Isolating trust in an industrial control system-on-chip architecture

摘要

A distributed industrial control system (ICS) also distributes trust across many software and hardware components. There is a need for some malware countermeasures to be independent of application, supervisory or driver software, which can introduce vulnerabilities. We describe the Trustworthy Autonomic Interface Guardian Architecture (TAIGA) that provides an on-chip, digital, security version of classic mechanical interlocks. In order to enhance trust in critical embedded processes, TAIGA redistributes responsibilities and authorities between a Programmable Logic Controller (PLC) processor and a hardware-implemented interface controller, simplifying PLC software without significantly degrading performance while separating trusted components from updatable software. The interface controller is synthesized from C code, formally analyzed, and permits runtime checked, authenticated updates to certain system parameters but not code. TAIGA's main focus is ensuring process stability even if this requires overriding commands from the processor or supervisory nodes. The TAIGA architecture is mapped to a commercial, configurable system-on-chip platform.