A truthful double auction framework for security-driven and deadline-aware task offloading in fog-cloud environment

Abstract

Emerging as an alternative to cloud computing, fog computing is expected to fulfill the resource demand of mobile users and provide low-latency and high-bandwidth services for ever-increasing Internet of Things (IoT) applications. Offloading time-critical and computation-intensive tasks to the fog can significantly reduce response time. Due to dynamicity, heterogeneity, and limited resources, fog may fail to provide services with guaranteed performance. This paper addresses the problem of task offloading and resource allocation in a fog-cloud environment for IoT applications. The proposed Truthful Double Auction Task Offloading (TDATO) framework provides incentives for fog providers to maximize their revenues, execute as many as possible requests with guaranteed performance, and maintain a high reputation. Depending on the user requirements in terms of security level and maximum latency tolerance for task execution, we propose two deadline- and security-related algorithms for winners' determination in the auction: Deadline-Minimum Security (DMS) and Deadline-Preferred Security (DPS). Moreover, two assignment algorithms are analyzed, Reputation-Based (RB) and Random Assignment (RA). The proposed double auction framework assures that all winning users' requirements in terms of deadlines and security level are satisfied while winning providers’ resources are allocated to the users that value them the most. Furthermore, the assignment mechanism aims to maximize resource utilization for fog providers with the highest reputation. We conducted a theoretical analysis to demonstrate that TDATO satisfies truthfulness, individual rationality, and budget balance and has polynomial-time computation complexity. Extensive simulation experiments are performed to evaluate performance.