The adoption of deep learning-based side-channel analysis(DL-SCA)is crucial for leak detection in secure products.Many previous studies have applied this method to break targets protected with countermeasures.Despite ...The adoption of deep learning-based side-channel analysis(DL-SCA)is crucial for leak detection in secure products.Many previous studies have applied this method to break targets protected with countermeasures.Despite the increasing number of studies,the problem of model overfitting.Recent research mainly focuses on exploring hyperparameters and network architectures,while offering limited insights into the effects of external factors on side-channel attacks,such as the number and type of models.This paper proposes a Side-channel Analysis method based on a Stacking ensemble,called Stacking-SCA.In our method,multiple models are deeply integrated.Through the extended application of base models and the meta-model,Stacking-SCA effectively improves the output class probabilities of the model,leading to better generalization.Furthermore,this method shows that the attack performance is sensitive to changes in the number of models.Next,five independent subsets are extracted from the original ASCAD database as multi-segment datasets,which are mutually independent.This method shows how these subsets are used as inputs for Stacking-SCA to enhance its attack convergence.The experimental results show that Stacking-SCA outperforms the current state-of-the-art results on several considered datasets,significantly reducing the number of attack traces required to achieve a guessing entropy of 1.Additionally,different hyperparameter sizes are adjusted to further validate the robustness of the method.展开更多
Recently,several PC oracle based side-channel attacks have been proposed against Kyber.However,most of them focus on unprotected implementations and masking is considered as a counter-measure.In this study,we extend P...Recently,several PC oracle based side-channel attacks have been proposed against Kyber.However,most of them focus on unprotected implementations and masking is considered as a counter-measure.In this study,we extend PC oracle based side-channel attacks to the second-order scenario and successfully conduct key-recovery attacks on the first-order masked Kyber.Firstly,we analyze the potential joint information leakage.Inspired by the binary PC oracle based attack proposed by Qin et al.at Asiacrypt 2021,we identify the 1-bit leakage scenario in the masked Keccak implementation.Moreover,we modify the ciphertexts construction described by Tanaka et al.at CHES 2023,extending the leakage scenario from 1-bit to 32-bit.With the assistance of TVLA,we validate these leakages through experiments.Secondly,for these two scenarios,we construct a binary PC oracle based on t-test and a multiple-valued PC oracle based on neural networks.Furthermore,we conduct practical side-channel attacks on masked Kyber by utilizing our oracles,with the implementation running on an ARM Cortex-M4 microcontroller.The demonstrated attacks require a minimum of 15788 and 648 traces to fully recover the key of Kyber768 in the 1-bit leakage scenario and the 32-bit leakage scenario,respectively.Our analysis may also be extended to attack other post-quantum schemes that use the same masked hash function.Finally,we apply the shuffling strategy to the first-order masked imple-mentation of the Kyber and perform leakage tests.Experimental results show that the combination strategy of shuffling and masking can effectively resist our proposed attacks.展开更多
A side-channel attack(SCA)-resistant AES S-box implementation is proposed,which is an improvement from the power-aware hiding(PAH)S-box but with higher security and a smaller area.We use the composite field approach a...A side-channel attack(SCA)-resistant AES S-box implementation is proposed,which is an improvement from the power-aware hiding(PAH)S-box but with higher security and a smaller area.We use the composite field approach and apply the PAH method to the inversion in the nonlinear kernel and a masking method to the other parts.In addition,a delaymatched enable control technique is used to suppress glitches in the masked parts.The evaluation results show that its area is contracted to 63.3%of the full PAH S-box,and its power-delay product is much lower than that of the masking implementation.The leakage assessment using simulation power traces concludes that it has no detectable leakage under t-test and that it at least can thwart the moment-correlation analysis using 665000 noiseless traces.展开更多
Side-channel attacks （SCA） may exploit leakage information to break cryptosystems. In this paper we present a new SCA resistant Elliptic Curve scalar multiplication algorithm. The proposed algorithm, builds a sequen...Side-channel attacks （SCA） may exploit leakage information to break cryptosystems. In this paper we present a new SCA resistant Elliptic Curve scalar multiplication algorithm. The proposed algorithm, builds a sequence of bit-strings representing the scalar k, characterized by the fact that all bit-strings are different from zero; this property will ensure a uniform computation behavior for the algorithm, and thus will make it secure against simple power analysis attacks （SPA）. With other randomization techniques, the proposed countermeasures do not penalize the computation time. The proposed scheme is more efficient than MOEller＇s one, its cost being about 5% to 10% smaller than MOEller＇s one.展开更多
An embedded cryptosystem needs higher reconfiguration capability and security. After analyzing the newly emerging side-channel attacks on elliptic curve cryptosystem （ECC）, an efficient fractional width-w NAF （FWNA...An embedded cryptosystem needs higher reconfiguration capability and security. After analyzing the newly emerging side-channel attacks on elliptic curve cryptosystem （ECC）, an efficient fractional width-w NAF （FWNAF） algorithm is proposed to secure ECC scalar multiplication from these attacks. This algorithm adopts the fractional window method and probabilistic SPA scheme to reconfigure the pre-computed table, and it allows designers to make a dynamic configuration on pre-computed table. And then, it is enhanced to resist SPA, DPA, RPA and ZPA attacks by using the random masking method. Compared with the WBRIP and EBRIP methods, our proposals has the lowest total computation cost and reduce the shake phenomenon due to sharp fluctuation on computation performance.展开更多
Co-residency of virtual machines(VMs) of different tenants on the same physical platform would possibly lead to cross-VM side-channel attacks in the cloud. While most of current countermeasures fail for real or immedi...Co-residency of virtual machines(VMs) of different tenants on the same physical platform would possibly lead to cross-VM side-channel attacks in the cloud. While most of current countermeasures fail for real or immediate deployment due to their requirement for modification of virtualization structure, we adopt dynamic migration, an inherent mechanism of the cloud platform, as a general defense against this kind of threats. To this end, we first set up a unified practical information leakage model which shows the factors affecting side channels and describes the way they influence the damage due to side-channel attacks. Since migration is adopted to limit the time duration of co-residency, we envision this defense as an optimization problem by setting up an Integer Linear Programming(ILP) to calculate optimal migration strategy, which is intractable due to high computational complexity. Therefore, we approximate the ILP with a baseline genetic algorithm, which is further improved for its optimality and scalability. Experimental results show that our migration-based defense can not only provide excellent security guarantees and affordable performance cost in both theoretical simulation and practical cloud environment, but also achieve better optimality and scalability than previous countermeasures.展开更多
