Recently, Kim reflected on 20 years of working with NCSA as a university professor, an industry researcher with Eli Lilly and Company, and an NSF administrator. He also shared his thoughts with NCSA's Karen Green on cyberinfrastructure and what it means to science, engineering, and the world.

Q: Before you came to NSF, you were a partner for several years. Can you tell us about that experience and why it was important to bring such a variety of disciplines together under the centers program?

A: I consider the opportunities I had to partner with the national partnership at NCSA as one of the highlights of my career. One of the especially memorable activities happened during the period from 1995 through 97 when I was partnered with Shankar Subramaniam [then at NCSA] in the computational biology team, and one of the things that we did–and at that time this was a revolutionary new idea–was use the power of the newly emerging cyberinfrastructure of the time, namely the World Wide Web and the Mosaic Web browser, to show that access to supercomputers didn't mean that you had to write a proposal for supercomputer time.

We had this vision that biologists could use something called the Biology Workbench and have access to supercomputers in a seamless fashion through the front end of a Web browser. In many ways, those ideas have become very commonplace and are the standard cyberinfrastructure models for today.

Q: Talk to us about cyberinfrastructure. What does NSF want to see as the end result of these efforts?

A: I'm struck by the fact that through the revolution in information technology we are able to deliver terabytes of data per day. In fact, in the very near future it might even be petabytes of data per day. But the scientific end users frequently complain that we are only extracting kilobytes of knowledge per month.

That paradox suggests that we still have a long way to go in creating a cyberinfrastructure that can address that bottleneck. How do we generate useful knowledge and insights from the torrents of data that we are creating? I think that challenge will continue for the foreseeable future. And so the promise and importance of cyberinfrastructure is predicated on the vision that we can create new tools and technologies so that the terabytes and petabytes of data that we generate can scale to the kind of knowledge that we know we can create.

Q: What do all these efforts–such as TeraGrid and cyberinfrastructure–mean to the world at large and how will they impact people’s lives?

A: What is especially exciting about cyberinfrastructure, and recent history shows this to be the case, is that the cyberinfrastructure that is created to enable science and engineering research has immediate applications in the everyday economy. That was true of the World Wide Web and the Internet, and I think it is also going to be true of the new and emerging cyberinfrastructure that we are creating to enable the new frontiers of science and engineering research. I would include in this high-speed networking technologies, new ideas in middleware, new ideas in data mining, data warehousing, and data storage.

Q: Do you want to go out on a limb and say precisely what those impacts will be?

A: I don't think it will be any one impact; it's going to be a broad range of impacts. But I think one could say that the digital revolution will continue and in many ways it will accelerate. Already we are starting to see intelligent cars and roads, intelligent manufacturing. I think one of the great frontiers in terms of our economy is going to be our ability to create exactly what we need, where we need it and when we need it, and thereby eliminate waste and inefficiencies. The United States has a $10 trillion economy, and it has inefficiencies in the supply chain and distribution. I see the next wave of the cyberinfrastructure and IT revolution having a tremendous impact in eliminating waste and ultimately fulfilling our vision of a greener economy.

Q: What are NSF's goals for the next five to 10 years in terms of shared computational infrastructure? What would you personally like to see happen?

A: The overarching goal in shared cyberinfrastructure–and the key word is shared–is that we believe there are infrastructures that are common across all science and engineering, and it would be inefficient and wasteful to duplicate or have redundant resources or siloed resources across multiple branches of science and engineering. That's actually the primary thinking behind this shared cyberinfrastructure concept. I think that is an evergreen vision.

Q: NCSA has had a successful relationship with NSF for 20 years. How do we keep that going in the future?

A: As an organization that delivers cyberinfrastructure, NCSA has a tremendous challenge to provide a balance between stability and innovation. Users of cyberinfrastructure expect stability and desire stability. At the same time, you are working in a world that is undergoing a very rapid sequence of technology disruptions, and therefore innovation continues to be important.

NCSA has historically met that challenge, and the historical lessons of how that innovation occurred while delivering stability is something that will be important. Perhaps a strategy for going forward can be gleaned from what worked well in the past.