Proposed Defense Research Articles

Stealthy DGoS Attack Against Network Tomography: The Role of Active Measurements

As a tool to infer the internal state of a network that cannot be measured directly, network tomography has been extensively studied under the assumption that the measurements truthfully reflect the end-to-end performance of measurement paths, which makes the resulting solutions vulnerable to manipulated measurements. In this work, we investigate the impact of manipulated measurements via a recently proposed attack model called the stealthy DeGrading of Service (DGoS) attack , which aims at maximally degrading the performance of targeted paths without exposing the manipulated links to network tomography. While existing studies on this attack assumed that network tomography only measures the paths actively used for data transfer (via passive measurements), our model allows network tomography to measure a larger set of paths, e.g., by sending probes on some paths not carrying data flows. By developing and analyzing the optimal attack strategy, we quantify the maximum damage of such an attack. We further develop a defense strategy by formulating and solving a Stackelberg game to select the best set of measurement paths under a budget constraint. Our evaluations on real topologies validate the efficacy of the proposed defense strategy while identifying areas for further improvement.

BiMorphing: A Bi-Directional Bursting Defense against Website Fingerprinting Attacks

Network traffic analysis has been increasingly used in various applications to either protect or threaten people, information, and systems. Website fingerprinting is a passive traffic analysis attack which threatens web navigation privacy. It is a set of techniques used to discover patterns from a sequence of network packets generated while a user accesses different websites. Internet users (such as online activists or journalists) may wish to hide their identity and online activity to protect their privacy. Typically, an anonymity network is utilized for this purpose. These anonymity networks such as Tor (The Onion Router) provide layers of data encryption which poses a challenge to the traffic analysis techniques. Although various defenses have been proposed to counteract this passive attack, they have been penetrated by new attacks that proved the ineffectiveness and/or impracticality of such defenses. In this work, we introduce a novel defense algorithm to counteract the website fingerprinting attacks. The proposed defense obfuscates original website traffic patterns through the use of double sampling and mathematical optimization techniques to deform packet sequences and destroy traffic flow dependency characteristics used by attackers to identify websites. We evaluate our defense against state-of-the-art studies and show its effectiveness with minimal overhead and zero-delay transmission to the real traffic.

Network Security Model Based on Active and Passive Defense Hybrid Strategy

This paper studies a new network information security system and its evaluation method based on the security threat situation network security model. This paper expounds the application of active defense system and passive defense system in network security, and uses advanced technical support platform and network security model. In this paper, the active defense system and the passive defense system are organically combined to form a new model of computer network security with characteristics. At the same time, based on the idea of security threat situation evaluation, this paper extends the traditional PDRR security model and proposes a network active defense security model and defense in depth security strategy. On this basis, a hybrid architecture of active and passive defense is proposed. The architecture integrates some new network active defense technologies, and constructs a comprehensive and deep network defense system together with network passive defense means. The experimental results show that the proposed defense model improves the security defense efficiency and effective defense rate of the traditional PDRR security model.

Detection of and Countermeasure Against Thermal Covert Channel in Many-Core Systems

The thermal covert channels (TCCs) in many-core systems can cause detrimental data breaches. In this article, we present a three-step scheme to detect and fight against such TCC attacks. Specifically, in the detection step, each core calculates the spectrum of its own CPU workload traces that are collected over a few fixed time intervals, and then it applies a frequency scanning method to detect if there exists any TCC attack. In the next positioning step, the logical cores running the transmitter threads are located. In the last step, the physical CPU cores suspiciously engaging in a TCC attack have to undertake dynamic voltage frequency scaling (DVFS) such that any possible TCC trace will be essentially wiped out. Our experiments have confirmed that on average 97&#x0025; of the TCC attacks can be detected, and with the proposed defense, the packet error rate (PER) of a TCC attack can soar to more than 70&#x0025;, literally shutting down the attack in practical terms. The performance penalty caused by the inclusion of the proposed DVFS countermeasures is found to be only 3&#x0025; for an <inline-formula> <tex-math notation="LaTeX">$8\times 8$ </tex-math></inline-formula> many-core system.

