EU to build €6.7B Europe-wide science cloud

The EU is to spend €6.7 billion on a “science cloud” to provide computing power and data storage, giving the continent's 1.7 million researchers access to a virtual environment to store, share and re-use their data across disciplines and borders.

The results of future Horizon 2020 projects would be go to the cloud "by default", said Commissioner for Research, Carlos Moedas.  

Of the €6.7 billion, about €2 billion will come from the Horizon 2020 programme for research and innovation, with the remainder coming from  sources of financing at EU and national level, and from industry.

A pilot platform will start at relatively small at €50 million. Only researchers involved in the pilot organisations will have access at first, but ultimately it will be made open to everyone.

By 2020, the cloud would also link in two prototype next-generation supercomputers, of which one will rank among the top three in the world, according to the Commission.

A key benefit of cloud computing for researchers is that they can upload their files and then download them again wherever they are. This can be useful for collaborative work, or for allowing multiple people to access the same files.

For instance, the Collaborative Cancer Cloud, an initiative launched last year by Intel and Oregon Health and Science University, enables hospitals and research institutions to share patient genomic, imaging, and clinical data securely.

Cloud services have created unprecedented demand for highly-educated engineers and mathematicians, who can build and operate these data centres, so the EU initiative could create a lot of extra jobs.

Fragmented infrastructure

“Europe is the largest producer of scientific data in the world, but insufficient and fragmented infrastructure means this big data is not being exploited to its full potential,” according to the Commission.

While there are multiple initiatives in Europe to provide cloud space for researchers, there is still not enough capacity, said Bob Jones, computer scientist at CERN. “If you look at the largest cloud infrastructure in Europe, it can handle maybe 40,000 - 50,000 users. But we need to satisfy 1.7 million researchers and 70 million science and technology professionals,” he said.

Jones coordinates the Helix Nebula project a portal involving 27 cloud providers and labs. Partners include some of the biggest producers of data in Europe, including the European Space Agency, CERN, the European Molecular Biology Laboratory, France’s National Centre for Scientific Research and Germany’s DESY.

Other large research clouds in Europe include ELIXIR, which manages biological data, and the European Plate Observing System, which monitors the earth’s crust.

The EU cloud hopes to bring these existing facilities together in a way that provides researchers with seamless access.

Opposition

Academic publishers view the initiative as a threat to their business. “I must admit that as a taxpayer I get very concerned,” said Derk Haank, chief executive officer of Springer. “All these initiatives to create infrastructures never came to much in my opinion. All our research is already in the cloud. Why create a new infrastructure? I don’t see the point,” he said.

Similarly, Swedish Member of the European Parliament (MEP) Christofer Fjellner is also unimpressed by the initiative. “There are hundreds of cloud services out there. Google Drive, Dropbox, OneDrive and iCloud are just [some] examples. Why is the Commission trying to re-invent the wheel?” he asked.

Jones says a public cloud would partner with private companies while bringing costs down and offering better protection for commercial ideas. “We want to profit from the private sector investment and innovation that have led to the examples listed by MEP Fjellner, but ensure that the intellectual property of publicly-funded research is respected and that commercial cloud services are offered under acceptable terms and conditionsm, which is not the case today.”

Benoit Hudzia, head of Irish operations at Stratoscale, a cloud provider, questioned the cost efficiency of creating a federated cloud. What the Commission is doing is akin to tying many small electricity generators together, when they should be investing in one big power centre, he said.

“If you really want to create a cloud, you need to give money to a lab that knows what it’s doing, instead of sprinkling money around to distributed centres,” Hudzia said.

While a pan-European science cloud may come with a big price tag, it should be viewed the context that European universities spend over €1.9 billion annually on big deals for subscription access to scientific journal content alone, says Susan Reilly, executive director of Liber, an association of European research libraries.

The cloud will “increase the impact of this expenditure on research by providing the infrastructure to enable every European citizen to be able to access the outcomes of publicly funded,” Reilly said. “€6 billion over the space of five years could easily be recouped in just the productivity by enabling researcher to do text and data mining at scale.”

The EU imprimatur on a cloud platform will increase trust among scientists, said Jones. “Researchers are relatively inexperienced with cloud services and some are concerned it’s a bit lawless.”

He acknowledged there may be some initial resistance to the idea of an open-science commons. “At the moment, the way the research community works, it’s not really about sharing research results.”