DAFNA: Decentralized Auction based Fog Node Allocation in 5G Era

Abstract

IoT applications such as smart city, smart healthcare, m-learning, m-tourism, augmented reality etc. require distributed edge computing technologies such as fog computing, mobile edge computing (MEC), cloudlet etc., to handle delay-sensitive applications. These applications are not feasible without 5G technologies because they need high-speed communication and the number of their users is also very large. In fog computing, fog nodes are allocated to the IoT users for running their heavy jobs. Many variants of a centralized reverse auction are reported in the literature for allocation of fog nodes. However, these auction methods have some major drawbacks such as one-point failure, utilization of only winner’s resources in a single round of auction, and a large number of message exchanges among participating entities in auction. Therefore, in this work, we have proposed a Decentralized Auction based Fog Node Allocation mechanism (DAFNA) that inherits all the properties of a traditional auction as well as improves the utilization of fog resources. In DAFNA, the concept of contention window and waiting time is used that makes the auction decentralized in nature. DAFNA allows fog resource providers to participate in more than one auction simultaneously for the same resource that improves the utilization of the resource. In addition, DAFNA also reduces the number of message exchanges among participating entities in auction in comparison with traditional auction. A case study is given to show the working of the proposed auction. The proposed auction mechanism is an application of 5G because it is decentralized in nature and requires speedy communication among participating entities.