Cognitive data augmentation for adversarial defense via pixel masking

• We proposed PixelMask: a novel data augmentation technique for adversarial robustness. • Extensive experiments showcase the effectiveness of the proposed defense algorithm. • Proposed PixelMask defense outperforms other strong data augmentation techniques. The vulnerability of deep networks towards adversarial perturbations has motivated the researchers to design detection and mitigation algorithms. Inspired by the dropout and dropconnect algorithms as well as augmentation techniques, this paper presents “PixelMask” based data augmentation as an efficient method of reducing the sensitivity of convolutional neural networks (CNNs) towards adversarial attacks. In the proposed approach, samples generated using PixelMask are used as augmented data, which helps in learning robust CNN models. Experiments performed with multiple databases and architectures show that the proposed PixelMask based data augmentation approach improves the classification performance on adversarially perturbed images. The proposed defense mechanism can be applied effectively for different adversarial attacks and can easily be combined with any deep neural network (DNN) architecture to increase the robustness. The effectiveness of the proposed defense is demonstrated in gray-box, white-box, and unseen train-test attack scenarios. For example, on the CIFAR-10 database under adaptive attack (i.e., projected gradient descent), the proposed PixelMask is able to improve the recognition performance of CNN by at-least 22.69%. Another advantage of the proposed algorithm over several existing defense algorithms is that the proposed defense either is able to retain or increase the classification accuracy of clean examples.

Block-Wise Image Transformation With Secret Key for Adversarially Robust Defense

In this paper, we propose a novel defensive transformation that enables us to maintain a high classification accuracy under the use of both clean images and adversarial examples for adversarially robust defense. The proposed transformation is a block-wise preprocessing technique with a secret key to input images. The proposed defense obfuscates gradients in the absence of the secret key unlike previously defeated obfuscating defenses. We developed three algorithms to realize the proposed transformation: Pixel Shuffling, Bit Flipping, and FFX Encryption. Experiments were carried out on the CIFAR-10 and ImageNet datasets by using both black-box and white-box attacks with various metrics including adaptive ones. The results show that the proposed defense achieves high accuracy close to that of using clean images even under adaptive attacks for the first time. In the best-case scenario, a model trained by using images transformed by FFX Encryption (block size of 4) yielded an accuracy of 92.30% on clean images and 91.48% under PGD attack with a noise distance of 8/255, which is close to the non-robust accuracy (95.45%) for the CIFAR-10 dataset, and it yielded an accuracy of 72.18% on clean images and 71.43% under the same attack, which is also close to the standard accuracy (73.70%) for the ImageNet dataset. Overall, all three proposed algorithms are demonstrated to outperform state-of-the-art defenses including adversarial training whether or not a model is under attack.

Stability-Based Analysis and Defense against Backdoor Attacks on Edge Computing Services

With the explosive development of mobile Internet and deep learning (DL), intelligent edge computing services based on collaborative learning are widely deployed in various application scenarios. These intelligent services include intelligent applications based on edge computing and DL-based optimization for edge computing (e.g., caching and communicating). However, in a wide variety of domains, DL has been found to be vulnerable to adversarial attacks, especially architecture-independent backdoor attacks. It embeds the attack pattern into the learned model and only performs the attack when it encounters the corresponding trigger. In this article, for the first time we analyze the impact of backdoor attacks on intelligent edge computing services. The simulation results demonstrate that once one or more edge nodes implement backdoor attacks, the embedded attack pattern will rapidly expand to all relevant edge nodes, which poses huge challenges to security-sensitive intelligent edge computing services. Subsequently, we analyze the trade-off between expected performance and ability to defend against backdoor attacks, which sheds new light on designing defense mechanisms for intelligent edge computing services. To address the challenges posed by backdoor attacks, we propose a stability-based defense mechanism. The experimental results demonstrate that the newly proposed defense mechanism can effectively defend against different levels of backdoor attacks without knowing whether there are adversaries, which is conducive to the deployment of the stability-based defense mechanism in real-world scenarios.

Defending Support Vector Machines Against Data Poisoning Attacks

Support Vector Machines (SVMs) are vulnerable to targeted training data manipulations such as poisoning attacks and label flips. By carefully manipulating a subset of training samples, the attacker forces the learner to compute an incorrect decision boundary, thereby causing misclassifications. Considering the increased importance of SVMs in engineering and life-critical applications, we develop a novel defense algorithm that improves resistance against such attacks. Local Intrinsic Dimensionality (LID) is a promising metric that characterizes the outlierness of data samples. In this work, we introduce a new approximation of LID called K-LID that uses kernel distance in the LID calculation, which allows LID to be calculated in high dimensional transformed spaces. We introduce a weighted SVM against such attacks using K-LID as a distinguishing characteristic that de-emphasizes the effect of suspicious data samples on the SVM decision boundary. Each sample is weighted on how likely its K-LID value is from the benign K-LID distribution rather than the attacked K-LID distribution. Experiments with benchmark data sets show that the proposed defense reduces classification error rates substantially (10% on average).

