On Fingerprinting Attacks and Length-Hiding Encryption

Website Fingerprinting (WF) attacks have posed a serious threat to the anonymity of the onion router (Tor) communication system, as attackers can passively pry into the encrypted traffic and infer the website visited by users. To defend against WF, recent studies focus on adversarial perturbations. However, most of them suffer from a high bandwidth overhead and a low defense performance. To address this problem, our basic idea is to generate perturbation only on the sensitive regions, which can effectively mask the website’s fingerprint, thus misleading the WF attack models and reducing the bandwidth overhead. In this paper, we formulate a joint optimization problem of perturbation position and magnitude by confining the perturbations within sensitive regions, which is rarely considered in the literature. We propose a robust low-overhead WF defense method based on reinforcement learning (RL), named RLpatch. RLpatch identifies the common sensitive regions of various surrogate models and adjusts perturbation according to the query result from a query WF model. It further employs the positional frequency of perturbations to generate a common perturbation paradigm for different traces of a same website. Experimental results show that RLpatch achieves higher defense performance, lower bandwidth overhead and better robustness against adversarial training compared to the state-of-the-art methods.

Website Fingerprinting (WF) attacks enable passive adversaries to identify the website a user visits over encrypted or anonymized network connections. WF attacks based on deep learning have achieved high accuracy in identifying websites based on abundant training traffic traces per website. However, collecting large-scale and fresh traces is quite cost-consuming and unrealistic. Morevoer, these deep-learning-based WF attacks lack flexibility because they require a long bootstrap time for retraining when facing new traffic traces with different distributions or newly added monitored websites. This paper proposes a high-accuracy WF attack named Contrastive Fingerprinting (CF), which leverages contrastive learning and data augmentation over a few training traces. The results of extensive experiments on challenging datasets over few-shot traces demonstrate the high accuracy of the CF attack and its robustness against WF defenses. For example, when each monitored website only has 20 training traces, CF identifies monitored websites with a high accuracy of 90.4% in the closed-world scenario and distinguishes monitored websites with a high True Positive Rate of 91.2% in the open-world scenario. The experimental results also show that CF outperforms two existing WF attacks with few-shot traces under different network conditions in real-world applications.

Although Tor aims to ensure anonymity, website fingerprinting (WF) attacks remain a critical threat to user privacy. Existing defenses, such as randomization and adversarial learning-based methods, face challenges including high bandwidth overhead, reliance on complete traffic traces, and limited adaptability. To address these limitations, we propose D-FOREST, a novel WF defense framework that leverages an enhanced Decision Transformer (DT)-based approach with online optimization capabilities. This method enables the DT, a reinforcement learning (RL)-inspired sequence model, to dynamically adapt its traffic perturbation strategy to evolving network conditions through real-time fine-tuning. Unlike prior adversarial learning approaches, D-FOREST utilizes the DT's autoregressive future-conditioned modeling to incrementally generate context-aware perturbations synchronized with real-time traffic bursts. This design eliminates dependence on complete traffic traces while ensuring robust adaptability. By framing perturbation generation as a sequence decision-making task, DFOREST optimizes the trade-off between defense effectiveness and bandwidth consumption through a combination of offline trajectory modeling and online adaptation. Experiments on synthetic datasets demonstrate that D-FOREST reduces WF attack accuracy to <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 5 \% - 2 0 \%}$</tex> with under <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{2 0 \%}$</tex> bandwidth overhead (BWO), outperforming existing methods. Meanwhile, D-FOREST demonstrates superior real-world adaptability compared to prior defense methods, maintaining stable protection efficacy throughout one month of continuous traffic while effectively eliminating catastrophic F1-score failures (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$&gt;50 \%$</tex>).

