Green Geographically Load Balancing Using Viability Of Distributed Data Centers’ Local Condition For Energy Efficiency

Nowadays, distributed Data Centers (DCs) are very common, because of providing more energy efficiency, higher system availability as well as flexibility. In a distributed cloud, each local DC responds to the specific portion of the incoming load which distributed based on different Geographically Load Balancing (GLB) policies. Taking advantages of diversity in conditions among distributed DCs, GLB has the promise of significant monetary saving. One of the basic assessment metrics among DCs is Power Usage Effectiveness (PUE). Conventionally, PUE is considered as a constant value which is the ratio of the total power required to operate and cool the DC vs. the power drawn by the IT equipment. However, in modern DCs equipped with free cooling and Renewable Energy (RE), the cooling power (as a big portion of total power) varies with the changes in environmental conditions as well as with the dynamically changes in IT power. Thereby, PUE is not only not fixed but also varies over time as well as IT load. In addition, while green energy is increasingly used in DCs, the conventional PUE is incapable of describing the green power efficiency. This paper considers the Green PUE as a power efficiency measurement in DCs, which is a function of time as well as IT load. We show that Green PUE can effectively reflect the changes in power efficiency (including the green power efficiency) caused by changes in DC s conditions. Using Green PUE as an assessment criterion, we propose a cost-effective GLB approach. Results show that the proposed