The PUE index was basically a data center power efficiency index introduced for the classification of data centers by the Green Grid, an international consortium of data center assessment. The PUE shows the total energy consumption of a data center compared to the built-in IT performance unit.The PUE indicator, based on former calculations, took into consideration the electrical capacity measured in the data center, but it is a question whether its examination is suf¬ficient regarding today’s technological development. It can easily happen that the technological possibilities have “out¬grown” this indicator regarding server systems, renewable energy, and regardig energy recovery from waste as well. In terms of today’s technological development, it is worth exploring the energy requirements of other, newly appear¬ing cooling systems besides the still commonly used “traditional” chiller systems requiring major electrical power supply, according to the type of energy used as well. Newly spreading technologies, such as refrigerators operating based on the principle of evaporative cooling, do not cool the warm refrigerant based on the formerly used principle, thus, using electricity, but they utilize the energy of water instead. Moreover, with the latest solutions, the refrigerant is not necessarily water any more, but air. Of course, the characteristics of the data center and the examination of the IT equipment to be installed have to be taken into consideration during technological studies, since the technology to be used may vary greatly depending on the power density of the equipment installed, which is dependent on the available space and the nature of the use as well.
Considering energy consumption, the most interesting factor is not necessarily the amount of total energy consumption alone, but also how much of this energy is produced locally. It is noteworthy that electricity cannot only be taken off the network, but from various alternative sources as well, that is, a data center can produce it for itself. Such production facilities are solar panels, which may constitute the covering of the building, or fuel cells, the utilization of wind energy, biomass and biogas in attached power plants. When talking of actual energy efficiency, these forms of self-produced energy are also worthy of taking into account.
It is also worth considering what can be utilized of the heat energy generated by the IT assets located in the data center which, until now, were simply distracted and released (into the air, for example). While cooling the IT equipment of today’s data centers, particularly when using liquid-based cooling, refrigerants are getting closer and closer to the IT element to be cooled (cabinet cooling, rack/row cooling, direct processor cooling). Therefore, it is becoming more and more possible to use a higher temperature differential where the retrieved “output” heat will become exploitable as thermal energy supply for other systems related to or independent of the data center. It could be performed with a heat pump or some other heat exchanger solution. Based on the above, the introduction of an energy efficiency index scheme should be considered that takes into account the energy sources (electricity, water, gas, etc.) used by the the data center as well, and which also shows the amount of energy produced by the data center for itself, along with the amount of energy that the data center can pass on to other facilities for future use. The GPUE indicator, that sets the direction for complex measurements, ranks energy sources based on their CO2 emission used per unit of electrical energy production (CO2/kWh), establishing a guideline for the measurement of non-sustainable, sustainable and renewable energy sources. However, this figure still does not take into account what happens if no electricity is used, or how to include in the efficiency of the data center the heat energy produced there, but passed on.
Another very important element of data center indicators is comparability, where the solution to the complex problem of the given energy efficiency indicator needs to be comple¬mented by a climatic and geographical index that takes into account the endowments of the data center location. For local conditions represent not only the infrastructural facil¬ities. They also show what technologies can be considered at all for use at a specific climate and geographic location, and what is the maximum energy efficiency that is realisti¬cally achievable regarding the location.
Author: KÁROLY NAGY
Project Manager, Persecutor Ltd.