HiPEAC Vision advocates disruptive change in computing systems research

The HiPEAC European Network of Excellence announces its fifth HiPEAC Vision. The 72-page document presents its recommendations for Horizon 2020 based on an analysis of market trends, a discussion of technology constraints and opportunities, and a review of Europe's strengths and weaknesses in the field of computing systems. The HiPEAC roadmap calls for particular attention to the EU's position as it poises to enter a new phase of scientific innovation.

"The key theme of this roadmap is the imminent end of uninterrupted exponential growth in computation, communication and storage capacity with potentially disruptive consequences for the global economy" explains Marc Duranton, a leading researcher at CEA in France and the chief editor of the vision document. "The current technological solutions do not scale well anymore due to the fundamental laws of physics, and there are no candidates that could replace them in the short term, and continue scaling at the same pace for another decade or more."

During the last decade, information technology has evolved faster than society can adapt, and change keeps accelerating. Consequently, "an increasing number of people feel uneasy about the effects of the progressively omnipresent use of information technology: loss of privacy, job destruction, environmental impact, security and safety issues" notes Koen De Bosschere, coordinator of the HiPEAC network. "Some people start agitating against innovations like e.g. Uber" he adds.

If technology cannot provide the ever-increasing compute power to develop new applications, and if the market does not embrace innovations anymore, it will lead to the end of computing as we know it. In order to change this trend we need new insights to continue performance scaling, and to make innovations acceptable by society. This will need more than "business as usual".

Duranton continues "The computing research community will have to leave its comfort zone, and search for disruptive solutions. If it does not find an affordable storage technology that continues scaling, the cost of big data might become prohibitively high. If it does not manage to find an effective solution for energy scaling, exascale systems will be unaffordable. If it does not find effective solutions to make our apps safe and secure, people might not use them".

The holy grail for the computing industry is to introduce the next killer application. It is hard to predict what that application will be, but it is very plausible that it will be connected to the internet, interact with the physical world, and be smart. In order to support all these functions, it will need more computational, storage and communication resources. In order to be acceptable for the users, it will have to be dependable, safe and secure.

To ensure the further growth of information technology and to address these concerns, several key challenges need to be tackled: energy efficiency, dependability, complexity. We also need models and techniques for information systems to naturally interact, analyze and interpret interactions with the physical world. On top of this we must develop better tools and models to empower non-computer scientists to design their own information systems. There are simply not enough information technology experts to solve all the problems society could benefit from it. Finally, we must search for and industrially develop alternative technologies to continue the performance growth of information systems.

More information:
More information: www.hipeac.net/roadmap

Provided by CORDIS