What exactly is DCIM?

So it goes without saying, DCIM is quite the ubiquitous term. To be even more confusing, the term “Data Center Infrastructure” is pretty open and encompassing itself. Therefore, ANYTHING that manages the infrastructure can be considered DCIM. And then to make things worse, there are a plethora of vendors out there to make all sorts of claims about compatibility with all sorts of control and monitoring platforms. It is important to understand that when we are talking about DCIM, we specify exactly what part of the entire DCIM “ecosystem” (I hate that word) we are referring to. At one end of the spectrum, there is the monitoring and control of applications, OS, hardware, individual computing hardware power, cooling, I/O rate, fan speed – then all the IT infrastructure information – and finally, matched with the facility information (power-in and heat-out) measures.

It’s not for Everyone

Even though there are many parts to the DCIM equation, it is safe to say that the main objective is to optimize the entire platform, but unfortunately, this isn’t always accomplished. Now there are those who have achieved great success and have implemented the entire suite of automation controls and monitoring. However, these are an elite few. There are also many who have tried this rigorous implementation yet have landed in the graveyard of DCIM failures. Then there are those that have successfully implemented the facility side of the DCIM equation – a Building Automation System (BMS or BAS) combined with an Electrical Monitoring and Control System (EMCS). And of course, there are also those that have implemented the IT side of DCIM. Integrating BOTH is a challenge that one must be committed to. After all, it is really an IT deployment, and the proper controls, objectives, planning, and budgets need to be developed up front.

Determining if DCIM is Right for you

So when considering DCIM implementation, here’s what you should start with: What is your objective? Is it to create a full lights out, automated facility? Or to simply get some monitoring of your facility status? Or something in between? All of these can be considered part of the overall DCIM implementation. Bottom-line: DCIM encompasses a lot. Here is the point of this writing – just because you are automating a part or all of your operation does not mean you are covered. Just like the global airlines are learning, pilots need to be able to comprehend the status and configuration of their aircraft while not on autopilot. They need to understand the basis for what the systems are telling them. In general, develop a “feel” for the geometry in any configuration. The same goes for a data center. What are the systems telling you and is your staff trained enough to understand what is going on? A simple way to determine this is to use our ACES problem solving model used for addressing trouble alerts: Can they, A – Communicate an Alert to others; C – Confirm what the problem is; E – Evaluate (determine the source of the problem); and, S – Sustain operations? In short, a High Reliability Organization will demand that they do this. Can your operation be considered a High Reliability Organization?