We talked to Peter Michielse, deputy director of the supercomputer centre SUFRsara, to get some more information.
There is no doubt that a number of Dutch scientists really do need the fastest supercomputers available to do their research and to stay at the forefront of their field. But they not only need supercomputers, but also fast networks, and Big Data infrastructures. In Europe, the term e-Infrastructure is used for all IT infrastructure that supports science. In the Netherlands, these large e-Infrastructures are mainly operated by the SURF organisations, SURFnet for the network, and SURFsara for the compute and data infrastructures. SURF intends to set aside money it yearly receives from the Dutch government for the next upgrades of the e-infrastructure.
"SURF misses about 8 million euro per year for future investments in the e-Infrastructure. Not only for the supercomputer but also for the networking and data e-Infrastructure. We did not get the money at the level that was intended some years ago. The Dutch government is aware of the problem and working on a solution", stated Peter Michielse. "The replacement for the Cartesius supercomputer is planned for 2017. If we do not have enough money by then, it will be difficult to have a TOP50 supercomputer in the Netherlands”, Peter Michielse continued.
The Dutch national supercomputer is big, and as such provides a lot of capacity for Dutch researchers. But some of the researchers need even more. In Europe this capacity - we are talking about the top supercomputers in Europe - is provided by PRACE. The Netherlands are also part of PRACE. If a researcher wants to get access to a top machine, he has to show his application really works on a large parallel supercomputer. "Running your application with success on a national supercomputer, is helping to show it really does”, stated Peter Michielse. "Without a national supercomputer as a stepping stone it will be very difficult to get access to Europe's top machines."
Dutch scientists have written an outline to the Royal Netherlands Academy of Arts and Sciences (KNAW) explaining their vision for the future about supercomputing and e-Infrastructure needs and how it would enable their scientific work at the cutting edge. The vision is for the time frame 2020-2025. "The 100 million euro mentioned in De Volkskrant article was connected to this vision”, stated Peter Michielse, "not to the short term investment in the supercomputer."
KNAW is not directly involved in running, planning or funding e-Infrastructures. Its main role is to shape the Science and Art research agenda in the Netherlands, and to advise for instance the Dutch government on that. So it is the right platform for discussing what is needed for the future of research in the Netherlands, but not so much about day-to-day investments and operations.
So the good news is there is vision on where to go with e-Infrastructures in the Netherlands in the long run, but on the short term there needs to be some money found soon to be able to renew the supercomputer and other e-Infrastructure parts in the near future.
[This is the second revised version of this article.]