Every datacenter has an invisible ceiling that constrain the quantity of IT infrastructure (servers, storage, network devices) you can place at a data center site.
High density computing is being driven by the need to maximise the use of valuable data center floor space. This concentrated computing power causes concentrated heat generation, and data centers must ensure adequate localised cooling capacity to match non-uniformly distributed heat loads. The power and cooling capacity of a data center are already largely set by the original MEP (Mechanical Electrical Plumbing) design and data center location choice. Power, cooling and floor space together form an invisible threshold that leads to the concept of stranded capacity. Stranded capacity is capacity that cannot be utilized by IT loads due to the design or configuration of the system. It is very costly that data centers can’t meet the operational and capacity requirements of their initial design and it is a hurdle to get your data center green.
At the Data Centers Europe 2011 event in Nice, Matt Warner from 6SigmaDC stated that proactive airflow management in data center operation, by using CFD simulation, could reclaim a considerable part of the stranded capacity and by this extend the life of a data center.
According to Matt Warner the problem is that each IT device has its own airflow and heat characteristics. Any IT deployed in a data center will disrupt the airflow and cooling even in empty zones of the room. And therefore each IT device has the potential to effect the resilience of every other device in the rack. The stacking of the IT equipment and the physical placement of the IT equipment in the room determines the thermal resilience. Some rack configurations or placements will cause hot spots or even overheating. These thermal hotspots will occur before the data center reaches capacity. So if your data center is running at 40% of design load, you do not have 60% capacity left! A typical reaction to overheated IT devices is by blocking gaps in the cabinet, by restacking the cabinet, or to reduce the cooling set points in the area of the devices. These are mostly expensive IT operations and also reduces the efficiency of the ACUs and reduces cooling capacity and increases energy costs of chiller units. With an example it was shown that using CFD simulation could halve the stranded capacity for a mixed infrastructure environment.
A strong case was made that to maintain a highly complex system such as a data center is, simulation software is needed to make underpinned decisions and prevent wrong decisions that will lead to more stranded capacity and will diminish the lifespan of a datacenter. And we can all agree on Matt Warner’s final statement “The greenest data center is the one you don’t need to build.”