Data center operations: four pitfalls to avoid

Whether it involves land or technical facilities (electricity, cooling, security), companies invest heavily to build or renovate their data centers. Conversely, they frequently underestimate the resources required for proper operations, particularly the human element. Generally relying on outsourcing, from the urbanization of IT rooms to the multi-technical maintenance of the building, the operating team is often reduced to the bare minimum with a knowledgeable employee who centralizes a wide spectrum of skills, covering both IT (networks and hardware) and the building's technical infrastructure, notably electricity and cooling. Versatile, they have in-depth knowledge and the site's history. But what happens when this person, who holds all the expertise of a data center, is absent or leaves the company?

Just like equipment redundancy, which is necessary to eliminate single points of failure ("SPOF"), strengthening operations teams reduces risk for companies and constitutes a profitable long-term investment. These "five-legged sheep" profiles, capable of both mastering IT and understanding the building, are however not very easy to recruit, due in particular to the image deficit attached to the world of data center operations.

Managing a data center effectively without previously defined procedures and without an accurate history of changes and interventions is illusory. Tools, particularly BIM (Building Information Management), despite the promises they bring, face the same limits as traditional documentation: if, due to a lack of time and resources, the data center operations monitoring elements are not updated regularly, the result is the same. In fine, partial operating procedures and poorly documented interventions amount to managing the data center "by sight," reacting to emergencies and incidents, and lead to haphazard maintenance in operational conditions.

Due to its mixed nature, combining technological and building infrastructure, the data center depends on the joint responsibility of the IT department and general or real estate services. This requires, a minima, regular and effective communication between these different departments as well as a clear distribution of their responsibilities and scopes of intervention across all processes (daily maintenance, security, response protocols in the event of a malfunction or major crisis).

Almost all companies today are "IT-dependent" and cannot suffer a momentary interruption of digital services. Therefore, when such an event occurs, other company departments turn to the IT department even if the malfunction results from a problem external to IT, affecting a building element, its electrical systems, or cooling systems, for example. Establishing a department responsible for end-to-end service continuity of information systems, including data center management, constitutes a preferred governance model.

Maintaining the information system in operational conditions is no longer a simple "nice to have" but a "must have." A data center out of service, for whatever reason, means an unusable information system and therefore a business that is at best slowed down, if degraded mode operating procedures have been planned, or at worst at a standstill. The service continuity of the information system can therefore no longer be limited to IT elements. It must now integrate building operations through the monitoring of a certain number of indicators (KPIs). The development of connected objects and the widespread use of artificial intelligence should, in the very near future, help companies have even finer data to further industrialize processes and improve predictive maintenance.

In other words, it is high time to raise the operations of the infrastructure's "lower layers" (building and IT) in the company's scale of priorities—less visible but no less essential than the application layers in digital business.