The number and creativity of side channel attacks have increased dramatically in recent years. Of particular interest are attacks leveraging power line communication to 1) gather information on power consumption from ...The number and creativity of side channel attacks have increased dramatically in recent years. Of particular interest are attacks leveraging power line communication to 1) gather information on power consumption from the victim and 2) exfiltrate data from compromised machines. Attack strategies of this nature on the greater power grid and building infrastructure levels have been shown to be a serious threat. This project further explores this concept of a novel attack vector by creating a new type of penetration testing tool: an USB power adapter capable of remote monitoring of device power consumption and communicating through powerline communications.展开更多
Side-channel attacks based on supervised learning require that the attacker have complete control over the cryptographic device and obtain a large number of labeled power traces.However,in real life,this requirement i...Side-channel attacks based on supervised learning require that the attacker have complete control over the cryptographic device and obtain a large number of labeled power traces.However,in real life,this requirement is usually not met.In this paper,an attack algorithm based on collaborative learning is proposed.The algorithm only needs to use a small number of labeled power traces to cooperate with the unlabeled power trace to realize the attack to cryptographic device.By experimenting with the DPA contest V4 dataset,the results show that the algorithm can improve the accuracy by about 20%compared with the pure supervised learning in the case of using only 10 labeled power traces.展开更多
Mobile devices within Fifth Generation(5G)networks,typically equipped with Android systems,serve as a bridge to connect digital gadgets such as global positioning system,mobile devices,and wireless routers,which are v...Mobile devices within Fifth Generation(5G)networks,typically equipped with Android systems,serve as a bridge to connect digital gadgets such as global positioning system,mobile devices,and wireless routers,which are vital in facilitating end-user communication requirements.However,the security of Android systems has been challenged by the sensitive data involved,leading to vulnerabilities in mobile devices used in 5G networks.These vulnerabilities expose mobile devices to cyber-attacks,primarily resulting from security gaps.Zero-permission apps in Android can exploit these channels to access sensitive information,including user identities,login credentials,and geolocation data.One such attack leverages“zero-permission”sensors like accelerometers and gyroscopes,enabling attackers to gather information about the smartphone’s user.This underscores the importance of fortifying mobile devices against potential future attacks.Our research focuses on a new recurrent neural network prediction model,which has proved highly effective for detecting sidechannel attacks in mobile devices in 5G networks.We conducted state-of-the-art comparative studies to validate our experimental approach.The results demonstrate that even a small amount of training data can accurately recognize 37.5%of previously unseen user-typed words.Moreover,our tap detection mechanism achieves a 92%accuracy rate,a crucial factor for text inference.These findings have significant practical implications,as they reinforce mobile device security in 5G networks,enhancing user privacy,and data protection.展开更多
The surge in smishing attacks underscores the urgent need for robust,real-time detection systems powered by advanced deep learning models.This paper introduces PhishNet,a novel ensemble learning framework that integra...The surge in smishing attacks underscores the urgent need for robust,real-time detection systems powered by advanced deep learning models.This paper introduces PhishNet,a novel ensemble learning framework that integrates transformer-based models(RoBERTa)and large language models(LLMs)(GPT-OSS 120B,LLaMA3.370B,and Qwen332B)to enhance smishing detection performance significantly.To mitigate class imbalance,we apply synthetic data augmentation using T5 and leverage various text preprocessing techniques.Our system employs a duallayer voting mechanism:weighted majority voting among LLMs and a final ensemble vote to classify messages as ham,spam,or smishing.Experimental results show an average accuracy improvement from 96%to 98.5%compared to the best standalone transformer,and from 93%to 98.5%when compared to LLMs across datasets.Furthermore,we present a real-time,user-friendly application to operationalize our detection model for practical use.PhishNet demonstrates superior scalability,usability,and detection accuracy,filling critical gaps in current smishing detection methodologies.展开更多
Zero-click attacks represent an advanced cybersecurity threat,capable of compromising devices without user interaction.High-profile examples such as Pegasus,Simjacker,Bluebugging,and Bluesnarfing exploit hidden vulner...Zero-click attacks represent an advanced cybersecurity threat,capable of compromising devices without user interaction.High-profile examples such as Pegasus,Simjacker,Bluebugging,and Bluesnarfing exploit hidden vulnerabilities in software and communication protocols to silently gain access,exfiltrate data,and enable long-term surveillance.Their stealth and ability to evade traditional defenses make detection and mitigation highly challenging.This paper addresses these threats by systematically mapping the tactics and techniques of zero-click attacks using the MITRE ATT&CK framework,a widely adopted standard for modeling adversarial behavior.Through this mapping,we categorize real-world attack vectors and better understand how such attacks operate across the cyber-kill chain.To support threat detection efforts,we propose an Active Learning-based method to efficiently label the Pegasus spyware dataset in alignment with the MITRE ATT&CK framework.This approach reduces the effort of manually annotating data while improving the quality of the labeled data,which is essential to train robust cybersecurity models.In addition,our analysis highlights the structured execution paths of zero-click attacks and reveals gaps in current defense strategies.The findings emphasize the importance of forward-looking strategies such as continuous surveillance,dynamic threat profiling,and security education.By bridging zero-click attack analysis with the MITRE ATT&CK framework and leveraging machine learning for dataset annotation,this work provides a foundation for more accurate threat detection and the development of more resilient and structured cybersecurity frameworks.展开更多
