Low-Latency Task Classification and Scheduling in Fog/Cloud based Critical e-Health Applications

Abstract

5G wireless networks have been designed to provide high reliability, ultra-low latency, and support of massive amount of connected devices, in the scope of the Internet of Things (IoT). Advances in electronics and networking are enabling the wide adoption of multiple types of verticals, under the umbrella of IoT. One area of IoT, which is gaining lots of attention, is e-Health. Within e-Health, users’ monitoring in general, and monitoring of vital signs, such as heartbeat rate, are very useful in saving many lives; however, they require ultra-low latency. Cloud-based networking and computing have been proposed to achieve the required low latency. Furthermore, fog computing was proposed to further decrease the achieved latency for critical tasks and services; however, this adds more complexity to the control of the network, in addition to the task scheduling among both the cloud and fog layer. In this paper, we tackle this problem by proposing a task classification and scheduling scheme in a fog-cloud networking environment, by considering comprehensively modeled characteristics of tasks, user profile, environmental exposure and networking features, targeting the improvement of overall latency for higher priority critical tasks. Moreover, we perform an analytical study on the execution comparison between cloud and fog computing services, which paves the way to further develop an orchestrator for task scheduling, among the multi-layer fog-cloud based e-Health systems. Simulation results show that the proposed task scheduling scheme outperforms the benchmark, by guaranteeing ultra-low latency for critical tasks (high-priority), while ensuring sufficient latency performance for latency-tolerant tasks.