We’re all familiar with phrase “lost in translation”. Since words in one language may not have direct equivalents, or even mean something entirely different, in another, they sometimes have undesirable connotations for their intended audience. For example, in some Asian countries Kentucky Fried Chicken’s “Finger Lickin’ Good” slogan translated into “Eat your fingers off” which was a bit of a barrier to successful market entry, and you probably won’t find Puff’s tissues on drugstore shelves the next time you visit Germany since the term is a colloquialism for a house of ill-repute. If this has some of you wondering about those Chinese characters that you had tattooed on your arm, my advice would be to avoid wearing sleeveless shirts next time you’re in downtown Beijing. The converse of these incidences of international misunderstandings, are words and phrases that maybe do a better job of expressing something than our own native language. To illustrate this concept, I give you the Japanese phrase, “Poka Yoke”, that describes any mechanism that helps an equipment operator avoid mistakes.
Despite the broad spectrum of bipedal occupants of the planet, “someone screwed up” is a universal concept. As I’ve said before, when it comes to data center outages the logical forensic line of inquiry should focus on “who”, not “what”. What unites us across the globe in general, and the data center business specifically, is how do we effectively deliver our own Poka Yoke?
I would argue that although we’ve made tremendous strides in our ability to build facilities faster, make them more energy efficient and maybe even manage them better—I always like to throw a bone to the DCIM guys—we don’t do a very good job of turning them over to the end user. Think about it. Since it’s estimated that approximately 70% of data center outages can be attributed to human error, we can probably do a little better when it comes to the old documentation and training department. This by no means that anyone wants bad things to happen, but if the owner’s manual for your car is more thorough than your SOPs you might want to consider giving technicians a little bit more to work with. It’s probably also a good idea to think about the packaging of said information. While it’s nice to have everything written down in one place, trying to refer to a three-ring binder, thick enough to hold the collected works of Shakespeare (including the sonnets) while attempting figure out why a PDU isn’t working is going to reduce the effectiveness of even the best of technicians.
Certainly, it seems like achieving an enhanced level of Poka Yoke is on everyone’s minds and organizations are taking different paths to reduce the potential for human induced catastrophe. Some have focused on the standardization of hardware platforms to simplify operations, while others have homed in on the people performing the work. For example, with our partner, Icarus Ops, we’ve worked to provide end users with detailed checklists, similar to those used in areas like flight operations, that are accessible from either a tablet, Android or X Glass to provide them with an easily accessible reference tool.
As they say, no one is perfect and, unfortunately, this maxim is the biggest threat to data center operations. Reducing the potential for human error in the data center is an on-going process, but, at least we have a name for it—in any language.