Cyber-physical microgrids are vulnerable to rootkit attacks that manipulate system dynamics to create instabilities in the network. Rootkits tend to hide their access level within microgrid system components to launch sudden attacks that prey on the slow response time of defenders to manipulate system trajectory. This problem can be formulated as a multi-stage, non-cooperative, zero-sum, Markov game with the attacker and the defender modeled as opposing players. To maximize defender utilities at all stages of the game, this paper proposes a deep reinforcement learning-based strategy that dynamically identifies rootkit access levels and isolates incoming manipulations by incorporating changes in the defense plan. Our proposed approach introduces a novel reward formulation mechanism that quantifies the level of microgrid instability in real time. This allows us to detect the presence of rootkits in the microgrid. The developed reward formulation also allows us to achieve scalability of the proposed solution to different grid sizes and configurations. A major advantage of the proposed strategy is its ability to establish resiliency without altering the physical transmission/distribution network topology, thereby diminishing potential instability issues. The paper presents several simulations and case studies to demonstrate the operating mechanism and robustness of the proposed strategy.
5 November 2024 in IEEE Transactions on Network Science and Engineering
With the usage of Machine Learning (ML) algorithms in modern-day Network Intrusion Detection Systems (NIDS), contemporary network communications are efficiently protected from cyber threats. However, these ML algorithms are starting to be compromised by adversarial attacks that ambush the ML pipeline. This paper demonstrates the feasibility of an adversarial attack called the Cosine Similarity Label Manipulation (CSLM) which is geared toward compromising training labels for ML-based NIDS. The paper develops two versions of CSLM attacks: Minimum CSLM (Min-CSLM) and Maximum CSLM (Max-CSLM). We demonstrate the attacks' efficacy towards single and multi-controller Software-defined Network (SDN) setups. Results indicate that the proposed attacks provide substantial deterioration of classifier performance in single SDNs, specifically, those that utilize Random Forests (RF), which deteriorate $\approx$ 50% under Min-CSLM attacks, and Support Vector Machines (SVM), which undergo $\approx$ 60% deterioration from a Max-CSLM attack. We also note that RF, SVM, and Multi-layer Perceptron (MLP) classifiers are also extensively vulnerable to these attacks in Multi-controller SDN setups (MSDN) as they incur the most observed utility deterioration. MLP-based uniform MSDNs incur the most deterioration under both proposed CSLM attacks with $\approx$ 28% decrease in performance, while SVM and RF-based variable MSDNs incur the most deterioration under both CSLM attacks with $\approx$ 30% and $\approx$ 35% decrease in performance, respectively.
18 December 2024 in IEEE Transactions on Network Science and Engineering
Machine learning (ML)-based network intrusion detection systems (NIDS) have become a prospective approach to efficiently protect network communications. However, ML models can be exploited by adversarial poisonings, like Random Label Manipulation (RLM), which can compromise multi-controller software-defined network (MSDN) operations. In this paper, we develop the Trans-controller Adversarial Perturbation Detection (TAPD) framework for NIDS for MSDNs. The detection framework takes advantage of the MSDN architecture and focuses on periodic transference of ML-based NIDS models across the SDN controllers in the topology, and validates the models using local datasets to calculate error rates. We demonstrate the efficacy of this framework in detecting RLM attacks in an MSDN setup. Results indicate efficient detection performance by the TAPD framework in determining the presence of RLM attacks and the localization of the compromised controllers. We find that the framework works well even when there is a significant number of compromised agents. However, the performance begins to deteriorate when more than 40% of the SDN controllers have become compromised.
Interdisciplinary developments like the smart grid (SG) provide enhanced functionality like efficient power delivery, reliability, and safety while ensuring the smooth integration of traditional and renewable sources of power. Unfortunately, the progressive utilization of the SG and its applications makes them a prime target for cyber attacks. In this article, we propose the grid cyber attack progression (GCAP) framework for SG security. The proposed framework provides higher flexibility in how SG organizations can prepare for complex cyber attacks by incorporating behavior patterns like the adversary skipping and/or backtracking stages during attack progression, while still maintaining a sequential and hierarchical structure. Unlike contemporary frameworks, GCAP provides us with cyber attack progression from the defender’s perspective. We illustrate the advantages of the proposed framework over established frameworks for SG security, after which we evaluate the framework on multiple case studies on real-world cyber attacks on power systems. Finally, we identify defensive solutions that can be laid down in every stage of the GCAP framework to provide more security and protection for SG infrastructures.
With the usage of machine learning (ML) algorithms in modern-day network intrusion detection systems (NIDS), contemporary network communications are efficiently protected from cyber threats. However, these ML algorithms are starting to be compromised by adversarial attacks that ambush the ML pipeline. In this paper, we demonstrate the feasibility of an adversarial attack called the Cosine Similarity Label Manipulation (CSLM), which is geared toward compromising training labels for ML-based NIDS, and how they can affect ML pipelines. We demonstrate the efficacy of the attacks towards both single and multi-controller software-defined network (SDN) setups. Results indicate that the proposed attacks provide substantial deterioration of classifier performance in single SDNs, specifically, those that utilize RF, which deteriorates ~50% under Min-CSLM attacks, and SVMs, which undergo ~60% deterioration from a Max-CSLM attack. We also note that RF, SVM, and MLP classifiers are also extensively vulnerable to these attacks in MSDNs as they incur the most observed utility deterioration. MLP-based uniform multi-controller setups (MSDN) incur the most deterioration under both proposed CSLM attacks with ~28% decrease in performance, while SVM and RF-based variable MSDNs incur the most deterioration under both CSLM attacks with ~30% and ~35% decrease in performance, respectively.
Malwares have historically been the central threat concerning cybersecurity. With the rise in modern technologies, malwares have become more lethal due to the increasing connectivity between various computing environments. To exacerbate this increasing threat, modern day anti-virus signature matching seems to provide limited effectiveness, which makes it possible that new malware variants can evade detection when their behavior does not correlate exactly with known malware signatures on a system. There is also a wide time gap between when a new malware variant is released, and its corresponding antivirus signature is developed and made available to the public. This time gap is when these new malware variants can cause havoc on computing environments. To counteract this threat, the paper proposes a novel malware detection methodology to analyze and detect malware and classify malware by family. This is performed through time series analysis of the sequence of API calls made, using Extreme Learning Machines (ELM) and Online Sequential Extreme Learning Machines (OS-ELM). The paper further analyzes the approach by accurately detecting and predicting malware family classification using a subset of the API call sequences. Simulation results show that the proposed approach can accurately detect and classify malware families with higher accuracy and greater speed than traditional LSTM. Our experiments show ELM and OS-ELM performing with 91% accuracy within just 3 seconds of learning time in contrast to other methods. Thus, our method is superior both in terms of training time and accuracy.