Security Frameworks and Attack Models in Smart Metering and Smart Grids: A Systematic Review of Threats and Countermeasures

Abstract

As Smart Grids evolve into highly interconnected Cyber-Physical Systems (CPS), smart meters face unprecedented security threats. While existing surveys address high-level grid security, there is a distinct lack of structured analysis regarding security frameworks operating under strict embedded hardware limits. This paper conducts a systematic review to map contemporary cyber-physical attack models against edge-deployed mitigation strategies, evaluating the architectural trade-offs between security optimization and device resource constraints. Following a PRISMA-aligned systematic literature review protocol, a rigorous multi-stage filtering process was executed across major academic databases. Peer-reviewed studies were critically screened and evaluated based on predefined criteria focusing exclusively on embedded smart meter endpoints. Aggregating empirical data from the surveyed literature, the analysis reveals a profound architectural polarization. Lightweight Cryptography (LWC) primitives demonstrate an optimized memory footprint but remain blind to semantic data manipulations. Conversely, Edge-AI anomaly detection models deliver superior intrusion detection rates (averaging ~98.4% across surveyed studies) but induce critical computational penalties, systematically spiking CPU utilization up to 85%. Furthermore, blockchain-based frameworks offer robust decentralized resilience but introduce propagation latencies that exceed real-time state estimation thresholds. This review exposes critical research gaps, including the absence of unified hardware-software security co-designs and a lack of preparation for post-quantum cryptography. Ultimately, a strategic technical roadmap is outlined, offering practical guidelines for utility providers and researchers to develop energy-efficient, quantum-resistant embedded architectures.