Cyclic Defense GAN Against Speech Adversarial Attacks

This paper proposes a new defense approach for counteracting state-of-the-art white and black-box adversarial attack algorithms. Our approach fits into the implicit reactive defense algorithm category since it does not directly manipulate the potentially malicious input signals. Instead, it reconstructs a similar signal with a synthesized spectrogram using a cyclic generative adversarial network. This cyclic framework helps to yield a stable generative model. Finally, we feed the reconstructed signal into the speech-to-text model for transcription. The conducted experiments on targeted and non-targeted adversarial attacks developed for attacking DeepSpeech, Kaldi, and Lingvo models demonstrate the proposed defense's effectiveness in adverse scenarios.

Load altering attack-tolerant defense strategy for load frequency control system

Cyber attacks become emerging threats to every information-oriented energy management system. By violating the cyber systems, the hacker can disrupt the security and stability due to the strong coupling between the cyber and physical facilities. In this paper, one type of cyber attacks designated as the load altering attack is studied for the power system frequency control, and corresponding defense strategies are proposed to improve the frequency control performance. Considering the difficulty of the application of model-based controller into large-scale power systems, a novel model-free defense framework is for the first time presented. Under this framework, both active defense and passive defense strategies are designed. The former assumes that the defender has the initiative to learn different attack scenarios. Adaptive defense strategies are implemented using the online attack identification information and off-line trained strategy pool. The latter assumes that the defender passively tolerates various attack scenarios via the pre-trained off-line strategy. Both approaches prove to be effective through validation based on the IEEE benchmark systems. The proposed defense framework and defense strategies can be extended to other energy control systems to enhance their attack tolerance capability.

Robust source camera identification against adversarial attacks

Application of Deep Neural Networks (DNN) has dramatically improved the performance of Source Camera Identification (SCI), but easily suffers from adversarial attacks. These attacks raise security problems by tampering the identified outcomes with imperceptible noise. To address this issue, we analyze the feature extraction mapping for DNN-based SCI models on manifolds and discover that the vulnerability comes from the oscillation of the mapping. In light of this, we take that the feature extraction mapping should satisfy locally smooth and information monotonicity as a new design principle for robust SCI, and accordingly developed a defensive scheme. The proposed scheme constructs local smooth mapping that guarantees information monotonicity and achieves sufficient statistics by minimizing Kullback Leibler Divergence (KLD) between the local statistic coordinates on two manifolds. To enhance the usability of our method, we implement it with a Pre-Defense Network (PDN) trained by a two-phase training strategy, which ensures robustness, accuracy, and portability. Experiments on Dresden Image Dataset demonstrate that the proposed defense method offers not only strong robustness for the DNN-based SCI model against adversarial attacks, but also yields comparable or even superior identification performance over existing defense methods. Moreover, PDN also shows defense effect when migrated to other DNN-based SCI models, without extra retraining.

Thwarting All Logic Locking Attacks: Dishonest Oracle With Truly Random Logic Locking

While logic locking is a promising defense to protect hardware designs, many attacks have been shown to undermine its security by retrieving the secret key. All the powerful attacks rely on a working chip, i.e., an oracle, and in particular, heavily use the test access. The proposed technique DisORC turns the oracle into a dishonest one whenever a potential attack is detected. DisORC works on the premise that structural testing of chips need not be performed with the correct functionality. We implement this capability by adding circuitry around a logic-locked design that reconfigures its functionality upon detecting access to scan chains. Any attempt to access scan chains disconnects the secret key from the circuit, and clears all of its traces, isolating and securing it. We also pair this defense with a truly random logic locking (TRLL) scheme that makes random decisions in inserting key gates and retaining signal polarities without relying on any logic synthesis technique to perform bubble pushing. Any netlist analysis-based attack, known or anticipated, will then learn nothing useful to infer the key values. The combined defense DisORC + TRLL thwarts oracle-based and netlist analysis-based attacks while delivering sufficient corruption levels at the outputs. We also show that the proposed defense is cost effective and can be integrated into the design flow easily. The proposed logic locking defense provides protection against untrusted foundry, testing facility, end users, and any combination of them colluding together.

