MobFogSim: Simulation of mobility and migration for fog computing

Carlo Puliafito,Diogo M. Gonçalves,Márcio M. Lopes,Leonardo L. Martins,Edmundo Madeira,Enzo Mingozzi,Omer Rana,Luiz F. Bittencourt
DOI: https://doi.org/10.1016/j.simpat.2019.102062
IF: 4.199
2020-05-01
Simulation Modelling Practice and Theory
Abstract:Fog computing is an extension of the cloud towards the network edge that brings resources and services of computing in closer proximity to end users. This proximity provides several benefits such as reduced latency that improves user experience. However, user mobility may limit such benefits in practice, as the distance to a fog service may vary as a user moves from one location to another. Migration of a fog service may be one possible mitigation strategy, enabling the service to always be close enough to a user. Although many simulators exist for evaluating application behaviour and performance within a fog computing environment, none allows evaluation of service migration solutions to support mobility. MobFogSim is presented in this work to overcome this limitation, it extends iFogSim to enable modelling of device mobility and service migration in fog computing. MobFogSim is validated by comparing simulation results with those obtained from a real test bed where fog services are implemented as containers. Additional experiments are carried out in MobFogSim taking account of various mobility patterns of a user, derived from Luxembourg SUMO Traffic (LuST). We use an experiment-based approach to study the impact of user mobility on container migration in fog computing.
computer science, interdisciplinary applications, software engineering
What problem does this paper attempt to address?