Hardware-Performance-Counters-based anomaly detection in massively deployed smart industrial devices

摘要: Energy providers are massively deploying devices to manage distributed resources or equipment. These used for example the energy of smart factories efficiently monitor infrastructure smart-grids. By design, they typically exhibit homogeneous behavior, with similar software and hardware architecture. Unfortunately, these also interest attackers aiming develop botnets compromise companies' security. This paper presents a new protection approach based on Hardware Performance Counters (HPC) detect anomalies in deployed devices. HPC processed using outlier detection algorithms. Compared existing solutions, we propose lightweight comparative analysis devices' without relying modeling applications running To assess relevance effectiveness approach, thorough experimental is carried out representative industrial-type environment, sampling data from 100 Raspberry Pi simulate about 10,000 simultaneously. The results show high performance efficiency under different profiles attack payloads. Moreover, calibration depends primarily rather than application It should ease its deployment an operational environment.