Graph Neural Networks(GNNs)have proven highly effective for graph classification across diverse fields such as social networks,bioinformatics,and finance,due to their capability to learn complex graph structures.Howev...Graph Neural Networks(GNNs)have proven highly effective for graph classification across diverse fields such as social networks,bioinformatics,and finance,due to their capability to learn complex graph structures.However,despite their success,GNNs remain vulnerable to adversarial attacks that can significantly degrade their classification accuracy.Existing adversarial attack strategies primarily rely on label information to guide the attacks,which limits their applicability in scenarios where such information is scarce or unavailable.This paper introduces an innovative unsupervised attack method for graph classification,which operates without relying on label information,thereby enhancing its applicability in a broad range of scenarios.Specifically,our method first leverages a graph contrastive learning loss to learn high-quality graph embeddings by comparing different stochastic augmented views of the graphs.To effectively perturb the graphs,we then introduce an implicit estimator that measures the impact of various modifications on graph structures.The proposed strategy identifies and flips edges with the top-K highest scores,determined by the estimator,to maximize the degradation of the model’s performance.In addition,to defend against such attack,we propose a lightweight regularization-based defense mechanism that is specifically tailored to mitigate the structural perturbations introduced by our attack strategy.It enhances model robustness by enforcing embedding consistency and edge-level smoothness during training.We conduct experiments on six public TU graph classification datasets:NCI1,NCI109,Mutagenicity,ENZYMES,COLLAB,and DBLP_v1,to evaluate the effectiveness of our attack and defense strategies.Under an attack budget of 3,the maximum reduction in model accuracy reaches 6.67%on the Graph Convolutional Network(GCN)and 11.67%on the Graph Attention Network(GAT)across different datasets,indicating that our unsupervised method induces degradation comparable to state-of-the-art supervised attacks.Meanwhile,our defense achieves the highest accuracy recovery of 3.89%(GCN)and 5.00%(GAT),demonstrating improved robustness against structural perturbations.展开更多
Transformer-based models have significantly advanced binary code similarity detection(BCSD)by leveraging their semantic encoding capabilities for efficient function matching across diverse compilation settings.Althoug...Transformer-based models have significantly advanced binary code similarity detection(BCSD)by leveraging their semantic encoding capabilities for efficient function matching across diverse compilation settings.Although adversarial examples can strategically undermine the accuracy of BCSD models and protect critical code,existing techniques predominantly depend on inserting artificial instructions,which incur high computational costs and offer limited diversity of perturbations.To address these limitations,we propose AIMA,a novel gradient-guided assembly instruction relocation method.Our method decouples the detection model into tokenization,embedding,and encoding layers to enable efficient gradient computation.Since token IDs of instructions are discrete and nondifferentiable,we compute gradients in the continuous embedding space to evaluate the influence of each token.The most critical tokens are identified by calculating the L2 norm of their embedding gradients.We then establish a mapping between instructions and their corresponding tokens to aggregate token-level importance into instructionlevel significance.To maximize adversarial impact,a sliding window algorithm selects the most influential contiguous segments for relocation,ensuring optimal perturbation with minimal length.This approach efficiently locates critical code regions without expensive search operations.The selected segments are relocated outside their original function boundaries via a jump mechanism,which preserves runtime control flow and functionality while introducing“deletion”effects in the static instruction sequence.Extensive experiments show that AIMA reduces similarity scores by up to 35.8%in state-of-the-art BCSD models.When incorporated into training data,it also enhances model robustness,achieving a 5.9%improvement in AUROC.展开更多
With the increasing emphasis on personal information protection,encryption through security protocols has emerged as a critical requirement in data transmission and reception processes.Nevertheless,IoT ecosystems comp...With the increasing emphasis on personal information protection,encryption through security protocols has emerged as a critical requirement in data transmission and reception processes.Nevertheless,IoT ecosystems comprise heterogeneous networks where outdated systems coexist with the latest devices,spanning a range of devices from non-encrypted ones to fully encrypted ones.Given the limited visibility into payloads in this context,this study investigates AI-based attack detection methods that leverage encrypted traffic metadata,eliminating the need for decryption and minimizing system performance degradation—especially in light of these heterogeneous devices.Using the UNSW-NB15 and CICIoT-2023 dataset,encrypted and unencrypted traffic were categorized according to security protocol,and AI-based intrusion detection experiments were conducted for each traffic type based on metadata.To mitigate the problem of class imbalance,eight different data sampling techniques were applied.The effectiveness of these sampling techniques was then comparatively analyzed using two ensemble models and three Deep Learning(DL)models from various perspectives.The experimental results confirmed that metadata-based attack detection is feasible using only encrypted traffic.In the UNSW-NB15 dataset,the f1-score of encrypted traffic was approximately 0.98,which is 4.3%higher than that of unencrypted traffic(approximately 0.94).In addition,analysis of the encrypted traffic in the CICIoT-2023 dataset using the same method showed a significantly lower f1-score of roughly 0.43,indicating that the quality of the dataset and the preprocessing approach have a substantial impact on detection performance.Furthermore,when data sampling techniques were applied to encrypted traffic,the recall in the UNSWNB15(Encrypted)dataset improved by up to 23.0%,and in the CICIoT-2023(Encrypted)dataset by 20.26%,showing a similar level of improvement.Notably,in CICIoT-2023,f1-score and Receiver Operation Characteristic-Area Under the Curve(ROC-AUC)increased by 59.0%and 55.94%,respectively.These results suggest that data sampling can have a positive effect even in encrypted environments.However,the extent of the improvement may vary depending on data quality,model architecture,and sampling strategy.展开更多
The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)an...The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)and Deep Learning(DL)techniques have demonstrated promising early detection capabilities.However,their effectiveness is limited when handling the vast volumes of IoT-generated data due to scalability constraints,high computational costs,and the costly time-intensive process of data labeling.To address these challenges,this study proposes a Federated Learning(FL)framework that leverages collaborative and hybrid supervised learning to enhance cyber threat detection in IoT networks.By employing Deep Neural Networks(DNNs)and decentralized model training,the approach reduces computational complexity while improving detection accuracy.The proposed model demonstrates robust performance,achieving accuracies of 94.34%,99.95%,and 87.94%on the publicly available kitsune,Bot-IoT,and UNSW-NB15 datasets,respectively.Furthermore,its ability to detect zero-day attacks is validated through evaluations on two additional benchmark datasets,TON-IoT and IoT-23,using a Deep Federated Learning(DFL)framework,underscoring the generalization and effectiveness of the model in heterogeneous and decentralized IoT environments.Experimental results demonstrate superior performance over existing methods,establishing the proposed framework as an efficient and scalable solution for IoT security.展开更多
