GridAttackSim: A Cyber Attack Simulation Framework for Smart Grids

Abstract

The smart grid system is one of the key infrastructures required to sustain our future society. It is a complex system that comprises two independent parts: power grids and communication networks. There have been several cyber attacks on smart grid systems in recent years that have caused significant consequences. Therefore, cybersecurity training specific to the smart grid system is essential in order to handle these security issues adequately. Unfortunately, concepts related to automation, ICT, smart grids, and other physical sectors are typically not covered by conventional training and education methods. These cybersecurity experiences can be achieved by conducting training using a smart grid co-simulation, which is the integration of at least two simulation models. However, there has been little effort to research attack simulation tools for smart grids. In this research, we first review the existing research in the field, and then propose a smart grid attack co-simulation framework called GridAttackSim based on the combination of GridLAB-D, ns-3, and FNCS. The proposed architecture allows us to simulate smart grid infrastructure features with various cybersecurity attacks and then visualize their consequences automatically. Furthermore, the simulator not only features a set of built-in attack profiles but also enables scientists and electric utilities interested in improving smart grid security to design new ones. Case studies were conducted to validate the key functionalities of the proposed framework. The simulation results are supported by relevant works in the field, and the system can potentially be deployed for cybersecurity training and research.

GridAttackAnalyzer: A Cyber Attack Analysis Framework for Smart Grids

“…At the node level, the metric is used to measure the attack cost for successfully compromising a node. For node t ∈ T of an attack tree and each inner node, the value of attack cost for a node is calculated by Equation (4). At the path level, the measure is the cost spent by an attacker to compromise the target over the attack path.…”

“…For the network model, there have been numerous network architecture designs for the smart grid system [30,31]. We have employed the IEEE feeders in our previous research in [4,5]. Hence, in the scope of this study, PNNL taxonomy feeders will be used.…”

“…According to the report by the European Union Agency for Cybersecurity (ENISA) [35], channel jamming, DNS attacks, injection attacks, and malicious code are among the attacks that should be considered in Smart Grid area. The simulation of these attack types are detailed in our related research in [4]. In this work, we focus on vulnerability exploitation attacks.…”

“…Fortunately, simulation testbeds and analytical techniques can be applied to conduct cybersecurity experimentation in the critical domains where testing on a real system is prohibited, such as smart grid. A smart grid simulation approach was applied in our related research in [4,5]. Besides, there is an increasing interest in the development of smart grid cyber-physical testbeds [6].…”

Detecting Anomalies using Overlapping Electrical Measurements in Smart Power Grids

“…In recent years, cyber attacks in critical infrastructures have become more frequent [1]- [4] and increased in complexity and sophistication. Smart power grids are vulnerable to these attacks in large part due to an increase in cyber-physical connectivity in Supervisory Control and Data Acquisition (SCADA) systems that are used to monitor and control parts of the power grid [5].…”