Information Technology Reference
In-Depth Information
18.4.3 PLANCK Mission
The PLANCK experiment is an ESA satellite mission aimed at mapping the
microwave sky by performing a number of complete sky surveys with an
unprecedented combination of sky and frequency coverage, accuracy, stability, and
sensitivity. The satellite flew in May 2009 and performed eight complete sky
surveys before its switch-off in October 2013. However, operations continue after
the switch-off, in particular, data reduction and analysis of the observed sky.
This mission is extremely demanding in terms of storage and computational
needs, since all the data measured by the instruments is reduced and analyzed more
than once in order to improve the data quality and to re
c
results. Moreover, a number of numerical simulations are performed. Simulations
are an essential step in this project for different reasons. First, they allow us to
understand how a given instrument behavior reflects onto the observed sky signal,
and also they represent a test bed for the construction of the reduction analysis
pipeline of the real data.
The simulation task selected to run through the STARnet portal is the PLANCK
simulations pipeline, i.e., LevelS. It produces realistic simulations of the two
PLANCK instruments, both in terms of observed signal (including, e.g. beam
smearing and pointing accuracy) and in terms of instrumental noise. The main goal
of these simulations is to validate the acquisition and reduction procedure. More-
over, they are the core software used to run Monte carlo simulations to estimate the
cosmological parameters.
LevelS simulations are quite expensive in terms of computing and storage needs,
and they can be successfully run on a DCI such as the EGI grid (Taffoni 2007).
DCIs are used also to store data produced by these simulations.
Implementing LevelS simulation on the STARnet portal may allow researchers,
who need to run a number of LevelS simulations with different parameters, to use a
simpli
ne the
final scienti
c problem and not on the use
of the computing infrastructures. The portal offers a simplified environment to
manage the computing tasks.
LevelS is structured as a pipeline and is composed of several elementary stages,
each stage being an algorithm applied onto an input dataset and producing an output
dataset. Stages are chained (e.g., the output of stage A is used as input for stage B),
but not necessarily linearly. The
ed environment focusing on the actual scienti
final output of LevelS is time-ordered data (TOD)
consisting of the time sequence of detector outputs. At any given time, each
detector observes the signal sky composed of a mixture of cosmic microwave
background (CMB), and galactic and extragalactic foreground emissions convolved
with the detector beam pattern. Instrumental noise is added according to detector
characteristics.
We select a very simple pipeline consisting, however, of different software
modules. The basic steps of this pipeline are described below:
The CMB power spectrum is created with cmbfast starting from cosmological
parameters.
Search WWH ::
Custom Search