Intrusion Detection-Based Ensemble Learning and Microservices for Zero Touch Networks

Abstract

The industry has undergone a digital transformation facilitated by the Industrial Internet of Things (IIoT) technology, ushering in the era of Industry 4.0. However, the widespread use of IIoT devices, such as sensors, robots, and other IIoT technologies, has made IIoT systems and associated services vulnerable to a range of network-based attacks, which can impede their performance and disrupt operations. Moreover, the automation process is essential for IoT-based industries to meet future demands. The concept of a zero-touch network has recently evolved to coordinate and manage network resources automatically. Machine learning (ML) plays a critical role in its architecture due to its ability to facilitate a close-loop automation process. With ML, analytical tasks and real-time predictions can be achieved effectively to build smart applications for the early detection of cyberattacks in IIoT systems. The ML components, such as preprocessing, training, and testing, can be subdivided into microservices to improve service while allowing interaction with edge and cloud services. In this article, we propose an ensemble learning-based intrusion detection system (IDS) for a zero-touch network automation process by leveraging ML and microservice technology to improve the trustworthiness of IIoT systems. More specifically, we use the feature selection technique to select significant features and pass them to different models, then blend their predictions based on a stacked ensemble learning approach. Experiments are conducted to evaluate the performance of the proposed framework compared to existing studies.