Why EMC Wants to Build a High Performance Data Center in Holyoke
Scene: An abandoned brick building along the Connecticut River. The image dissolves, then reforms to show a new, ultra-modern factory in its place. Move to interior shot of computers and server banks. Brilliant academics ponder the future. Highly trained young professionals walk purposefully, the future alive in their eyes.
I have no idea whether the scene depicted above is accurate. But that’s the vision planted in my mind by the announcement earlier this month that the state of Massachusetts and a coalition of corporate and academic institutions—MIT, Boston University, the University of Massachusetts, Cisco, and EMC—had banded together to plan a High Performance Computing Center (HPCC) for Holyoke, MA.
Xconomy broke the news about the HPCC on June 9. The general idea is that a state-of-the-art “green” data and computing center in Holyoke, cooled by the Connecticut’s waters and powered by affordable, renewable hydropower, will establish Massachusetts as a leader in next-generation computing technologies and bring jobs and dollars to the depressed Holyoke region. The center would presumably house the networked servers and storage systems needed to address complex research problems in fields from genomics to climate modeling, as well as various commercial projects. “The potential for breakthrough technologies and research is enormous,” said Governor Deval “the Govinnovator” Patrick in a statement, “and both the center and this collaboration will undoubtedly serve to lift up the City of Holyoke and regional economies throughout Western Massachusetts.”
That’s the essence of the coverage to date. But, of course, that is only the beginning of the story—since the plan hasn’t even been finalized yet. I was especially interested in the corporate angle to all this, since corporate innovation is an issue I’ve spent a book or two exploring. And while to date this column has dealt mainly with entrepreneurs, venture capitalists, and startups, innovation really does happen in big companies, too. And so, against the backdrop of the envisioned High Performance Computing Center in Holyoke, this is a look at such an attempt.
My guest for today’s column is Jeff Nick, SVP and chief technology officer for EMC. The story he tells isn’t just of a data storage and management company wanting to spur the construction of an advanced computing center to grow its current businesses. Rather, Nick laid out a multi-tiered strategy of how EMC plans to combine its shorter-term business ambitions with efforts to explore the future in virtual computing, cloud computing, data management, security, and more within the confines of the fully operational, fully commercial data center.
At a high level, Nick characterized the project as offering “three dimensions of opportunity” for EMC and its R&D strategy. The first lies in the opportunity to test new techniques involving the center’s physical infrastructure, what Nick calls “ping, power, and pipe.” (“Ping” is a new term for me that Nick says refers to the way cables are physically laid down to network computers together.)
Systems infrastructure, as opposed to the physical infrastructure, is the second area of opportunity EMC’s chief technology officer sees. This covers areas such as virtualization—the specialty of VMware, a California company that is majority-owned by EMC—and cloud computing, the domain of EMC’s Cloud Infrastructure and Services Division.
These technologies will be both the undergirding for, and to some extent the subject of, the research projects envisioned for the new center. For instance, Nick says the HPCC will rely heavily on virtualization, in which one computer simulates the action of many, or many computers simulate the action of one massive machine. “We really can create much greater efficiencies in the utilization of resources, because you can run multiple virtual machines on the same physical machine, and therefore drive up utilization dramatically without any sacrifice in performance,” he says. Cloud computing refers to allowing customers to tap the center’s computing power or storage capacity as needed—sharing data in some cases and keeping it private in others, but all within the confines of the center’s systems, not their own. “[It's] really enabling a multi-tenant environment where different companies, different customers, different universities maintain their own private, secure environment for their data and their resources, and also enable multi-party collaboration,” Nick says.
Finding better ways to support this kind of operation and collaboration is a top priority for companies like EMC, VMware, RSA (EMC’s security division), Cisco (the other corporate partner on this project), and others, Nick says.
This is where the R&D aspect of the plan really comes in. The center would be a “friendly test bed” to pursue advanced research and incubate new technologies, improve them, and then “bring them into the marketplace,” Nick says. The university partners are key to this, he explains, because university researchers are often early adopters of new technology. “If we create this environment where they get to play with these toys and give us feedback and then begin to build out additional capability on top of the technology that we supply them, which is what researchers do, then we’re able to accelerate the innovation curve.”
The third of Nick’s “dimensions of opportunity” has to do with the longer-term fruits of the research collaborations enabled by the center as they advance applications in a wide variety of industries. This is the most open-ended of the elements to EMC’s strategy, because it essentially says that good but unpredictable things will ultimately come out of a center that brings together researchers in the life sciences, climate modeling, health care, renewable energy, finance, and other fields. “This is really a great opportunity to allow the researchers from different universities to kind of work side by side,” says Nick. “It really creates a rich Petri dish, if you will, from which new ideas and new discoveries form.”
I asked Nick whether this meant that researchers would be physically present in the HPCC. “I don’t know,” he says, pointing out the center is still in the planning stages and such issues aren’t resolved. “But I don’t really think it matters, as long as the researchers have access to the information and the IT resources.”
A lot of other issues are also still up in the air during this planning phase, and so only time will tell how well EMC’s vision matches reality. But whether it involves a physical presence or not, a research melting pot that tests and hopefully helps pioneer new computing systems, techniques, and architectures even as it helps the bottom line is very difficult for any single corporation to create these days. If you have partners to help bear the cost and risk, and if other partners help you test your technologies while doing their jobs, you might just do what Nick intends to do—pick up the pace of innovation.