Not too long ago, Gartner made a splash with the research headline, “The Data Center is Dead”. The point wasn’t that enterprise data centers were suddenly obsolete, but that the days of a traditional enterprise-owned data center serving as the sole location to source IT services were waning. With a growing reliance on externally provided services, from cloud providers, application hosting providers, or workloads housed in colocation centers, more and more enterprises were likely to hold the line, or actually reduce their footprint in the enterprise data center. This is a far cry from shuttering all enterprise data centers, as not all applications are ideally placed in the cloud at this time, and constraints imposed by the “three P’s” (legacy, latency, and legal) maintain the relevance and value of an on-prem estate. So the questions naturally arise, “what do we do with our data centers?” “Should we continue to update them?” “How should we plan for this?”
In his research note titled “How to Turn Old Data Centers Into Critical IT Assets”, David Cappuccio addresses the enterprise need to have a strategy regarding older data centers, and argues that they can, and at times should be updated to support new and emerging business services while reducing operating costs. Given that a percentage of applications will remain “on-prem” for the foreseeable future, it turns out that a wave of optimization can improve efficiency and operations to an extent that it is worth at least one more cycle of updates and upgrades.
He describes three practices I&O leaders should consider, namely enhancing delivery, maximizing space, and reinventing the fundamentals of data center design. Enhancing delivery describes using such designs as self contained rack clusters to introduce more efficient sub-units of highly dense data center footprint, with their own self-contained power and cooling infrastructure. Maximizing space focuses on such techniques as hot aisle containment to improve efficiency and free up additional space. Lastly, reinvention tackles the status quo though the use of new technologies such as rear door heat exchangers and liquid cooling to increase density while reducing the power requirements of cooling IT loads. In short, they are all about increasing efficiency in a modular fashion, allowing for incremental upgrades rather than a very costly complete overhaul. Bringing modern technology and design patterns to incrementally improve the data center’s ability to handle new and emerging workloads can not only save money, but help to avoid building new data centers by increasing the capacity of the old ones.
While the trend towards use of externally provided services such as those from cloud continues, there are workloads that need to remain on-premise. Using the strategy and tactics highlighted in this research note, enterprises have another efficient option than just moving everything to a combination of cloud and colocation.