Calibration of LOFAR ELAIS-N1 data in the Amazon cloud

The International LOFAR Telescope is a next-generation software-driven telescope operating in the poorly explored 30–240 MHz frequency range. With an unprecedented field of view, and multiple beams, LOFAR is opening up a completely new phase of radio astrophysics, as well as being both a scientific and technical pathfinder instrument for the SKA.

 

However, the calibration of LOFAR data presents several challenges:

  • a) the effects of the ionosphere introduce systematic errors that can not be overcome with the standard calibration pipelines and prevent continued integration to reach deep fields,
  • b) the huge amount of data and processing power involved requires the use of a parallelizable solution, and,
  • c) the management of the specialized software is not trivial.

All of these challenges will also apply to SKA data.

 

Recently, an innovative calibration strategy, able to reach the theoretical noise limit, has been developed. The cloud infrastructure of Amazon Web Services (AWS) provides a parallelizable (distributed) solution ideal to solve the problems with the data management, the computing constrains and the way the specialized software is dealt with. With this project we aim to:

  1. Adapt the state of the art calibration pipeline to the Amazon cloud and release it.
  2. Identify and publish the best strategy to handle data and resources in the cloud for the calibration.
  3. Process the existing data for the ELAIS-N1 field (LOFAR’s First Public Deep Field) to produce a wide-field image down to the theoretical limit, facilitating a vast array of science.
  4. Investigate the feasibility of AWS to process all LOFAR Survey Fields.

This project, which is led by the University of Edinburgh, has been granted in the AstroCompute in the Cloud call by Amazon Web Services and the SKA Office.

 

AMIGA participates in this project providing its expertise on porting astronomical applications to distributed computing infrastructures as the Cloud systems.