Optimizing the Computational Offloading Decision in Cloud-Fog Environment

Internet of Things is set to revolutionize our lives and it has already penetrated every sphere of our lives. However, IoT applications are in need of various resources like compute, storage, networking, etc. IoT devices being resource- constrained are mainly dependent on Cloud services for their resource requirements. But as the number of IoT devices is increasing rapidly, Cloud resources are not growing proportionally and in near future demands of IoT devices will exceed the capabilities of the Cloud resources. Keeping this shortcoming of Cloud into mind, Fog computing was introduced as a complementary system paradigm for Cloud computing. Fog computing uses the resources of the geographically closer devices to serve the demands of IoT devices. Our work focuses on the efficient utilization of the Cloud-Fog resources by distributing the application modules among Fog devices and cloud data centers. Placing the application modules on Fog devices improves performance parameters like response time, latency, energy consumption, etc. We have proposed a load balancing algorithm whose performance has been evaluated on the iFogSim simulator and has achieved a 21% reduction in network consumption when compared to the default policy of iFogSim. Our approach is generic which can be used in the majority of IoT applications.