Techno-Economic Study on The Alternative Power and Cooling Systems Design for Cost & Energy-Efficient Edge Cloud Data Center(s)

Sammanfattning: The 5G technology has enabled performance-sensitive applications with low latency and high bandwidth requirements, which has put more low latency requirements on computing services. To answer this need, a small-scale data center called edge cloud is predicted to grow fast in the future. Due to its nature of being close to the end-users, the growth of edge clouds in the populated area may cause a problem with the existing power system. Besides this power system challenge, the edge cloud also requires a higher resource cost than the hyper-scale data center because of the economies of scale. In this thesis, four viable alternative power and cooling technologies are introduced to address those challenges. These four technologies are solar PV, Vertical Axis Wind Turbine (VAWT), Rear Door Heat Exchanger (RDHx), and immersion cooling. Detailed data of edge cloud are required to understand the contribution of these four technologies. However, due to the infancy state of edge cloud, those data are unavailable, and assumptions regarding data are made. Besides that, a cost model for an edge cloud is also required to show how significant the contribution of those alternative technologies is if compared to the total cost of ownership. In this thesis, the cost model for the edge cloud is extended for the alternative power and cooling system scenarios. Along with the assumed data of an edge cloud, sensitivity analysis is performed to determine whether the alternative power and cooling technologies can bring down the cost of edge cloud resources or not. Through the cost modeling, it was found out that VAWT and immersion cooling is not feasible for the particular assumed data center. On the other hand, solar PV can save 4.55% of data center electricity consumption (equal to 0.21% reduction of the total expense when calculated using the current electricity price). Furthermore, RDHx performed better with 22.73% of data center electricity expenses (equivalent to 8.35% of saving from total cost when calculated using the current electricity price).