Implementation and detection of novel attacks to the PLC memory of a clean water supply system

Abstract

Critical infrastructures such as nuclear plants and water supply systems are mainly managed through electronic control systems. Such control systems comprise of a number of elements, such as programmable logic controllers (PLC), networking devices, sensors and actuators. With the development of online and networking solutions, such control systems can be managed online. Even though network connected control systems permit users to keep up to date with system operation, it also opens the door to attackers taking advantages of such availability. In this paper, a novel attack vector for modifying PLC memory is proposed, which affects the perceived values of sensors, such as a water flow meter, or the operation of actuators, such as a pump. In addition, this attack vector can also manipulate control variables located in the PLC working memory, reprogramming decision making rules. To show the impact of the attacks in a real scenario, a model of a clean water supply system is implemented on a Festo MPA rig. The results show that the attacks on the PLC memory can have a significant detrimental effect on control system operations. Further, a mechanism of detecting such attacks on the PLC memory is proposed based on monitoring energy consumption and electrical signals using current-measurement sensors. The results show the successful implementation of the novel PLC attacks as well as the feasibility of detecting such attacks.