Website fingerprinting (WF) attacks pose a threat to privacy of web activity, especially on anonymity networks such as Tor. Recent studies show that the deep neural network (DNN) significantly improves the impact of website fingerprinting attacks. Especially, DNN-based attack undermines the existing defense methods which are mainly rely on the manually designed rule. In this paper, we present a novel defense that generates universal perturbation that can transform original examples to adversarial examples which is effectively defending against a specific WF model. The proposed defense is evaluated on state-of-the-art DNN attack over a public Tor traffic dataset. The experimental results show our adversarial example generation method performs better than the baseline methods. The proposed defense defeats all existing WF attacks based on deep neural networks with a low overhead. Comparing with state-of-the-art defenses such as Walkie-Talkie and WTF-PAD with a lower bound of 31% and 64% overheads, the proposed defense achieves identical defense performance with at least 50% bandwidth overhead saving.

Website Fingerprinting (WF) attacks based on Deep Neural Networks (DNN) pose serious threats to user privacy in anonymous communication systems such as Tor. Existing defenses show limited effectiveness in black-box scenarios, facing challenges including poor perturbation transferability, lack of directional and real-time constraints, limited universality, and high overhead. To address these issues, this paper proposes HAAD (Heuristic Ant-based Adversarial Defense), a black-box defense framework based on an ant colony algorithm. HAAD generates perturbations by combining an ensemble substitute model with ant colony optimization, avoiding reliance on gradient information and enhancing transferability to unknown models. Additionally, a per-packet injection method with directional constraints is designed to achieve real-time defense without extra latency. A universal perturbation algorithm is introduced, optimizing perturbations for diverse traffic scenarios via a scoring mechanism, effectively reducing communication overhead. Experimental results show that under various adversarial attack models, HAAD achieves a Defense Success Rate (DSR) exceeding 90% while requiring only 30 injected dummy packets—incurring minimal overhead. When deployed on a P4 switch, the maximum end-to-end latency is only 200 nanoseconds. We further implement HAAD in the Tor pt-plugin, where the full-page load median latency is merely 0.02 seconds. These results highlight HAAD's strong defensive capability and practical applicability.

One of the primary filtering methods that the Great Firewall of China (GFW) relies on is poisoning DNS responses for certain domains. When a DNS request is poisoned by the GFW, multiple DNS responses are received - both legitimate and poisoned responses. While most prior research into the GFW focuses on the poisoned responses, ours also considers the legitimate responses from the DNS servers themselves. We find that even when we ignored the immediate poisoned responses, the cache from the DNS servers themselves are also poisoned. We also find and discuss the IP addresses within the DNS responses we get; in particular 9 IP addresses that are returned as a result for many different poisoned domains. We present the argument that this type of attack may not be primarily targeted directly at users, but at the underlying DNS infrastructure within China.

Website Fingerprinting (WF) attacks pose a serious threat to users' online privacy, including for users of the Tor anonymity system. By exploiting recent advances in deep learning, WF attacks like Deep Fingerprinting (DF) have reached up to 98% accuracy. The DF attack, however, requires large amounts of training data that needs to be updated regularly, making it less practical for the weaker attacker model typically assumed in WF. Moreover, research on WF attacks has been criticized for not demonstrating attack effectiveness under more realistic and more challenging scenarios. Most research on WF attacks assumes that the testing and training data have similar distributions and are collected from the same type of network at about the same time. In this paper, we examine how an attacker could leverage N-shot learning---a machine learning technique requiring just a few training samples to identify a given class---to reduce the effort of gathering and training with a large WF dataset as well as mitigate the adverse effects of dealing with different network conditions. In particular, we propose a new WF attack called Triplet Fingerprinting (TF) that uses triplet networks for N-shot learning. We evaluate this attack in challenging settings such as where the training and testing data are collected multiple years apart on different networks, and we find that the TF attack remains effective in such settings with 85% accuracy or better. We also show that the TF attack is also effective in the open world and outperforms traditional transfer learning. On top of that, the attack requires only five examples to recognize a website, making it dangerous in a wide variety of scenarios where gathering and training on a complete dataset would be impractical.

