Energy Consumption Minimization in Data Centers for Cloud-RAN

Data centers are to become a vital part of mobilenetworks when exploring new architectures such as cloud radioaccess network, which is seen as a more energy efficient alternative to today’s mobile network installations. However, the scaleof a data center affects its energy efficiency, with larger datacenters having more resources for energy-saving measures butat the same time different challenges than those faced by datacenters of smaller scale. Hence, this work analyzes and comparesenergy efficiency of small, medium, and large data centers, andexplores the energy minimization opportunities for cloud radioaccess network data centers. When moving processing from theradio access network to a data center to save energy in the radioaccess network, the mobile network operator does not want tomove the energy consumption from the radio access network tothe data center. Hence, energy consumption must be reduced inall segments of the network, and not moved to another segment.In the case study provided, cloud radio access network datacenters are categorized as mid-scale and small-scale dependingon the size of area they cover due to strict latency requirementsof mobile network traffic flow. Thus, even though larger DCsprove to be more energy efficient, other factors, which in thecase of mobile networks will be latency, will impact the size ofthe data center. Hence, this is a major driver to consider howenergy consumption can be minimized in small data centers.