An Intrusion Detection System for the OneM2M Service Layer Based on Edge Machine Learning

摘要

The number of connected Things is growing at a frantic pace, which has led to vertical, proprietary Internet of Things (IoT) solutions. To ensure a horizontal IoT cross-industry interoperability, eight of the word's leading ICT standards bodies introduce the oneM2M standard. Its main goal is to satisfy the need for a common M2M Service Layer that guarantees the communication between heterogeneous devices and applications. Various security mechanisms have been proposed in the oneM2M specifications to protect the IoT solutions. As a complementary security level, we propose the first generic Intrusion Detection System (IDS) for the oneM2M Service Layer based on Edge Machine Leaning (ML). This oneM2M-IDS can be added to the basic architecture of oneM2M or can be added as a plugin to existing systems based on oneM2M. In this work, we define and implement oneM2M attack scenarios related to the service availability. Moreover, we propose an edge IDS architecture and we detail ML features selection. The performance of the proposed IDS is studied through multiple experiments with different ML algorithms.