, will survey approximately 109 stars or 1% of the galactic stellar population over a 5.5 year period. The main purpose of this mission is micro-arcsecond astrometry, that would yield important insights into the kinematics of the galaxy, its evolution, as well as provide important additional findings, including a updated coordinate reference system to that provided by the ICRS. Gaia performs its observations using two telescopes with fields of view separated by 106.5 degrees, spinning around an orthogonal axis at about 6 hours per day. The spin axis itself precesses: it is always oriented at 45 degrees from the sun, and precesses around the sun every 63 days. Thus each part of the sky is observed approximately every 63 days. The 6-hour spin, or scan-rate matches the CCD readout rate. The amount of data to process per day - 50-130 Gigabytes - corresponds to over 30 million stellar sources. To perform this processing, the Gaia Data Processing and Analysis Consortium (DPAC) have developed approximately 2 million lines of software, divided into subsystems specific to a given functional need, that are run across 6 different Data Processing Centres (DPCs). The final result being a catalog including the 109 stars observed. Most of the daily processing is performed at the DPC in ESAC, Spain (DPCE), which runs 3 main subsystems, the MOC Interface Task (MIT), the Initial Data Treatment (IDT), and First Look (FL). The MIT ingests the initial data provided by the MOC in the form of binary data and writes (amongst other things) `star packets' containing the raw stellar information needed for IDT, which provides a basic level of processing, including stellar positions, photometry, radial velocities, cross match and catalogue updates. FL determines the payload health (e.g, the health for the 106 CCDs, geometric calibration) and astrometric performance via the one day astrometric solution. This presentation provides an overview of the DPAC software as a whole, and focuses on the daily pipeline processing: the systems used, the teams involved, the challenges during development and operations, and lessons learned.
This document describes the uplink commanding system for the ESA Gaia mission. The need for commanding, the main actors, data flow and systems involved are described. The system architecture is explained in detail, including the different levels of configuration control, software systems and data models. A particular subsystem, the automatic interpreter of human-readable onboard activity templates, is also carefully described. Many lessons have been learned during the commissioning and are also reported, because they could be useful for future space survey missions.