An Incentive-compatible Offloading Mechanism in Fog-Cloud Environments Using Second-price Sealed-bid Auction
Description
In a fog-cloud environment, the tasks submitted by end-users are first sent to intermediate nodes called fog nodes. If the computing resources in fog nodes are insufficient, the nodes will offload tasks to the remote cloud. Naturally, intermediate nodes are reluctant to offload tasks to upstream entities, which in turn results in a degradation in network performance. To motivate such reluctant nodes, some previous researches have used game theoretical approaches. We believe that auction theory is one of the most important mathematical tools to motivate fog nodes to participate in offloading operations. In this paper, we propose a second-price sealed-bid auction mechanism for offloading optimization. In our model, the service unit plays the role of the commodity. Also, fog nodes and cloud datacenter play the role of bidders and auctioneers, respectively. We prove that the proposed auction mechanism has two important properties of incentive compatibility (IC) and incentive rationality (IR). Since network actors are not necessarily truthful, the existence of these two properties is crucial. We compare the proposed mechanism with state-of-the-art methods. Experimental evaluations using the iFogSim simulator indicated that the proposed method is much better than baseline methods in terms of significant criteria such as execution time, energy consumption, and network usage.
Files
Steps to reproduce
Please read he "README" file.