SYSTEMS

By Steve Fisher, Editor In Chief -- Always a heavy hitter at SCXX shows and in high performance computing in general, an interview from the show with Sun Microsystems personnel was a must. Supercomputing Online spoke with Fred Kohout, director of marketing for Sun’s technical products group and Joerg Schwarz, group manager, scientific engineering & computing about the company’s focus on grid computing and more.
Supercomputing: Gentlemen, thank you for the interview. How's your show been thus far? SCHWARZ: This is a great show for Sun Microsystems. We have on the show floor a demonstration of grid computing. You can submit a compute job on a technical compute portal at the Sun booth and it will be executed both at the San Diego Supercomputer Center and Ohio Supercomputer Center. We have a lot of cross virtualization with our partners, supercomputing centers and computing centers from all over the world. We had 120 customers signed up for the pre-show Sun HPC Consortium meeting. We have a big attendance scheduled for our Grid Breakfast on Wednesday morning, we have all our briefings booked, etc. There’s a huge interest in both the academic and the commercial worlds in what Sun is doing and Sun’s vision, SunONE architecture, Sun open network environment architecture, and grid computing. Supercomputing: How about a comment on one of the really hot topics these days...grid computing. In particular how it relates to the academic and scientific space. KOHOUT: I think it’s safe to say that grid computing has been the flavor of the month for the last couple of months. We recognize that’s the case, but we’ve actually had grid computing on our radar screen from both a customer requirement and a product and technology requirement for well over two years. We actually acquired a company called GridWare over a year ago and released some software based on that acquisition that we branded Sun Grid Engine. This is just one example of Sun’s strategy around grid computing. We’re going to own and invest in the IT to enable grid computing. This is a very fundamental difference between Sun and other system companies. We’re going to do for grid computing what Sun did for the Internet. We’re going to deliver the technologies whether it’s hardware or software to enable this to occur. When Sun talks about grid computing we identify three clear stages of deployment in the way that people approach and deploy grids for problem solving. First is the cluster grid, the basic foundation of grid computing. We think of this as supporting one team of people, one set of resources. So that in a nutshell kind of defines cluster grids. The second stage is campus grids. Here we’re still inside the organization, the campus grid is multiple teams of people and multiple sets of resources. Here we’re sharing resources. Teams are working collaboratively in the sharing of resources so of course that brings in some organizational issues, people working together, budgets across teams and so forth. The third stage is global grids. Here we’re talking about multiple organizations, we talking about jumping across organizational boundaries in deploying grids. So we see those as three substantial boundaries. You start with cluster grids to get into the model, to get the benefit of the model. With today’s pressure for ROI people are increasingly using cluster grid technology and this where our Sun Grid Engine Enterprise Edition product is focused. Supercomputing: What are a couple major foci here at this year's show? How about for the year ahead? KOHOUT: Number one, the three stages of the grid: cluster, campus and global. It’s the software stack needed to enable grid computing. And it’s the partnerships needed to enable grid computing. In terms of looking in the year ahead I think you’re going to see continuing emphasis on the software stack and new developments in hardware. SCHWARZ: I agree. We’ll also see a very strong adoption of the SunONE architecture including the directories space, name space which is a fundamental requirement for successful grid computing infrastructure. Supercomputing: Where do you see Sun being more successful in 2002, the academic/research space or the commercial space? How about your government business particularly in areas such as intelligence and defense? KOHOUT: As a company we enjoyed a real fine position in academic research and we’ve been able to develop this company over the last almost twenty years into being not defined by academic and research alone, but as being a category player in network computing. What Cisco is to switches and routers, we are to network computing. It doesn’t matter what industry you’re in. SCHWARZ: I have a little different perspective on that. My perspective is that with the economic conditions we have right now, of course the academic and research and government business is critical for Sun. What we are experiencing right now is that most of the projects are still going on. Governments, universities, research centers are still investing. We see the governments in growth areas like bioinformatics invest a lot of money to basically prepare their economies for the next boom, which as many people agree is in the area of computational biology and life sciences. So we see a lot of investment, a lot of things going on and I believe we will be very successful in the academic and research space and I think we will also be successful in the commercial space. Supercomputing: As a company, what is Sun most proud of at SC2001? How about you personally? SCHWARZ: What makes me proud is that we can not only talk about grid computing and ‘Sun powers the grid’ but also show that we are a little bit more advanced than some of our friends and can show what we do with our customers and how we do it and I’m proud that our customers understand our technology contribution and our collaboration and demonstrate how we work together at the supercomputing show. KOHOUT: If you ask me what I’m really most proud of I think it’s the fact that after almost twenty years we’re still approaching the market the same way we always did which is we’re a product and technologies company, we look to partner with other companies and institutions to enable them to solve problems or enable a market to come to fruition. And that’s in stark contrast to how others would approach the market and ask you to buy only from them and only what they deem to be good. So I really see after twenty years we’re still very much a collaborative partner in the technical computing space. ---------- Supercomputing Online wishes to thank both Frank Kohout and Joerg Schwarz for their time and insights. ----------