HUNTINGTON, NY — New York later this year will knit into a single network the hundreds of high-performance computers scattered among businesses and academic and nonprofit institutions to enable faster and easier research in the life sciences and other fields, the state’s top technology development official told BioRegion News last week.
The initiative was originally intended to begin this year, though it was cancelled after its funding was removed from the state’s budget. Officials said this year’s plan, which is expected to be launched by the fall, will cost a fraction of the original estimate.
Edward Reinfurt, executive director of the New York State Foundation for Science, Technology, and Innovation, or NYSTAR, said his agency is currently evaluating responses it received last month to a request for information from academic institutions and businesses interested in participating.
“NYSTAR would expect these services to be provided at a minimum of five different locations throughout the state, either through individual awards or through a virtual center,” the RFI stated.
The responses will help state officials fine-tune a more formal request for proposals to run the network, which would include the computational biology consortium announced last year [BRN, June 4, 2007]. That consortium was to be based on two IBM Blue Gene/L supercomputer systems, situated at the State University of New York—Stony Brook University and Rensselaer Polytechnic Institute. Both schools joined Columbia University, Cornell University, New York University, SUNY University at Albany, and SUNY University at Buffalo to form the consortium.
Since then, NYSTAR has either discussed with, or heard expressions of interest from, Marist College, Rochester Institute of Technology, and SUNY Binghamton University. Some urged the state not to define categories of scientific research that will participate in the network, while others tried to align the way in which they plan to use the network with a state list of key industries, which includes the life sciences, Reinfurt said.
“We’re going to work with the physicist, the biologist, the chemist, whoever wants to run applications.”
Also since last year, state officials have expanded their vision for the high-speed computer network beyond biology to include other areas of science — especially long-term, cross-disciplinary collaborations that could enable researchers to make discoveries across a number of scientific fields.
“Our goal is to probably get the RFP going in the next two to three weeks,” said Reinfurt, who spoke with BRN minutes after addressing the 2008 Life Sciences Industry Summit, presented by the Long Island Life Sciences Initiative, and held here last week. “We don’t see the need for a longer response period than a month from that time, and foresee making the decision, I would say, in the August time frame.”
RFI respondents were required to share their thoughts with NYSTAR on several questions, including:
- What do researchers at academic and not-for-profit research institutions and technology companies need to successfully allow them to take advantage of these assets?
- What state-owned resources are already in place across New York that can be taken advantage of for these purposes?
- What would be the most effective funding structure to achieve the goals: Individual awards to individual institutions? Designation of a lead institution who would work with others? Or establishing a virtual center that will oversee funding and activities to be undertaken by partner institutions?
- What staff resources would the service providers need to allow them to successfully perform these activities?
- What should the programmatic requirements be to allow NYSTAR to successfully achieve its objectives?
- In what ways can NYSTAR serve as an incentive to multi-disciplinary collaborations on the use of high performance computing in New York State through this RFI?
- On what specific industries or sciences, if any, should NYSTAR focus its high-performance computing resources, and why?
- Are there any unforeseen barriers or challenges that can cause this proposal to not fulfill its objectives?
“We will be funding an initiative that will not limit the participants at all in terms of any membership structure,” Reinfurt. “We believe that they will become guiding members of a group of both private firms and universities in developing our strategy going forward.”
NYSTAR plans to serve the network through another entity, but has not decided whether it would be a new not-for-profit or an existing body. “We’re going to work with the physicist, the biologist, the chemist, whoever wants to run applications.”
The multidisciplinary focus is expected to help resolve a longstanding issue: how to maximize the use of the state’s supercomputers. “We have, I think, in the case of Stony Brook, 30 percent use of the capacity, which we can assign,” Reinfurt said. “The person that may want to use it is generally not a computational scientist [and] doesn’t know how to use a BlueGene computer.
“We’d love to have the problem of having to ration the areas of use. We’re not there yet. We do want a diversity of use,” Reinfurt added. “We have some priority areas that we will be identifying, and we want to complement that.”
NYSTAR also wants to synchronize its computational policy with that of national laboratories, which have generally focused resources on assisting users rather that buying equipment, Reinfurt said.
Reinfurt told BRN that state officials had initially intended to launch the program this year with an allocation of $4 million from the state’s fiscal 2008 budget — half of the $8 million initially sought by the computational biology consortium when Stony Brook and RPI announced the initiative last year.
“We were well along the way with the $4 million of creating a staffing structure which would basically have been the type of hands-on assistance that various experts in various disciplines would need to take advantage of the supercomputers,” Reinfurt said. “There would be, maybe, post-doc[toral] students with visualization, simulation skills, algorhithm skills, basically taking all the data that a professor had compiled for 10 years and putting in a code format that could run through a super[computer], and then teaching them the application features.”
But when the state’s budget shortfall reached $4.6 billion earlier this year during the crafting of a new $121.7 billion spending plan, officials responded by shelving a wave of new programs. Among them were the consortium and two other programs intended to boost the state’s life-science effort: hiring of 250 new researchers or “eminent scholars” among 2,000 new full-time faculty members for the State University of New York system, and creating a $3 billion, 10-year research funding program to be paid for by partially privatizing the state lottery [See Around the Regions, this issue].
Now, NYSTAR plans to use $1 million of its own funds to launch the computational network “to provide hands-on assistance to users of the facilities,” Reinfurt said.
Then as now, the goal of the computational consortium is to knit into a cohesive cluster the crowds of computers used by New York’s far-flung biotech sites — an enduring promise by officials and life-science leaders that has largely remained unfulfilled.
Reinfurt said the network could serve as a valuable tool in the state’s ongoing efforts to attract new life-science and other tech-based jobs and retain existing ones.
That would gibe with one of the original purposes of the initial consortium discussed last year: Making supercomputers more easily available and easier to use for corporate users. The original consortium also discussed the need for support staff to help run the machines, as well as developing statewide regional centers where support crews could work directly with researchers.
“We very much see this as an economic-development tool to the extent that supercomputing can provide breakthroughs in products or make them more competitive or unique,” Reinfurt said. “We would like this to augment the product-development cycles of companies.”