Existing Side-Channel Attacks （SCAs） have several limitations and, rather than to be real attack methods, can only be considered to be security evaluation methods. Their limitations are mainly related to the samplin...Existing Side-Channel Attacks （SCAs） have several limitations and, rather than to be real attack methods, can only be considered to be security evaluation methods. Their limitations are mainly related to the sampling conditions, such as the trigger signal embedded in the source code of the encryption device, and the acquisition device that serves as the encryption-device controller. Apart from it being very difficult for an attacker to add a trigger into the original design before making an attack or to control the encryption device, there is a big gap in the capacity of existing SCAs to pose real threats to cipher devices. In this paper, we propose a new method, the sliding window SCA （SW-SCA）, which can be applied in scenarios in which the acquisition device is independent of the encryption device and for which the encryption source code requires no trigger signal or modification. First, we describe the main issues in existing SCAs, then we theoretically analyze the effectiveness and complexity of our proposed SW-SCA --a method that can incorporate a sliding-window mechanism into almost all of the existing non-profiled SCAs. The experimental results for both simulated and physical traces verify the effectiveness of the SW-SCA and the appropriateness of its theoretical complexity.展开更多
The security of cryptographic systems is a major concern for cryptosystem designers, even though cryptography algorithms have been improved. Side-channel attacks, by taking advantage of physical vulnerabilities of cry...The security of cryptographic systems is a major concern for cryptosystem designers, even though cryptography algorithms have been improved. Side-channel attacks, by taking advantage of physical vulnerabilities of cryptosystems, aim to gain secret information. Several approaches have been proposed to analyze side-channel information, among which machine learning is known as a promising method. Machine learning in terms of neural networks learns the signature （power consumption and electromagnetic emission） of an instruction, and then recognizes it automatically. In this paper, a novel experimental investigation was conducted on field-programmable gate array （FPGA） implementation of elliptic curve cryptography （ECC）, to explore the efficiency of side-channel information characterization based on a learning vector quantization （LVQ） neural network. The main characteristics of LVQ as a multi-class classifier are that it has the ability to learn complex non-linear input-output relationships, use sequential training procedures, and adapt to the data. Experimental results show the performance of multi-class classification based on LVQ as a powerful and promising approach of side-channel data characterization.展开更多
Fault attacks have emerged as an increasingly effective approach for integrated circuit security attacks due to their short execution time and minimal data requirement.However,the lack of a unified leakage model remai...Fault attacks have emerged as an increasingly effective approach for integrated circuit security attacks due to their short execution time and minimal data requirement.However,the lack of a unified leakage model remains a critical challenge,as existing methods often rely on algorithm-specific details or prior knowledge of plaintexts and intermediate values.This paper proposes the Fault Probability Model based on Hamming Weight(FPHW)to address this.This novel statistical framework quantifies fault attacks by solely analyzing the statistical response of the target device,eliminating the need for attack algorithm details or implementation specifics.Building on this model,a Fault Injection Attack method based on Mutual Information(FPMIA)is introduced,which recovers keys by leveraging the mutual information between measured fault probability traces and simulated leakage derived from Hamming weight,reducing data requirements by at least 44%compared to the existing Mutual Information Analysis method while achieving a high correlation coefficient of 0.9403 between measured and modeled fault probabilities.Experimental validation on an AES-128 implementation via a Microcontroller Unit demonstrates that FPHW accurately captures the data dependence of fault probability and FPMIA achieves efficient key recovery with robust noise tolerance,establishing a unified and efficient framework that surpasses traditional methods in terms of generality,data efficiency,and practical applicability.展开更多
In cloud storage,client-side deduplication is widely used to reduce storage and communication costs.In client-side deduplication,if the cloud server detects that the user’s outsourced data have been stored,then clien...In cloud storage,client-side deduplication is widely used to reduce storage and communication costs.In client-side deduplication,if the cloud server detects that the user’s outsourced data have been stored,then clients will not need to reupload the data.However,the information on whether data need to be uploaded can be used as a side-channel,which can consequently be exploited by adversaries to compromise data privacy.In this paper,we propose a new threat model against side-channel attacks.Different from existing schemes,the adversary could learn the approximate ratio of stored chunks to unstored chunks in outsourced files,and this ratio will affect the probability that the adversary compromises the data privacy through side-channel attacks.Under this threat model,we design two defense schemes to minimize privacy leakage,both of which design interaction protocols between clients and the server during deduplication checks to reduce the probability that the adversary compromises data privacy.We analyze the security of our schemes,and evaluate their performances based on a real-world dataset.Compared with existing schemes,our schemes can better mitigate data privacy leakage and have a slightly lower communication cost.展开更多
Co-residency of different tenants’ virtual machines(VMs) in cloud provides a good chance for side-channel attacks, which results in information leakage. However, most of current defense suffers from the generality or...Co-residency of different tenants’ virtual machines(VMs) in cloud provides a good chance for side-channel attacks, which results in information leakage. However, most of current defense suffers from the generality or compatibility problem, thus failing in immediate real-world deployment. VM migration, an inherit mechanism of cloud systems, envisions a promising countermeasure, which limits co-residency by moving VMs between servers. Therefore, we first set up a unified practical adversary model, where the attacker focuses on effective side channels. Then we propose Driftor, a new cloud system that contains VMs of a multi-executor structure where only one executor is active to provide service through a proxy, thus reducing possible information leakage. Active state is periodically switched between executors to simulate defensive effect of VM migration. To enhance the defense, real VM migration is enabled at the same time. Instead of solving the migration satisfiability problem with intractable CIRCUIT-SAT, a greedy-like heuristic algorithm is proposed to search for a viable solution by gradually expanding an initial has-to-migrate set of VMs. Experimental results show that Driftor can not only defend against practical fast side-channel attack, but also bring about reasonable impacts on real-world cloud applications.展开更多
基金supported by the Hunan Provincial Natural Science Foundation of China(2022JJ30103)“the 14th Five-Year Plan”Key Disciplines and Application-Oriented Special Disciplines of Hunan Province(Xiangjiaotong[2022]351)the Science and Technology Innovation Program of Hunan Province(2016TP1020).
