A Utilization-Based Approach for Schedulability Analysis in Wireless Control Systems

Abstract

Recent advancements in industrial Internet-of-Things (IoT), more specifically, the development of industrial wireless standards such as WirelessHART and ISA100, are paving the way for the fourth industrial revolution, Industry 4.0. These wireless standards specify highly reliable and real-time communications as key requirements in industrial wireless sensor-actuator networks. Schedulability analysis remains the cornerstone for analyzing the real-time performance of these networks. While it is well-explored in the domain of CPU scheduling, schedulability analysis for multi-hop wireless networks has seen little progress till date. Existing work mostly focuses on worst-case delay analysis that runs in pseudo-polynomial time, making it is less suitable under frequent network dynamics which are quite common in industrial IoT. To address this, in this paper, we develop a schedulability analysis based on utilization bound for multi-hop, multi-channel industrial wireless sensor-actuator networks. Because of its extremely low runtime overhead, utilization-based schedulability test is considered to be one of the highly efficient and effective schedulability analyses. However, no work has been done yet on utilization-based analysis for multi-hop wireless network. The key challenge for a utilization-based test for multi-hop wireless network arises from the fact that wireless network is subject to transmission conflict and network dynamics which are not present in CPU scheduling. We address this challenge by bridging the gap between wireless domain and CPU task scheduling. We have evaluated our result through simulations using TOSSIM that shows that our schedulability analysis is safe and effective in practice.