Of course, others would argue that energy, water, or increasing power density might profoundly change data centers before “big data” and storage become too big. In fact, I heard these points made the previous day in a panel discussion organized by the CFRT’s Bruce Myatt of M+W. The panel included Bloom Energy’s Peter Gross, Peter Gross, Ron Hughes, director of the California Office of Technology Services, Zahl Limbuwala of BCS Chartered Institute for IT, and Ben Radhakrishnan of the Lawrence Berkeley National Laboratory, so its conclusions are not to be taken lightly.
Uptime has been ahead of the curve before, as when it identified increasing power density as a problem, which eventually led the industry to develop containment, modularity, PUE, and DCIM, among other solutions. I’m afraid that virtualization made possible extreme power densities that we see in some facilities, and the meltdown caused by Moore’s Law was just an unintended consequence.
Unintended consequences have a way of disrupting everything in their wake.
Recently, for instance, Sam Macrane of Daikin McQuayrecommended The Box to me. Subtitled “How the Shipping Container Made the World Smaller and the World Economy Bigger, The Box details the changes that the introduction of the container caused to shipping, trucking, and railroads, port cities, longshoremen, stevedores, the worldwide economy, and more. I understood, for example, what happened to London’s dormant east end, which was the center of the city’s once-thriving shipping industry and which city officials hope the Olympics can revitalize. According to Levinson, the container drove shipping costs to almost zero, which made possible worldwide adoption of just-in-time manufacturing and reduced consumer costs. Most of these consequences never occurred to Malcolm McLean, who pioneered the idea in the 1950s.
Events like DatacenterDynamics, where I met Bruce, and 7x24Exchange, where I spoke to Sam are essential to the exchange of ideas. Sessions at both events were interesting and thought provoking, and I am glad to be part of these conversations. But I do want to ask what you think will be the primary driver?
The Box narrative describes a world in which bigger is better, and barriers of time, distance, and cost thought to be unbreakable are soon surpassed by the new economics of even larger tankers. The evolution of these superships does not follow a path as smooth as that described by data centers. But the lessons of size and scale can be examined. Does complexity and economics mean that data center containers and modules will replace stick built facilities? Will data center ownership devolve into just a few hands, who build larger and larger data centers?
Or will the cloud and latency issues press IT into more dispersed facilities, which could be small and easily managed? Perhaps the supersize data centers being built today are dinosaurs because of their huge energy and water demands?
What if Moore’s Law finally fails? And does it matter if Moore’s Law fails because a future generation of chips uses more or less power with the increased concentration of processing power than Moore’s Law predicts.
Increased growth in data storage could be a factor that drives technological innovation in data centers. Oddly, though, at the end of the day, many of these considerations relate to the industry’s inability to forecast IT capacity, which has been a problem since day one.