Computational Offloading in Fog Computing An approach for scheduling Internet of Things workflows in Cloud and Fog hybrid infrastructure

Abstract

Information and communication technology has witnessed unprecedented changes with the newlineintroduction of the Internet of Things(IoT). However, the implementation of IoT applications newlineis mainly dependant on Cloud services like compute, storage, networking, etc. Fog computing newlinehas been introduced as a complement to the Cloud infrastructure because in the near future newlinedemands of the IoT devices will exceed the capabilities of Cloud computing. Moreover, newlinemany IoT applications require services that Cloud cannot provide or the services provided newlinemay not be optimal. This work deliberates on how Fog computing is introduced as a layer newlinebetween the Cloud and the IoT applications so as to fulfil the demand for various services of newlineIoT applications. The Cloud-Fog hybrid infrastructure provides flexibility to the IoT newlineapplications to utilize the services of Fog devices or Cloud data centers as per the newlinerequirement of the IoT applications. newlineThe introduction of the Fog computing layer introduces various challenges and this research newlineis focussed on one of the important challenges that is computational offloading. This work newlinediscusses how application modules can be distributed among the Fog devices and the Cloud newlinedata centers and proposes multiple computational offloading algorithms which try to optimize newlinethe QoS parameters. But the optimization of one QoS parameters may be achieved at the newlineexpense of the degradation of other QoS parameter. The proposed computational offloading newlinealgorithm combines the four performance parameters and optimizes the results of all the newlineparameters. This algorithm is the first of its kind which focuses on four performance newlineparameters together. newlineA mathematical model has been proposed to measure the performance parameters which have newlinebeen used by the computational offloading algorithms and to evaluate the performance of newlinethese computational algorithms, a simulator named iFogSim has been used. The simulator has newlinebeen modified to suit the requirements of this work. The algorithms use four performance newlineparam