– Bob Romero, senior director of information technology at Johns Hopkins University (www.hopkinsmedicine.org) , says:
Perhaps the proper term is Data “Explosion”. All aspects of data generation and storage are on the rise with digital imaging leading the way. It is not uncommon to see certain data types grow 30%-50% per year requiring data centers to house larger storage solutions. The multiplier effect for backups and testing representations of this data compound the problem in both disk and tape solutions. The net result is a substantial footprint for storage devices and corresponding power consumption for these elements. Data de-duplication will have a profound impact on stemming the affects of the data explosion in both primary and secondary storage.
Johns Hopkins, as with most Academic Medical Centers, has embarked on cost containment and energy conservation initiatives over the past several years. Substantial improvements have been noted from the deployment of server virtualization technologies that have reduced the capital and operating costs of large fleets of servers. The same approach is being investigated for similar savings in the client workstation/desktop arena throughout the Hospital. The direct impact to the data center will be the backend servers that house the virtual clients that will be deployed. We project a sizeable deployment of servers to host these virtual clients.
Most data centers are facing the real limitations of power availability and increasing costs. The growth of server farms and the associated storage growth as outlined above continue to demand more power that is at a premium in some geographies. Major capital investments must be made to upgrade power and cooling plants to house these systems. The combination of points made above only serve to increase the power consumption in the data center with additional application and infrastructure servers. Further virtualization and consolidation of computing platforms on more efficient technologies will continue to be the focus of containment to flatten the growth curve in many areas of the data center.