Game Theoretic Energy Allocation for Renewable Powered In-Situ Server Systems

Junlong Zhou, Jianfei Chen, Kun Cao, Tongquan Wei, Mingsong Chen

Abstract: In-situ server systems are deployed in very special operating environment to handle in-situ workloads that are normally generated from environmentally sensitive areas or remote places that lack established utility infrastructure. This very special operating environment of in-situ servers urges such systems to be 100 percent powered by renewable energy. However, existing energy management schemes assume a hybrid supply of grid and renewable energy, hence are not well suited for 100 percent renewable powered in-situ server systems. In this paper, we tackle the problem of allocating harvested energy to 100 percent renewable powered server systems for optimizing both the overall system throughput and throughput of individual servers. From a game theoretic perspective, we model the energy allocation problem as a cooperative game among multiple servers and derive a Nash bargaining solution. Based on the Nash bargaining solution, we then propose a heuristic algorithm that determines the energy allocation strategies according to system energy states. Experimental results show that our proposed game theoretic approach achieves a high throughput from perspectives of both the overall system and individual servers.