Energy and Latency-aware Resource Reconfiguration in Fog Environments

With the increase of small devices with networking capabilities, Fog Computing emerged as a paradigm to improve Quality of Service. Yet, since these devices are resource constrained, there is a need to offload computational services and to maintain connectivity while moving. In this work, we aim to improve users experience by creating a new path after mobility occurs and the connection to their resources cannot be used anymore. We propose a MILP formulation and a heuristic to minimize energy consumption and latency to obtain a tradeoff solution between these objectives. These approaches find a single path composed by migration of resources to a new device and data communication path to that device. We compare both approaches, formulation and heuristic, and a baseline first-fit approach on randomly generated topologies, resources and users. The results show that the heuristic obtains solutions that are closer to the optimal, while rejecting less users and using the same number of hops than the baseline.