A Model-Based Intrusion Detection System for Smart Meters

Smart meters have gained wide adoption as an integral part of the smart grid. However, their security remains problematic as many attacks are discovered against them. Smart meters are embedded devices that are constrained in terms of computing power and memory. They are also deployed on a large scale which imposes specific requirements (e.g., no false positives) on any IDS developed for them. In this paper, we propose a model-based technique for building intrusion detection systems (IDS) for smart meters, that takes these constraints into account. We implement our IDS on an open source smart meter platform. We show that our IDS imposes little performance overhead, even under severe memory constraints, and effectively detects a wide range of both known and unknown attacks. In comparison, existing IDSs incur unacceptable performance overheads on the meter.