Are you driving your Data Center forward? Looking at the road ahead rather than in the rear view mirror…..
I have over the past few years been exposed to several modeling tools in the Data Center industry. Coming from an aeronautics background, modeling was a given and always used in the decision making process. However, I was surprised to find that the Data Center Industry still lagged behind in this respect, though I think it has made considerable progress over the past few years. Of the few I have come across, namely, Data Center CFD packages, pipework modelling tools, and building thermal simulation software, there is only one tool that considers energy performance and resilience across the entire M&E chain.
All these tools are useful, but they have a time and a place. No doubt they are all costly prospects for any CFO, but as designers, contractors and end users, we do not need all of them, all of the time. CFD modeling has made great inroads, and I have been an advocate of this for a while. But as far as designers and contractors go, once we move towards air flow containment then it really becomes a tool for the end user to adopt in optimally locating their servers within the data hall. I don’t mean to knock CFD modeling, but often it gives a lot of information about only a fraction of our estate, and as far as energy savings go, only the CRACs or AHUs are considered.
What we really need is for people to start using Data Center system predictive modeling tools that consider the upstream and downstream effect of M&E plant within the facility; the data hall IT and CRAC units being only a part of this chain. Remember airflow containment may allow you to recoup savings from reduced fan power, but also more prominent are the primary cooling system operating parameters along with reducing the Chiller compressor work. Once you have a systems model of your facility then it is easy to forecast IT growth, plant deployment, the effect of power cost, location, resilience, efficiency; all at a click of a mouse button.
My belief is that all these tools fall under DCPM (Data Center Predictive Modeling) and will become more and more important, as it is the only way of forecasting and informing you where you will be with your decisions, as opposed to the widely used DCIM tools that tell you what has already happened.
Ehsaan Farsimadan is a Mission Critical Consultant at Romonet Ltd, and has a PhD in Mechanical Engineering.