Distributed User-centric Service Migration for Edge-Enabled Networks

Edge-enabled scenarios are composed of all sorts of heterogeneous devices, networks, and services. Such devices can cooperate in providing ubiquitous service availability for users in both fixed and mobile networks. A multi-tier edge-enabled network is a suitable scheme for provisioning cloud services ubiquitously with QoS support. However, in mobile scenarios, the network may need to migrate services and applications to follow user mobility and avoid disruptions, posing a great challenge for the network and cloud mobility management solution. This paper proposes an intelligent service migration and resource management algorithm for intra- and inter-tier communication in edge-enabled environments, called Migration On Edge-Enabled Scenarios (MOES). MOES considers mobility prediction, service requirements, and monetary costs to perform an intelligent service migration decision. MOES also manages service migrations as devices leave the coverage area of their networks. MOES was implemented in a realistic edge-enabled heterogeneous scenario in which simulation results show its efficiency in delivering the necessary service requirements by up to 72% in terms of latency, at a cost 58% lower in terms of the number of migrations compared to state-of-the-art algorithms.