With the widespread popularity of smart home devices and the emergence of smart home integration platforms such as Google, Amazon, and Xiaomi, the smart home industry is in a stage of vigorous development. While smart homes provide users with convenient and intelligent living, the problem of smart home devices leaking user privacy has become increasingly prominent. Smart home devices give users the ability to remotely control home devices, but they also reflect user home activities in traffic data, which brings the risk of privacy leaks. Potential attackers can use traffic classification technology to analyze traffic characteristics during traffic transmission (e.g., at the traffic exit of a smart home gateway) and infer users’ private information, such as their home activities, causing serious consequences of privacy leaks. To address the above problems, this paper focuses on research on privacy protection technology based on traffic obfuscation. By using traffic obfuscation technology to obscure the true traffic of smart home devices, it can prevent malicious traffic listeners from analyzing user privacy information based on traffic characteristics. We propose an enhanced smart home traffic obfuscation method called SHTObfuscator (Smart Home Traffic Obfuscator) based on the virtual user technology concept and a virtual user behavior construction method based on logical integrity. By injecting traffic fingerprints of different device activities into the real traffic environment of smart homes as obfuscating traffic, attackers cannot distinguish between the real device working status and user behavior privacy in the current home, effectively reducing the effect of traffic classification attack models. The protection level can be manually or automatically adjusted, achieving a balance between privacy protection and bandwidth overhead. The experimental results show that under the highest obfuscation level, the obfuscation method proposed in this paper can effectively reduce the classification effect of the attack model from 95% to 25%.

Fully homomorphic encryption (FHE) allows us to perform computations directly over encrypted data and can be widely used in some highly regulated industries. Gentry’s bootstrapping procedure is used to refresh noisy ciphertexts and is the only way to achieve the goal of FHE up to now. In this paper, we optimize the LWE-based GSW-type bootstrapping procedure. Our optimization decreases the lattice approximation factor for the underlying worst-case lattice assumption from \(\tilde{O}(N^{2.5})\) to \(\tilde{O}(N^{2})\), and is time-efficient by a \(O(\lambda )\) factor. Our scheme can also achieve the best factor in prior works on bootstrapping of standard lattice-based FHE by taking a larger lattice dimension, which makes our scheme as secure as the standard lattice-based PKE. Furthermore, in this work we present a technique to perform more operations per bootstrapping in the LWE-based FHE scheme. Although there have been studies to evaluate large FHE gates using schemes over ideal lattices, (i.e. using FHEW or TFHE), we are the first to study how to perform complex functions homomorphically over standard lattices.

We revisit the problem of privacy-preserving range search and sort queries on encrypted data in the face of an untrusted data store. Our new protocol RASP has several advantages over existing work. First, RASP strengthens privacy by ensuring forward security: after a query for range [a, b], any new record added to the data store is indistinguishable from random, even if the new record falls within range [a, b]. We are able to accomplish this using only traditional hash and block cipher operations, abstaining from expensive asymmetric cryptography and bilinear pairings. Consequently, RASP is highly practical, even for large database sizes. Additionally, we require only cloud storage and not a computational cloud like related works, which can reduce monetary costs significantly. At the heart of RASP, we develop a new update-oblivious bucket-based data structure. We allow for data to be added to buckets without leaking into which bucket it has been added. As long as a bucket is not explicitly queried, the data store does not learn anything about bucket contents. Furthermore, no information is leaked about data additions following a query. Besides formally proving RASP’s privacy, we also present a practical evaluation of RASP on Amazon Dynamo, demonstrating its efficiency and real world applicability.

To secure cloud storage and enforce access control, data encryption has become essential, given the ever increasing cyber threat everywhere. Attribute-based Encryption (ABE) crypto systems are widely considered as a promising solution under such a context for its security strength, scalability and control flexibility. One major challenge, however, for applying ABE-based techniques in real world applications is its high overhead in various aspects. In this research, we are particularly concerned with the storage size expansion in existing ABE schemes. This combined with the vast-size nature of the cloud data poses an enormous challenge to the effective usage of the cloud data storage space and affects the utility of data deduplication. Normally, data deduplication is carried out based on identifying similar and even identical contents both within and between data files, however, these patterns will be destroyed after performing data encryption using any semantically secure encryption scheme including ABE. In this research, we focus on ciphertexts deduplication under ABE, which to our best knowledge is the first of such an effort. Our fundamental observation stems from the structure of ABE ciphertexts and the possible similarities among different access structures. We show how to design a secure ciphertext deduplication scheme based on a classical CP-ABE scheme by innovatively modifying the construction with a recursive algorithm, eliminating the duplicated secrets and adding additional randomness to some certain ciphertext. We then give a detailed analysis on the proposed scheme with respect to both efficiency and security. To thoroughly assess the performance of the proposed scheme, we also implement a prototype system and conduct comprehensive experiments, which shows that our ciphertext reduplication scheme could reduce up to 80% computation and storage cost in the best case.

