Deadline Aware Data Offloading in Fog Computing

Abstract

Fog computing is an extension of cloud computing where services are provided at the edge of a network. With the growth of the Internet of Things (IoT), different applications have emerged. Many of these applications such as the Intelligent Transportation system (ITS), Tactile Internet applications (telerobot and robot surgery) and healthcare emergency application are delay-sensitive and need high reliability. Fog computing becomes an essential and integral part of cloud computing to provide a better quality of service for these applications to fulfill latency requirements and needed reliability. Under the fog computing environment, different resource allocation, task scheduling algorithms, and task offloading techniques are proposed to manage applications. These deployed algorithms and techniques enhance resource utilization, power consumption and overall response time fully or partially. But it is still difficult to accomplish the required latency requirement of applications with these algorithms. To solve this problem, with the use of a design science approach a deadline-aware data offloading model as well as algorithm is proposed and implemented to fog nodes. The proposed algorithm considers the deadline of a task, computation power of fog nodes, task nature and transmission delay during scheduling. After the evaluation of these parameters, the node controller offloads the task at the appropriate fog node. The evaluation of the proposed algorithm is done with a simulator. The effectiveness of the proposed algorithm is evaluated and compared with the existing first come first served (FCFS) and dynamic task offloading algorithms. According to the literature these algorithms are found to be the best algorithms from other scheduling algorithms. The experimental results show that the proposed algorithm achieves 24% and 17% less overall delay time when compared to FCFS and Dynamic Offloading algorithms respectively.