IBM and ASTRON (ASTRonomisch Onderzoek in Nederland, the Netherlands Institute for Radio Astronomy) announced an agreement for a collaboration that will lead to the construction of a new supercomputer for the collection of data from the SKA (Square Kilometre Array), the consortium that aims to build a radio telescope consisting of the largest array of the world. ASTRON allocated an initial sum of 32.9 million Euros for a five years collaboration.
ASTRON is one of the most important scientific partners in the consortium that is developing the SKA. In 2024, when the array is supposed to be completed, it will be the most sensitive radio telescope in the world and will be used to investigate the evolution of galaxies, dark matter and even the origins of the universe.
An instrument such as the SKA will requires a computer with performances superior even to those currently existing in supercomputers. In fact, the millions of antennas that will compose it will detect radio signals in an area equivalent to one square kilometer but they will be scattered in an area roughly the size of the continental USA. The SKA will be fifty times more sensitive than the existing radio equipment and more than ten thousand times faster than current instruments.
It’s estimated that the data the SKA will detect every day will create an amount of data twice today’s global Internet traffic. The storage capacity required after the data will be processed is estimated between 300 and 1,500 petabytes per year. To have a comparison, the Large Hadron Collider at CERN produces about 15 petabytes of data a year.
It’s therefore necessary to build a supercomputer that achieves a new level of computing power, the one called exascale computing, in order to handle the data flow that will come from the SKA. The collaboration between IBM and ASTRON, called DOME, has this purpose.
The challenge is enormous even for a computer science giant such as IBM. In fact, the technologies required to build a supercomputer suitable are more advanced than those currently existing. In essence, the problem isn’t to build a bigger supercomputer but to further develop existing technologies or even to develop new technologies.
IBM researchers have a number of ideas, we’ll see which ones will actually be developed. This will be really a scientific and technological extraordinary enterprise and, as always happens in these cases, in the coming years new technological developments will come in one way or another in our homes.