文摘The adoption of deep learning-based side-channel analysis(DL-SCA)is crucial for leak detection in secure products.Many previous studies have applied this method to break targets protected with countermeasures.Despite the increasing number of studies,the problem of model overfitting.Recent research mainly focuses on exploring hyperparameters and network architectures,while offering limited insights into the effects of external factors on side-channel attacks,such as the number and type of models.This paper proposes a Side-channel Analysis method based on a Stacking ensemble,called Stacking-SCA.In our method,multiple models are deeply integrated.Through the extended application of base models and the meta-model,Stacking-SCA effectively improves the output class probabilities of the model,leading to better generalization.Furthermore,this method shows that the attack performance is sensitive to changes in the number of models.Next,five independent subsets are extracted from the original ASCAD database as multi-segment datasets,which are mutually independent.This method shows how these subsets are used as inputs for Stacking-SCA to enhance its attack convergence.The experimental results show that Stacking-SCA outperforms the current state-of-the-art results on several considered datasets,significantly reducing the number of attack traces required to achieve a guessing entropy of 1.Additionally,different hyperparameter sizes are adjusted to further validate the robustness of the method.
基金National Natural Science Foundation of China(62472397)Innovation Program for Quantum Science and Technology(2021ZD0302902)。
文摘Recently,several PC oracle based side-channel attacks have been proposed against Kyber.However,most of them focus on unprotected implementations and masking is considered as a counter-measure.In this study,we extend PC oracle based side-channel attacks to the second-order scenario and successfully conduct key-recovery attacks on the first-order masked Kyber.Firstly,we analyze the potential joint information leakage.Inspired by the binary PC oracle based attack proposed by Qin et al.at Asiacrypt 2021,we identify the 1-bit leakage scenario in the masked Keccak implementation.Moreover,we modify the ciphertexts construction described by Tanaka et al.at CHES 2023,extending the leakage scenario from 1-bit to 32-bit.With the assistance of TVLA,we validate these leakages through experiments.Secondly,for these two scenarios,we construct a binary PC oracle based on t-test and a multiple-valued PC oracle based on neural networks.Furthermore,we conduct practical side-channel attacks on masked Kyber by utilizing our oracles,with the implementation running on an ARM Cortex-M4 microcontroller.The demonstrated attacks require a minimum of 15788 and 648 traces to fully recover the key of Kyber768 in the 1-bit leakage scenario and the 32-bit leakage scenario,respectively.Our analysis may also be extended to attack other post-quantum schemes that use the same masked hash function.Finally,we apply the shuffling strategy to the first-order masked imple-mentation of the Kyber and perform leakage tests.Experimental results show that the combination strategy of shuffling and masking can effectively resist our proposed attacks.
基金This work was supported by the National Science and Technology Major Project of China(2017ZX01030301).
文摘A side-channel attack(SCA)-resistant AES S-box implementation is proposed,which is an improvement from the power-aware hiding(PAH)S-box but with higher security and a smaller area.We use the composite field approach and apply the PAH method to the inversion in the nonlinear kernel and a masking method to the other parts.In addition,a delaymatched enable control technique is used to suppress glitches in the masked parts.The evaluation results show that its area is contracted to 63.3%of the full PAH S-box,and its power-delay product is much lower than that of the masking implementation.The leakage assessment using simulation power traces concludes that it has no detectable leakage under t-test and that it at least can thwart the moment-correlation analysis using 665000 noiseless traces.
基金Supported by the National Natural ScienceFoundation of China (60473029)
文摘Side-channel attacks （SCA） may exploit leakage information to break cryptosystems. In this paper we present a new SCA resistant Elliptic Curve scalar multiplication algorithm. The proposed algorithm, builds a sequence of bit-strings representing the scalar k, characterized by the fact that all bit-strings are different from zero; this property will ensure a uniform computation behavior for the algorithm, and thus will make it secure against simple power analysis attacks （SPA）. With other randomization techniques, the proposed countermeasures do not penalize the computation time. The proposed scheme is more efficient than MOEller＇s one, its cost being about 5% to 10% smaller than MOEller＇s one.
基金supported by the National Natural Science Foundation of China(60373109)Ministry of Science and Technologyof China and the National Commercial Cryptography Application Technology Architecture and Application DemonstrationProject(2008BAA22B02).
文摘An embedded cryptosystem needs higher reconfiguration capability and security. After analyzing the newly emerging side-channel attacks on elliptic curve cryptosystem （ECC）, an efficient fractional width-w NAF （FWNAF） algorithm is proposed to secure ECC scalar multiplication from these attacks. This algorithm adopts the fractional window method and probabilistic SPA scheme to reconfigure the pre-computed table, and it allows designers to make a dynamic configuration on pre-computed table. And then, it is enhanced to resist SPA, DPA, RPA and ZPA attacks by using the random masking method. Compared with the WBRIP and EBRIP methods, our proposals has the lowest total computation cost and reduce the shake phenomenon due to sharp fluctuation on computation performance.
