Cloud Computing Vs. $100 Million Data Center

Citizens of Holyoke, Mass., had reason to celebrate this week as the governor of Massachusetts and other dignitaries--including Cisco CEO John Chambers, EMC CEO Joe Tucci, and the presidents of Boston University, MIT, and the University of Massachusetts--announced plans to build a $100 million data center in their town. It's an ambitious proposal, but is it necessary?The Holyoke high-performance computer center--at this point, it's a mere glean in their eye--would support academic and corporate research, create local jobs, and serve as a state-of-the-art facility for green computing. Advocates enthusiastically describe it as a "cutting edge concept" and "absolute game changer" for the region. Governor Deval Patrick said the center would "lift up" Holyoke and surrounding economies.

Yet, it's hard to fathom why such a tech-smart group would sink $100 million into a data center construction project at the same time that many others see the emergence of on-demand cloud computing services as a faster, better, cheaper alternative, one not requiring capital investment, bulldozers, and hydroelectric gear. This is the stomping grounds of "Big Switch" author Nicholas Carr, after all.

As I see it, the folks behind the proposed Holyoke data center really have two scenarios to consider: One in which they, as proposed, build and operate a new data center, in essence becoming a cloud services provider; and a second in which they leave that work to Amazon, IBM-Google, or another experienced data center operator and pay for IT services only as needed.

Another idea: Let Cisco or EMC operate the data center, thrusting themselves directly into the cloud services market.

The partners in this venture have 120 days to come up with a plan, during which they need to pick a site, figure out how they will pay for the project, and all the other details that go into building a data center from scratch. Their evaluation should include an assessment of whether third-party cloud services would support the research needs of the universities and others involved.

The drawback to a cloud services approach is that Holyoke wouldn't get the new construction and other related jobs that local politicians and residents understandably hope for. But job creation is a bad rationale for building a data center, anyway. Better to focus on research, innovation, and the myriad business opportunities created by on-demand IT resources regardless of how they're delivered.