| Abstract |
As sensing technologies become more prevalent, it has become increasingly common for edge and Internet of Things (IoT) devices to produce and consume vast quantities of data. Unfortunately, the massive increase in data sharing has far outpaced the growth of networking capabilities, thus necessitating a shift away from the classic client/server relationship to distributed systems. Past works have helped to address this problem though the use of fog computing, however, such solutions continue to suffer from the server becoming a bottleneck. In this paper, we develop EdgeCache, an edge computing system that employs game theory to optimally process and cache data by leveraging the computational, storage, and networking capabilities of devices that participate in the system. We implement EdgeCache in a crowd video sharing application that entails scalability, Quality of Experience (QoE), and time-sensitivity constraints. Our results demonstrate that EdgeCache is capable of delivering superior performance while consuming fewer network resources compared to state-of-the-art baselines. © 2019 IEEE. |
