DCIM and CFD can be a powerful tool when used together, though CFD systems pose issues of in-house training and decisions about the type and sophistication of the modeling software that will be purchased.
The percentage of data centers that use predictive simulation is small, Wade says, and many people do not understand the value of adding CFD to DCIM.
"For example, DCIM cannot tell you if you have issues with air bypass or recirculation. These are imbalances between supply and demand with poor air management," Wade says. "But the two together, DCIM and CFD, will be able to tell you about that."
CFD systems are installed by outside providers who can then train people inside the organization on how to manage the modeling tools. The capacity management team, usually a few people in an organization "who track and trend power, space, and cooling in the data center, are many times responsible for the installation of hardware," says Wade. The people running these systems have specialized expertise, he says.
There is a cost to keeping a calibrated model running, but the benefit is huge, in terms of creating efficiencies and ensuring uptime, say sources. Once the model is installed, people in-house can be trained to update the model. Using CFD modeling has an added benefit of forcing facilities and IT people into a common framework and language when considering various scenarios, Koomey says. For example, if IT says it wants 100 servers, the facility manager can turn around and say, "Let's put that in the model to see if we have capacity."
"This facilitates conversations between parts of the organization that do not talk to each other well," Koomey says.
Different types of modeling software have plusses and minuses associated with what they can do, explains Evanko. "There are various levels of software and sophistication associated with running these models. The more robust software tends to be more expensive but more useful in [creating] true energy efficiency," he says. Such software also requires a higher level of input and sophistication on the part of the end-user as well.
Server manufacturers are building equipment that can run at higher temperatures, which also drives a need for more sophisticated modeling — so that the data center manager can know exactly what the temperatures are near the computers. "You don't want to run at 68 degrees when you can go to 80 degrees," Evanko says. "But you have to have more precision with temperature, air flow, and static pressure." Those with more sophistication and the modeling tools to go with it can be more aggressive in pursuing energy efficiency, he says.
Maryellen Lo Bosco is a freelance writer who covers facility management and technology. She is a contributing editor for Building Operating Management.
CFD Modeling Can Help Maximize Data Center Capacity
CFD’s 3-D Detail Can Help Avoid Stranded Capacity
CFD Poses Issues of In-House Training, Software Purchases
