Energy-aware and adaptive fog storage mechanism with data replication ruled by spatio-temporal content popularity

摘要

Data traffic demand increases at a very fast pace in edge networking environments, with strict requisites on latency and throughput. To fulfil these requirements, among others, this paper proposes a fog storage system that incorporates mobile nodes as content providers. This fog storage system has a hybrid design because it does not only bring data closer to edge consumers but, as a novelty, it also incorporates in the system other relevant functional aspects. These novel aspects are the user data demand, the energy consumption, and the node distance. In this way, the decision whether to replicate data is based on an original edge service managed by an adaptive distance metric for node clustering. The adaptive distance is evaluated from several important system parameters like, distance from consumer to the data storage location, spatio-temporal data popularity, and the autonomy of each battery-powered node. Testbed results evidence that this flexible cluster-based proposal offers a more responsive data access to consumers, reduces core traffic, and depletes in a fair way the available battery energy of edge nodes.