Defending Hardware-Based Malware Detectors Against Adversarial Attacks

In the era of Internet of Things (IoT), Malware has been proliferating exponentially over the past decade. Traditional anti-virus software are ineffective against modern complex Malware. In order to address this challenge, researchers have proposed hardware-assisted Malware detection (HMD) using hardware performance counters (HPCs). The HPCs are used to train a set of machine learning (ML) classifiers, which in turn, are used to distinguish benign programs from Malware. Recently, adversarial attacks have been designed by introducing perturbations in the HPC traces using an adversarial sample predictor to misclassify a program for specific HPCs. These attacks are designed with the basic assumption that the attacker is aware of the HPCs being used to detect Malware. Since modern processors consist of hundreds of HPCs, restricting to only a few of them for Malware detection aids the attacker. In this article, we propose a moving target defense (MTD) for this adversarial attack by designing multiple ML classifiers trained on different sets of HPCs. The MTD randomly selects a classifier; thus, confusing the attacker about the HPCs or the number of classifiers applied. We have developed an analytical model which proves that the probability of an attacker to guess the perfect HPC-classifier combination for MTD is extremely low (in the range of $10^{-1864}$ for a system with 20 HPCs). Our experimental results prove that the proposed defense is able to improve the classification accuracy of HPC traces that have been modified through an adversarial sample generator by up to 31.5%, for a near perfect (99.4%) restoration of the original accuracy.

Defense Strategy Against Load Redistribution Attacks on Power Systems Considering Insider Threats

In recent years, cybersecurity is emerging as one of the most critical problems to the normal operation of power systems. Meanwhile, insider threats are considered as a strategic and serious challenge in the cybersecurity research and have attracted increasing attention. In this article, the detailed model of load redistribution (LR) attacks against the power systems considering the presence of insider threats is developed. The LR attack problem is formulated with a security resource allocation game model. Information leakage of the system operator's defense strategy by the insider to the external attacker is considered in the proposed game model. The optimal strategies of both the system operator and attacker in the presence of the insider are calculated to maximize their own payoffs. In so doing, the impacts of the insider in the LR attacks can be investigated with the proposed security resource allocation game model. Case studies based on the IEEE 39-bus and IEEE 118-bus test systems were conducted to validate the proposed model. The results of the case studies show that the information leakage by the insider will increase the payoff of the attacker in the LR attacks. The damage on the grid can be considerable even if the information leakage probability is small. The proposed defense strategy is able to reduce the expected operation cost of the system under the LR attacks with the information leakage due to the insider threats.

False Data Injection Attack and Defense Method on Load Frequency Control

The load frequency control (LFC) system is an essential component of the smart grid and is located between cyberspace information and the physical infrastructure. Because the operation of LFC depends on the modern network communication technology, it is vulnerable to network attacks. The false data injection attack (FDIA) injects attack data into the data collected by the LFC system, affecting the control center's calculation and causing errors in the control signals, which affects the stable operation of the power system. For the LFC system, this article builds a cyber-physical model. Based on the model, an FDIA detection and defense mechanism based on the GAN network is proposed. By calculating, contrasting, and replacing the control signals, the proposed method can detect and defend against FDIAs. The experimental results on the two-area interconnected power system, the IEEE 39-bus, and IEEE 118-bus power system show that the proposed defense method can ensure the power system frequency and tie-line power deviation remain within the set value when suffered from FDIA.

Observer-Based Fixed-Time Secure Tracking Consensus for Networked High-Order Multiagent Systems Against DoS Attacks.

This article studies the secure tracking consensus problem of nonlinear multiagent systems (MASs) against denial-of-service (DoS) attacks. Two types of DoS attacks, i.e., connectivity-maintained attacks and connectivity-broken attacks, are considered. The resulting topologies caused by DoS attacks may destabilize the consensus performance of MASs. Especially under connectivity-broken attacks, the connectivity between agents is destroyed. To deal with these difficulties, a novel defense strategy consisting of distributed observation and decentralized control is proposed. First, a distributed fixed-time observer (DFTO) is prepared for the case of connectivity-maintained attacks, which can quickly and accurately estimate the leader's information for each follower. Besides, the adverse impact of DoS attacks is completely eliminated. Furthermore, to cope with the problem arising from connectivity-broken attacks, by using an online algorithm of updating label information, an improved resilient DFTO (RDFTO) is further developed, which can preserve those followers having directed paths from the leader to quickly and accurately estimate the leader's information, without being affected by DoS attacks. The developed DFTO and RDFTO have successfully eliminated or weakened the adverse effects caused by DoS attacks. Subsequently, based on the proposed DFTO/RDFTO with the power integrator technique, a fixed-time controller is finally constructed, which realizes the desired transient performance of consensus tracking in the finite-time interval. The effectiveness of the proposed defense strategy is verified by stability analysis and simulation examples.