基金supported by the National Key Research and Development Program of China (2018YFB0804004)the Foundation of the National Natural Science Foundation of China (61602509)+1 种基金the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (61521003)the Key Technologies Research and Development Program of Henan Province of China (172102210615)
文摘Co-residency of virtual machines(VMs) of different tenants on the same physical platform would possibly lead to cross-VM side-channel attacks in the cloud. While most of current countermeasures fail for real or immediate deployment due to their requirement for modification of virtualization structure, we adopt dynamic migration, an inherent mechanism of the cloud platform, as a general defense against this kind of threats. To this end, we first set up a unified practical information leakage model which shows the factors affecting side channels and describes the way they influence the damage due to side-channel attacks. Since migration is adopted to limit the time duration of co-residency, we envision this defense as an optimization problem by setting up an Integer Linear Programming(ILP) to calculate optimal migration strategy, which is intractable due to high computational complexity. Therefore, we approximate the ILP with a baseline genetic algorithm, which is further improved for its optimality and scalability. Experimental results show that our migration-based defense can not only provide excellent security guarantees and affordable performance cost in both theoretical simulation and practical cloud environment, but also achieve better optimality and scalability than previous countermeasures.
文摘The number and creativity of side channel attacks have increased dramatically in recent years. Of particular interest are attacks leveraging power line communication to 1) gather information on power consumption from the victim and 2) exfiltrate data from compromised machines. Attack strategies of this nature on the greater power grid and building infrastructure levels have been shown to be a serious threat. This project further explores this concept of a novel attack vector by creating a new type of penetration testing tool: an USB power adapter capable of remote monitoring of device power consumption and communicating through powerline communications.
文摘Side-channel attacks based on supervised learning require that the attacker have complete control over the cryptographic device and obtain a large number of labeled power traces.However,in real life,this requirement is usually not met.In this paper,an attack algorithm based on collaborative learning is proposed.The algorithm only needs to use a small number of labeled power traces to cooperate with the unlabeled power trace to realize the attack to cryptographic device.By experimenting with the DPA contest V4 dataset,the results show that the algorithm can improve the accuracy by about 20%compared with the pure supervised learning in the case of using only 10 labeled power traces.
基金supported by Universiti Kebangsaan Malaysia(No.GUP 2023-010).
文摘Mobile devices within Fifth Generation(5G)networks,typically equipped with Android systems,serve as a bridge to connect digital gadgets such as global positioning system,mobile devices,and wireless routers,which are vital in facilitating end-user communication requirements.However,the security of Android systems has been challenged by the sensitive data involved,leading to vulnerabilities in mobile devices used in 5G networks.These vulnerabilities expose mobile devices to cyber-attacks,primarily resulting from security gaps.Zero-permission apps in Android can exploit these channels to access sensitive information,including user identities,login credentials,and geolocation data.One such attack leverages“zero-permission”sensors like accelerometers and gyroscopes,enabling attackers to gather information about the smartphone’s user.This underscores the importance of fortifying mobile devices against potential future attacks.Our research focuses on a new recurrent neural network prediction model,which has proved highly effective for detecting sidechannel attacks in mobile devices in 5G networks.We conducted state-of-the-art comparative studies to validate our experimental approach.The results demonstrate that even a small amount of training data can accurately recognize 37.5%of previously unseen user-typed words.Moreover,our tap detection mechanism achieves a 92%accuracy rate,a crucial factor for text inference.These findings have significant practical implications,as they reinforce mobile device security in 5G networks,enhancing user privacy,and data protection.
基金funded by the Deanship of Scientific Research(DSR)at King Abdulaziz University,Jeddah,under Grant No.(GPIP:1074-612-2024).
文摘The surge in smishing attacks underscores the urgent need for robust,real-time detection systems powered by advanced deep learning models.This paper introduces PhishNet,a novel ensemble learning framework that integrates transformer-based models(RoBERTa)and large language models(LLMs)(GPT-OSS 120B,LLaMA3.370B,and Qwen332B)to enhance smishing detection performance significantly.To mitigate class imbalance,we apply synthetic data augmentation using T5 and leverage various text preprocessing techniques.Our system employs a duallayer voting mechanism:weighted majority voting among LLMs and a final ensemble vote to classify messages as ham,spam,or smishing.Experimental results show an average accuracy improvement from 96%to 98.5%compared to the best standalone transformer,and from 93%to 98.5%when compared to LLMs across datasets.Furthermore,we present a real-time,user-friendly application to operationalize our detection model for practical use.PhishNet demonstrates superior scalability,usability,and detection accuracy,filling critical gaps in current smishing detection methodologies.
文摘Zero-click attacks represent an advanced cybersecurity threat,capable of compromising devices without user interaction.High-profile examples such as Pegasus,Simjacker,Bluebugging,and Bluesnarfing exploit hidden vulnerabilities in software and communication protocols to silently gain access,exfiltrate data,and enable long-term surveillance.Their stealth and ability to evade traditional defenses make detection and mitigation highly challenging.This paper addresses these threats by systematically mapping the tactics and techniques of zero-click attacks using the MITRE ATT&CK framework,a widely adopted standard for modeling adversarial behavior.Through this mapping,we categorize real-world attack vectors and better understand how such attacks operate across the cyber-kill chain.To support threat detection efforts,we propose an Active Learning-based method to efficiently label the Pegasus spyware dataset in alignment with the MITRE ATT&CK framework.This approach reduces the effort of manually annotating data while improving the quality of the labeled data,which is essential to train robust cybersecurity models.In addition,our analysis highlights the structured execution paths of zero-click attacks and reveals gaps in current defense strategies.The findings emphasize the importance of forward-looking strategies such as continuous surveillance,dynamic threat profiling,and security education.By bridging zero-click attack analysis with the MITRE ATT&CK framework and leveraging machine learning for dataset annotation,this work provides a foundation for more accurate threat detection and the development of more resilient and structured cybersecurity frameworks.
基金funded by the National Key Research and Development Program of China(Grant No.2024YFE0209000)the NSFC(Grant No.U23B2019).
