NEW METRICS FOR THE MEASUREMENT OF ENERGY EFFICIENCY IN DATA CENTERS

ABSTRACT

The last decades witnessed an explosive increase of the demand for digital data in the economy. This led to a significant growth in the number and the sizes of data centers.
The energy used by the data centers has increased dramatically. This situation determined the quest for metrics that measures energy efficiency in data centers.
A classical metric for measuring energy efficiency in data centers is PUE (Power Usage Effectiveness). It was introduced in 2006 by C. Malone and Christian Belady. It is defined as the ratio between the PDC (total power used in the data center) and PIT (the power used for the IT equipment). PDC is the energy requirement for the major IT components of the data center plus the energy for supporting the infrastructure. The infrastructure is composed by: power delivery components, cooling system components and other miscellaneous component loads, such as data center lighting. PIT is the energy requirements associated with all the major IT components of the data center, such as server, storage, network, and other devices (KVM switches, monitors, and workstations/laptops) used to monitor or otherwise control the data center. PUE is a very popular metric for measuring energy efficiency in data centers. It has advantages and disadvantages.
The present paper proposes an analysis of new metrics for measuring energy efficiency in data centers. This thing is necessary in order to attain sustainability goals.
A globally accepted metric on the energy efficiency of computer servers in business applications is still missing, despite a great interest from stakeholders.
Some complementary metrics and new metrics for improving PUE are synthesized. This analysis can help researchers to identify the main trends of energy efficiency assessment in data centers. The paper provides also a comprehensive and critical survey of energy efficiency metrics in data centers. It motivates the quest for new metrics for the assessment of data centers energy efficiency.

Keywords: metrics for energy efficiency, data centers, PUE, energy