K.I.S.S., Einstein and data centers

It is pretty normal, practically the rule, to find data center projects every day that, instead of meeting the needs of the business, meet the needs of what we could call an engineering ego. Such an ego has a purpose other than the continuity of the business: it wants to be the one that builds the most complex and imposing data center.

We have forgotten that throughout the years, experience has shown us that the art of engineering lies in simplicity. In a world of engineers, where most of the work is limited by norms, standards, good practices, calculations, and models, it is important not to lose sight of a principle we must follow, as said by Einstein: “Everything should be as simple as possible, but not simpler.”

The reason why I am referring to simplicity as an art in engineering is because this is the hardest to attain. In our search for high availability, the lack of simplicity is leading us to build more complex infrastructures, which means greater complexity in the operations.

The question is not whether a data center will have a problem, if the equipment will fail, or if its availability will be interrupted, because we know with all certainty that at some point of its life cycle an incident will occur that could interrupt the continuity of the business. The question is when will this happen, and if we are ready to deal with it. This is the point where large and complex infrastructures created to backup IT systems cease to be important—and simple infrastructures that are easy to operate and recover become essential.

It is important to be clear about something: many times, due to the lack of proper knowledge, topologies are designed to comply with norms and standards (TIER, for example) that turn away from the essence of these good practices by “over specifying” infrastructure and trying to increase availability. These objectives are not really being accomplished; instead, they are causing problems with future operations. Clearly, a specialist with the correct design can generate the synergy between what’s required by TIER norms, the necessary availability, and simplicity.

We should always remember the KISS principle, which states that systems achieve optimum operation if they remain simple. This principle stands out because of its clear message: “Keep It Simple, Stupid.” If we do not understand the essence of Einstein’s quote, then we can easily understand this direct message.

When leading a design project, it is important to always question if we are really creating a design based on what we need. To what extent is it important to increase the complexity of the system unnecessarily? Even if it is true that as engineers in an environment of constant innovation we are expected to be one step ahead—to think “outside the box” and implement new technologies whenever possible—it is also our obligation to be knowledgeable enough to identify the limits of a design, and to remember that data centers are a tool to help us reach a goal, not the goal itself.

The challenge lies in giving up the idea that data centers are simple electromechanical infrastructures and that they are all the same. When you think this way, you are undoubtedly getting to the point of professional obsolescence, in which “if all you have is a hammer, everything looks like a nail” (Abraham Maslow). Additionally, we forget the art principle of engineering, which is simplicity.

Therefore, it is essential for us to develop methods, objectives, and design criteria with simplicity as the basic principle, thus avoiding unnecessary complexity. Let’s remember that “the mind is like a parachute: it only works when it’s open.”

By: Octavio Delgado