文摘Graph Neural Networks(GNNs)have proven highly effective for graph classification across diverse fields such as social networks,bioinformatics,and finance,due to their capability to learn complex graph structures.However,despite their success,GNNs remain vulnerable to adversarial attacks that can significantly degrade their classification accuracy.Existing adversarial attack strategies primarily rely on label information to guide the attacks,which limits their applicability in scenarios where such information is scarce or unavailable.This paper introduces an innovative unsupervised attack method for graph classification,which operates without relying on label information,thereby enhancing its applicability in a broad range of scenarios.Specifically,our method first leverages a graph contrastive learning loss to learn high-quality graph embeddings by comparing different stochastic augmented views of the graphs.To effectively perturb the graphs,we then introduce an implicit estimator that measures the impact of various modifications on graph structures.The proposed strategy identifies and flips edges with the top-K highest scores,determined by the estimator,to maximize the degradation of the model’s performance.In addition,to defend against such attack,we propose a lightweight regularization-based defense mechanism that is specifically tailored to mitigate the structural perturbations introduced by our attack strategy.It enhances model robustness by enforcing embedding consistency and edge-level smoothness during training.We conduct experiments on six public TU graph classification datasets:NCI1,NCI109,Mutagenicity,ENZYMES,COLLAB,and DBLP_v1,to evaluate the effectiveness of our attack and defense strategies.Under an attack budget of 3,the maximum reduction in model accuracy reaches 6.67%on the Graph Convolutional Network(GCN)and 11.67%on the Graph Attention Network(GAT)across different datasets,indicating that our unsupervised method induces degradation comparable to state-of-the-art supervised attacks.Meanwhile,our defense achieves the highest accuracy recovery of 3.89%(GCN)and 5.00%(GAT),demonstrating improved robustness against structural perturbations.
基金supported by Key Laboratory of Cyberspace Security,Ministry of Education,China。
文摘Transformer-based models have significantly advanced binary code similarity detection(BCSD)by leveraging their semantic encoding capabilities for efficient function matching across diverse compilation settings.Although adversarial examples can strategically undermine the accuracy of BCSD models and protect critical code,existing techniques predominantly depend on inserting artificial instructions,which incur high computational costs and offer limited diversity of perturbations.To address these limitations,we propose AIMA,a novel gradient-guided assembly instruction relocation method.Our method decouples the detection model into tokenization,embedding,and encoding layers to enable efficient gradient computation.Since token IDs of instructions are discrete and nondifferentiable,we compute gradients in the continuous embedding space to evaluate the influence of each token.The most critical tokens are identified by calculating the L2 norm of their embedding gradients.We then establish a mapping between instructions and their corresponding tokens to aggregate token-level importance into instructionlevel significance.To maximize adversarial impact,a sliding window algorithm selects the most influential contiguous segments for relocation,ensuring optimal perturbation with minimal length.This approach efficiently locates critical code regions without expensive search operations.The selected segments are relocated outside their original function boundaries via a jump mechanism,which preserves runtime control flow and functionality while introducing“deletion”effects in the static instruction sequence.Extensive experiments show that AIMA reduces similarity scores by up to 35.8%in state-of-the-art BCSD models.When incorporated into training data,it also enhances model robustness,achieving a 5.9%improvement in AUROC.
基金supported by the Institute of Information&Communications Technology Planning&Evaluation(IITP)grant funded by the Korea government(MSIT)(No.RS-2023-00235509Development of security monitoring technology based network behavior against encrypted cyber threats in ICT convergence environment).
文摘With the increasing emphasis on personal information protection,encryption through security protocols has emerged as a critical requirement in data transmission and reception processes.Nevertheless,IoT ecosystems comprise heterogeneous networks where outdated systems coexist with the latest devices,spanning a range of devices from non-encrypted ones to fully encrypted ones.Given the limited visibility into payloads in this context,this study investigates AI-based attack detection methods that leverage encrypted traffic metadata,eliminating the need for decryption and minimizing system performance degradation—especially in light of these heterogeneous devices.Using the UNSW-NB15 and CICIoT-2023 dataset,encrypted and unencrypted traffic were categorized according to security protocol,and AI-based intrusion detection experiments were conducted for each traffic type based on metadata.To mitigate the problem of class imbalance,eight different data sampling techniques were applied.The effectiveness of these sampling techniques was then comparatively analyzed using two ensemble models and three Deep Learning(DL)models from various perspectives.The experimental results confirmed that metadata-based attack detection is feasible using only encrypted traffic.In the UNSW-NB15 dataset,the f1-score of encrypted traffic was approximately 0.98,which is 4.3%higher than that of unencrypted traffic(approximately 0.94).In addition,analysis of the encrypted traffic in the CICIoT-2023 dataset using the same method showed a significantly lower f1-score of roughly 0.43,indicating that the quality of the dataset and the preprocessing approach have a substantial impact on detection performance.Furthermore,when data sampling techniques were applied to encrypted traffic,the recall in the UNSWNB15(Encrypted)dataset improved by up to 23.0%,and in the CICIoT-2023(Encrypted)dataset by 20.26%,showing a similar level of improvement.Notably,in CICIoT-2023,f1-score and Receiver Operation Characteristic-Area Under the Curve(ROC-AUC)increased by 59.0%and 55.94%,respectively.These results suggest that data sampling can have a positive effect even in encrypted environments.However,the extent of the improvement may vary depending on data quality,model architecture,and sampling strategy.
基金supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project Number(PNURSP2025R97)Princess Nourah bint Abdulrahman University,Riyadh,Saudi Arabia.
文摘The exponential growth of the Internet of Things(IoT)has introduced significant security challenges,with zero-day attacks emerging as one of the most critical and challenging threats.Traditional Machine Learning(ML)and Deep Learning(DL)techniques have demonstrated promising early detection capabilities.However,their effectiveness is limited when handling the vast volumes of IoT-generated data due to scalability constraints,high computational costs,and the costly time-intensive process of data labeling.To address these challenges,this study proposes a Federated Learning(FL)framework that leverages collaborative and hybrid supervised learning to enhance cyber threat detection in IoT networks.By employing Deep Neural Networks(DNNs)and decentralized model training,the approach reduces computational complexity while improving detection accuracy.The proposed model demonstrates robust performance,achieving accuracies of 94.34%,99.95%,and 87.94%on the publicly available kitsune,Bot-IoT,and UNSW-NB15 datasets,respectively.Furthermore,its ability to detect zero-day attacks is validated through evaluations on two additional benchmark datasets,TON-IoT and IoT-23,using a Deep Federated Learning(DFL)framework,underscoring the generalization and effectiveness of the model in heterogeneous and decentralized IoT environments.Experimental results demonstrate superior performance over existing methods,establishing the proposed framework as an efficient and scalable solution for IoT security.
