An adaptive rank-based approach for dynamic controller selection in Fog Computing

Abstract

The deployment of a dynamic and cooperative fog control plane, where controllers are selected on-demand among the most suitable underlying resources, has been recently proposed as a Control-as-a-Service (CaaS) model. Albeit it is expected that real-time applications shall benefit from this concept, mechanisms for QoS-aware controllers election is yet an open issue. In this work, we propose an adaptive rank-based controller selection strategy that is capable of tuning the weights employed for each characteristic of interest in order to cope with the environment dynamism. Results have shown an average controller exchange reduction of 37% when compared with a preliminary approach employing fixed weights in a dynamic scenario, as well as its efficiency in battery and memory usage by controllers.