This tutorial session will introduce the basic concepts of Data Center modeling, which can help to estimate and predict key performance indicators. Mission critical facilities present a high coupling between IT equipment and their physical environment. The energy consumed in a data center is related to the workload processed, affecting the physical environment during operation. If a data center does not process any workload, it will consume a fixed amount of energy in order to maintain the availability of all the required resources waiting for use. As workload increases, the power and airflow required increases, reaching a maximum, where the time required for processing the workload may also increase. The model serves as the basis to develop simulations and predictions to evaluate data centers under different scenarios. Parameters such as workloads and IT equipment specifications are used to estimate quality of service, power, airflow, energy, and key performance indicators through the model. The formulation describes the different parameters and simplifications considered. The attendees will be shown how the results help to assess strategies for end-to-end resource management and key performance indicators improvement. Examples will be presented to develop an understanding of the data center modeling and simulations processes and to draw practical conclusions.
Moises Levy (bio)
Director, Grupo Tecnosel
PhD(c), Florida Atlantic University