基金upported by the National Natural Science Foundation of China (No. 61472292)the Technological Innovation of Hubei Province (No. 2018AAA046)the Key Technology Research of New-Generation HighSpeed and High-Level Security Chip for Smart Grid (No. 526816160015)
文摘Existing Side-Channel Attacks （SCAs） have several limitations and, rather than to be real attack methods, can only be considered to be security evaluation methods. Their limitations are mainly related to the sampling conditions, such as the trigger signal embedded in the source code of the encryption device, and the acquisition device that serves as the encryption-device controller. Apart from it being very difficult for an attacker to add a trigger into the original design before making an attack or to control the encryption device, there is a big gap in the capacity of existing SCAs to pose real threats to cipher devices. In this paper, we propose a new method, the sliding window SCA （SW-SCA）, which can be applied in scenarios in which the acquisition device is independent of the encryption device and for which the encryption source code requires no trigger signal or modification. First, we describe the main issues in existing SCAs, then we theoretically analyze the effectiveness and complexity of our proposed SW-SCA --a method that can incorporate a sliding-window mechanism into almost all of the existing non-profiled SCAs. The experimental results for both simulated and physical traces verify the effectiveness of the SW-SCA and the appropriateness of its theoretical complexity.
文摘The security of cryptographic systems is a major concern for cryptosystem designers, even though cryptography algorithms have been improved. Side-channel attacks, by taking advantage of physical vulnerabilities of cryptosystems, aim to gain secret information. Several approaches have been proposed to analyze side-channel information, among which machine learning is known as a promising method. Machine learning in terms of neural networks learns the signature （power consumption and electromagnetic emission） of an instruction, and then recognizes it automatically. In this paper, a novel experimental investigation was conducted on field-programmable gate array （FPGA） implementation of elliptic curve cryptography （ECC）, to explore the efficiency of side-channel information characterization based on a learning vector quantization （LVQ） neural network. The main characteristics of LVQ as a multi-class classifier are that it has the ability to learn complex non-linear input-output relationships, use sequential training procedures, and adapt to the data. Experimental results show the performance of multi-class classification based on LVQ as a powerful and promising approach of side-channel data characterization.
文摘Fault attacks have emerged as an increasingly effective approach for integrated circuit security attacks due to their short execution time and minimal data requirement.However,the lack of a unified leakage model remains a critical challenge,as existing methods often rely on algorithm-specific details or prior knowledge of plaintexts and intermediate values.This paper proposes the Fault Probability Model based on Hamming Weight(FPHW)to address this.This novel statistical framework quantifies fault attacks by solely analyzing the statistical response of the target device,eliminating the need for attack algorithm details or implementation specifics.Building on this model,a Fault Injection Attack method based on Mutual Information(FPMIA)is introduced,which recovers keys by leveraging the mutual information between measured fault probability traces and simulated leakage derived from Hamming weight,reducing data requirements by at least 44%compared to the existing Mutual Information Analysis method while achieving a high correlation coefficient of 0.9403 between measured and modeled fault probabilities.Experimental validation on an AES-128 implementation via a Microcontroller Unit demonstrates that FPHW accurately captures the data dependence of fault probability and FPMIA achieves efficient key recovery with robust noise tolerance,establishing a unified and efficient framework that surpasses traditional methods in terms of generality,data efficiency,and practical applicability.
基金supported by the National Key R&D Program of China (No.2018YFA0704703)National Natural Science Foundation of China (Nos.61972215,61972073,and 62172238)Natural Science Foundation of Tianjin (No.20JCZDJC00640).
文摘In cloud storage,client-side deduplication is widely used to reduce storage and communication costs.In client-side deduplication,if the cloud server detects that the user’s outsourced data have been stored,then clients will not need to reupload the data.However,the information on whether data need to be uploaded can be used as a side-channel,which can consequently be exploited by adversaries to compromise data privacy.In this paper,we propose a new threat model against side-channel attacks.Different from existing schemes,the adversary could learn the approximate ratio of stored chunks to unstored chunks in outsourced files,and this ratio will affect the probability that the adversary compromises the data privacy through side-channel attacks.Under this threat model,we design two defense schemes to minimize privacy leakage,both of which design interaction protocols between clients and the server during deduplication checks to reduce the probability that the adversary compromises data privacy.We analyze the security of our schemes,and evaluate their performances based on a real-world dataset.Compared with existing schemes,our schemes can better mitigate data privacy leakage and have a slightly lower communication cost.
基金the National Natural Science Foundation of China (Nos. 61521003 and 61602509)the National Key Research and Development Program of China (Nos. 2016YFB0800100 and 2016YFB0800101)the Key Technologies Research and Development Program of Henan Province of China (No. 172102210615).
文摘Co-residency of different tenants’ virtual machines(VMs) in cloud provides a good chance for side-channel attacks, which results in information leakage. However, most of current defense suffers from the generality or compatibility problem, thus failing in immediate real-world deployment. VM migration, an inherit mechanism of cloud systems, envisions a promising countermeasure, which limits co-residency by moving VMs between servers. Therefore, we first set up a unified practical adversary model, where the attacker focuses on effective side channels. Then we propose Driftor, a new cloud system that contains VMs of a multi-executor structure where only one executor is active to provide service through a proxy, thus reducing possible information leakage. Active state is periodically switched between executors to simulate defensive effect of VM migration. To enhance the defense, real VM migration is enabled at the same time. Instead of solving the migration satisfiability problem with intractable CIRCUIT-SAT, a greedy-like heuristic algorithm is proposed to search for a viable solution by gradually expanding an initial has-to-migrate set of VMs. Experimental results show that Driftor can not only defend against practical fast side-channel attack, but also bring about reasonable impacts on real-world cloud applications.
