SPARSE: Spatially Aware LFI Resilient State Machine Encoding

Abstract

As finite state machines (FSMs) control the behavior of sequential circuits, they can be a target for attacks. With laser-based fault injection (LFI), an adversary may attain unauthorized access to sensitive states by altering the values of individual state flip-flops (FFs). Although standard error correction/detection techniques improve FSM resiliency, all states and FFs of an FSM are assumed equally critical to protect, incurring significant overhead. In this paper, we introduce a novel spatial vulnerability metric to aid the security analysis, which precisely manifests the susceptibility of FSM designs to LFI based on state FF sensitivity and placement. A novel encoding and spatially aware physical design framework (SPARSE) are then proposed that co-optimize the FSM encoding and state FF placement to minimize LFI susceptibility. SPARSE’s encoding uses the minimum number of FFs by placing security-sensitive FFs a sufficient distance apart from other FFs. SPARSE is demonstrated on 5 benchmarks using commercial CAD tools and outperforms other FSM encoding schemes in terms of security, area, and PDP.