CogniGron Seminar by Chris Broekema (ASTRON): Affordable and Sustainable Data- and Compute-Intensive Radio Astronomy

Get Ready to Surf the Cosmic Data Wave!

Chris Broekema, researcher at ASTRON (the Netherlands Institute for Radio Astronomy), will visit CogniGron to talk about designing efficient and sustainable data-transport and data-processing infrastructures for radio telescopes. 

Research abstract

Radio astronomy is a relatively new area of science that relies heavily on the abundant availability of data- and compute-capacity. One could even argue that the advent of affordable compute power, in the form of general purpose mini-computers in the 1970s, made the construction of aperture synthesis arrays like the one in Westerbork possible. More recently, in the early 2000s, ASTRON designed and built the LOFAR telescope to take advantage of the steady increase of available compute power. This is one of the very first 'software telescopes', where a massive amount of complexity is offloaded from the hardware into the software (with all the appropriate growing pains we are only now starting to really understand).

This predictable and inevitable increase in compute capability can no longer be relied upon. Worse, we are now faced with instruments that can produce far more data than we can affordably process. For future telescopes, like the Square Kilometre Array and even more ambitious later telescopes we could easily be overwhelmed by the virtual deluge of data being foisted on the scientist. In this talk I will go into some detail on the anatomy of a modern large-scale distributed radio telescope, what processing needs to be done with the collected data, and how we are preparing to process the fire-hose of data to be collected by modern instruments. New and emerging technologies play a central role in this strategy.