Faster exploration of data centre cooling using thermal influence indices

Abstract

Cooling is an important issue in data centre design and operation. Accurate evaluation of a design or operational parameter choice for cooling is difficult as it requires several runs of computationally intensive computational fluid dynamics (CFD) based models. Therefore there is need for an exploration method that does not incur enormous computation. In addition, the exploration should also provide insights that enable informed decision making. Given these twin goals of reduced computation and improved insights, we present a novel approach to data centre cooling exploration. The key idea is to do a local search around the current design or operation of a data centre to obtain better design or operation parameters subject to the desired constraints. To do this, all the microscopic information about airflow and temperature in data centre available from a single run of CFD computation is converted into macroscopic metrics called influence indices. The influence indices, which characterize the causal relationship between heat sources and sinks, are used to refine the design or operation of the data centre either manually or programmatically. New designs are evaluated with further CFD runs to compute new influence indices and the process is repeated to yield improved design or operation as per the computation budget available. We have carried out guided explorations of design and then operation of a realistic data centre using this methodology. Specifically, we considered maximization of the heat load (design parameter) and supply temperatures of CRAC (operating parameter) in the data centre subject to the constraints that: (1) servers are kept at appropriate temperatures and (2) overloading of CRACs is avoided. Our evaluation for a 1500sq.ft. data centre shows that the use of influence indices cuts down the exploration time for design by 80% as compared to unguided explorations. It is demonstrated quantitatively that this solution is close to optimal. A guideline was evolved to reduce the effect of initial configuration on the final solution. Proposed methodology provides an additional 10% reduction in operational cost over existing methods of unguided explorations.