Randomizing AMI configuration for proactive defense in smart grid

Smart grids are capable of bi-directional communication between smart meters and headend systems. It is a core feature of smart grid provided by the underlying Advanced Metering Infrastructure (AMI). Due to the critical nature of AMI, threats have been targeted towards it and a need of tailored defense mechanism has been highlighted in the recent studies. Since limited number of protocols and applications are supported in the AMI, it exhibits a predictable and deterministic behavior. This predictability makes it easier for an attacker to learn the network behavior and launch an evasion attack by generating similar traffic. To combat this, we present a mutable AMI configuration technique for Proactive Defense. We randomize three different configuration parameters in order to make the AMI behavior unpredictable. While providing the randomization, the approach stays deterministic for the AMI devices like smart collector in order to detect any evasion attempts. We use real-world dataset comprising of approximately 2000 meters collected at an AMI of a leading utility provider for analysis. We also use smart grid testbed to show the effectiveness of our mutation technique.