A report on the state of Canadian high performance computing (HPC) and cyberinfrastructure (CI) wants a small group of experts convened quickly to determine future needs or risk falling behind other nations that have made them a priority for developing their knowledge-based economies. The report — commissioned by the Canada Foundation for Innovation (CFI) — recommends following the lead of the US and UK in establishing an Office of Cyberinfrastructure to develop a national vision and engage the international community.
Concerned that there is an apparently insatiable demand for new research computing capacity, the CFI commissioned the report as a 10-year foresight exercise to better understand the state of pertinent infrastructure, emerging trends and the need for future investments in CI and HPC. The report is being reviewed and a decision on whether it will be released publicly will be made at a later date. It's estimated that CFI has spent $170 million supporting HPC, including $60 million to the seven regional HPC centres that comprise Compute Canada (R$, September 1/10).
"Canada has to start moving and get its act together … There has got to be a strategy," says report co-author Bill St Arnaud, a green information technology consultant and former chief research officer for CANARIE (R$, December 21/09). "The rest of the world has seen this as a critical component of their national digital strategy. In the US, it's at the core of its research and innovation strategy."
Canada has two key organizations engaged in aspects HPC and CI — CANARIE and Compute Canada — but the reports argues that neither is ideally structured or mandated to assist in their coordination and development. Establishing an office to provide leadership in these areas and coordinate investments by the CFI, the granting councils and others would represent a major step forward in ensuring that Canada has a base upon which to build.
"We need a serious re-thinking external to the current milieu … The situation in Canada for cyberinfrastructure is not optimal," says report co-author Dr Denis Therién, VP research at the Canadian Institute for Advanced Research (CIFAR), a professor of computer science at McGill Univ and its former VP research and international relations. "We need to have this conversation of what needs to be done with a few carefully chosen people. The national interest has to be the focus."
To compile information for the report, St Arnaud and Therién interviewed several people engaged in CI in Canada, the US, UK, Belgium, Australia and elsewhere to determine current trends and how other countries are harnessing the potential of HPC and CI for social and economic benefit.
"Other countries are doing things and it's not clear what is being done in Canada or if we even have the tools," says Therién, adding that his views on the issue are not those of any organizations he's involved with. "CANARIE is a networking organization so it's not their role to lead. Compute Canada could play this role as HPC specialists but the (necessary) structure is not there. HPC is a niche within research computing."
In the US, CI and HPC are high on the government's S&T priority ranking, especially since the December/10 release of a report — Designing a Digital Future: Federally Funded Research and Development in Networking and Information Technology — by the President's Council on Science and Technology (PCAST). President Barack Obama has repeatedly highlighted the importance of CI to the nation's prosperity and recently created an Office of Cyberinfrastructure to champion its expansion.
The UK has also established an office for cyberinfrastructure leadership through its research and education network, JANET (UK), which recently unveiled its Connected Shared Research Cloud platform — a shared HPC resource that enables universities to access more real-time research processing than they can afford individually, with the aim of making participating institutions more competitive globally.
Therién says a small group representing researchers, funders, intellectual leaders and public policy groups should be convened to brainstorm on the issue of Canada's future in HPC and CI. They should be given access to international experts from countries that have already gone through this process of examination and quickly produce a report upon which policy can be drafted.
Therién says the PCAST report on advanced networking and information technology (NIT) is one of the best available and reflects "great clarity of thinking". The report lists dozens of applications — from ebook readers and digital photography to cloud computing, scientific supercomputers and speech recognition — that will benefit from further investments in NIT.
"Recent technological and societal trends place the further advancement and application of NIT squarely at the centre of our Nation's ability to achieve essentially all of our priorities and to address essentially all of our challenges," states the PCAST report. "There are enormous opportunities for future transformations. To meet the challenge of change, America must continue to make R&D investments in new areas of NIT."
PCAST recommends that the US "increase investment in those fundamental NIT research frontiers that will accelerate progress across a broad range of priorities" and that "the effectiveness of government coordination of NIT R&D should be enhanced".
In the UK, a major new report commissioned by the Joint Information Systems Committee — Study of early adopters of shared services and cloud computing within Higher and Further Education — estimates that $4 billion can be saved annually by the higher education sector if universities move quickly on outsourcing and shared services of cloud computing and networks are implemented. The report recommends that universities work to reduce their reliance on government funding for support by "increasing their level of engagement with private business in order to diversify their income streams, and seek to take a more entrepreneurial approach to generating revenue from commercial sources".
"Nobody has any money but there are cost savings associated with cyberinfrastructure," says St Arnaud. "The first step is to recognize it as an area of critical importance. We need to get a small group together and then figure out how to pay for it later, but there are significant costs savings as well"
R$