An anomaly based distributed detection system for DDoS attacks in Tier-2 ISP networks

In the present computer era, the vulnerabilities inherent in the Internet architecture enable various kinds of attacks. Distributed Denial of Service (DDoS) is one of such prominent attack that is a lethal threat to Internet domain that harnesses its computing and communication resources. The increase in network traffic rates of legitimate traffic and its flow similarity with attack traffic has made the DDoS detection very difficult despite deployment of diversified defense solutions. The ISPs are bound to invest heavily to counter such problems which has a significant impact on company finances. To provide uninterrupted quality services to the end users, ISPs needs to deploy a distributed solution for timely detection and discrimination of attack and behaviorally similar flash events (FE) traffic. Such distributed defense systems can be deployed at source-end, intermediate network-end or at the victim-end location. Since the volume of traffic to be analyzed is very large, the detection accuracy and low computational complexity of the proposed defense solution is always a challenging problem. This paper proposes an ISP level distributed, collaborative and automated (D-CAD) defense system for detecting DDoS attacks and FEs, and has the capability to effectively distinguishing the two. Additionally, D-CAD defense system is also capable of categorizing FE traffic and has low computational complexity. The proposed system is validated in novel software defined networks (SDN) using Mininet emulator. The results show that D-CAD defense system outperformed its existing counterparts on various detection system evaluation metrics.

Hardening Random Forest Cyber Detectors Against Adversarial Attacks

Machine learning algorithms are effective in several applications, but they are not as much successful when applied to intrusion detection in cyber security. Due to the high sensitivity to their training data, cyber detectors based on machine learning are vulnerable to targeted adversarial attacks that involve the perturbation of initial samples. Existing defenses assume unrealistic scenarios; their results are underwhelming in non-adversarial settings; or they can be applied only to machine learning algorithms that perform poorly for cyber security. We present an original methodology for countering adversarial perturbations targeting intrusion detection systems based on random forests. As a practical application, we integrate the proposed defense method in a cyber detector analyzing network traffic. The experimental results on millions of labelled network flows show that the new detector has a twofold value: it outperforms state-of-the-art detectors that are subject to adversarial attacks; it exhibits robust results both in adversarial and non-adversarial scenarios.

Flexible Power System Defense Strategies in an Isolated Microgrid System with High Renewable Power Generation

This work develops an underfrequency preventive control strategy for an islanded power system with a high penetration of wind power generation. First, the preventive control strategy uses the frequency nadir forecasting module to analyze the frequency stability under largest diesel generator tripping (N-1) contingency events. If predicted frequency nadir is too low, four frequency support methods are then analyzed and used for preventing potential frequency stability problem. They include generator rescheduling (GR), the use of battery energy storage system (BESS), direct load control (DLC) and emergency demand response program (EDRP). In terms of the GR method, the optimal diesel generator dispatch is obtained, with sufficient frequency stability and minimal fuel cost and start-up cost. In the BESS method, the optimal instantaneous power output from BESS is obtained based on its frequency support capability. With the DLC or EDRP method, the optimal contract-based load-shedding or the load-reduction to provide frequency support is obtained, respectively. Then, the operating costs of each method to support frequency are analyzed. The research methods and simulation results are very useful to the low-frequency protection of actual power systems with high renewable power generation. This work proposed a complete defense strategy in a microgrid system. It combines GR, BESS, DLC and EDRP. Therefore, the system operators have many options to implement their defense strategies, based on the operating costs of various methods. In other words, the proposed defense strategy provides a more flexible solution for the protection of micro grids with a high renewable power penetration. Therefore, the solution considers the system safety and economical aspects.

An Investigation on Feasibility and Security for Cyberattacks on Generator Synchronization Process

Electric power supply is crucial for the effective operation of many sectors such as manufacturing, health care, etc. Further, electrification of the transportation network including metro-rail systems and personal electric vehicles necessitates capacity expansion of power distribution system (PDS). Capacity expansion in PDSs is seeing a clear paradigm shift from component upgrades to the incorporation of distributed energy resources. Though renewable energy sources are attractive from the sustainability aspect, conventional generators are also essential for many applications such as maintaining the grid stability. In this article, for the first time, we report a new type of attack on the generators; attacks on the synchronization process, targeting the availability of generators. The threat model considered in this article is similar to the popular real-world attacks, such as “Stuxnet worm” or programmable logic controller root-kit attacks such as “Harvey,” however, the reported attack goal is novel. We propose a defense mechanism that can be decoupled from the network data and is named as physical layer security (PLS). We present theoretical analysis, simulation results, and experimental validation for the reported attack. We also propose an alternative implementation named as Level 0–1 security for cases where PLS may not be practical. The proposed defense is generic and can be extended to many other applications/attack vectors.