Multi-Objective Service Provisioning in Fog: A Trade-Off Between Delay and Cost Using Goal Programming

Abstract

Internet of Things (IoT) has enabled new possibilities for some newly emerged applications such as automated cars and smart city applications. By emerging IoT applications with different requirements compared to existing services, cloud computing could not satisfy these needs anymore. Fog computing was introduced in 2012 and it brought a new collaborative computing model to make the growing of IoT possible. This paper presents a multi-objective framework to find eligible fog nodes to dynamically deploy the IoT applications on them. The proposed framework could be employed to achieve a trade-off between the cost of resources and average service delay. The multi-objective dynamic service provisioning (MDSP) problem is formulated as a mixed-integer linear programming (MILP) model and the weighted goal programming is applied to solve the multi-objective problem. In addition, an evaluation of the proposed multiobjective framework against two relevant single-objective approaches is performed.