Website fingerprinting attacks attempt to apply deep learning technology to identify websites corresponding to encrypted traffic data. Unfortunately, to the best of our knowledge, once the total number of encrypted traffic data becomes insufficient, the identification accuracy in most existing works will drop dramatically. This phenomenon grows worse because the statistical features of the encrypted traffic data are not always stable but irregularly varying in different time periods. Even a deep learning model requires good performance to capture the statistical features, its accuracy usually diminishes in a short period of time because the changes of the statistical features technically put the training and testing data into two non-identical distributions. In this paper, we first propose a convolutional neural network-based website fingerprinting attack (CWFA) scheme. This scheme integrates packet direction with the timing sequence from the encrypted traffic data to improve the accuracy of analysis as much as possible on few data samples. We then design a new fine-tuning mechanism for the CWFA (FM-CWFA) scheme based on transfer learning. This mechanism enables the proposed FM-CWFA scheme to support the changes in the statistical patterns. The experimental results in closed-world and open-world settings show that the effectiveness of the CWFA scheme is better than previous researches, with the slowest performance degradation when the number of data decreases, and the FM-CWFA scheme can remain effective when the statistical features change.

Uncertain data streams can have tuples with both value and existential uncertainty. A tuple has value uncertainty when it can assume multiple possible values. A tuple is existentially uncertain when the sum of the probabilities of its possible values is $$<$$ 1. A situation where existential uncertainty can arise is when applying relational operators to streams with value uncertainty. Several prior works have focused on querying and mining data streams with both value and existential uncertainty. However, none of them have studied, in depth, the implications of existential uncertainty on sliding window processing, even though it naturally arises when processing uncertain data. In this work, we study the challenges arising from existential uncertainty, more specifically the management of count-based sliding windows, which are a basic building block of stream processing applications. We extend the semantics of sliding window to define the novel concept of uncertain sliding windows and provide both exact and approximate algorithms for managing windows under existential uncertainty. We also show how current state-of-the-art techniques for answering similarity join queries can be easily adapted to be used with uncertain sliding windows. We evaluate our proposed techniques under a variety of configurations using real data. The results show that the algorithms used to maintain uncertain sliding windows can efficiently operate while providing a high-quality approximation in query answering. In addition, we show that sort-based similarity join algorithms can perform better than index-based techniques (on 17 real datasets) when the number of possible values per tuple is low, as in many real-world applications.

In this paper, we study distributed solutions for joint power-spectrum resource allocation among delay sensitive users over multi-carrier networks. Our work differs from prior research in two ways. First, unlike prior works that only consider the impact of effective throughput of a user, our work also considers the source traffic characteristics of applications and applies queuing analysis to quantify the packet transmission delay, which is especially important to a delay sensitive user. Secondly, most prior works focus on the equilibrium of the decentralized solutions that usually require global information in a static network. Instead, we focus on interactive adaptation methods for users to dynamically maximize their expected utility based on the local observed information. We propose an interactive learning framework for users to adapt their power/channel selection to the wireless networks with interference coupling among users. Our simulation results show that the proposed interactive learning approach with information exchange among users improves the total utility significantly.

This study examines the role of perceived feedback and motivations on the use for contribution in Knowledge Management Systems (KMS). An individual expects responses from others when he or she posts the knowledge or questions of soliciting knowledge in KMS. If an individual posts a message and there are many responses, motivation to use KMS will develop. Prior research has considered feedback a potential factor that would affect an individual's motivation, such as self-efficacy, but there is few prior works that have tested empirically on how feedback can influence an individual's motivations as well as the use for contribution in KMS. Perceived feedback can be categorized into feedback on posted knowledge and feedback on knowledge request postings. The results of the study show that motivations such as organizational reward, knowledge self-efficacy, and enjoyment of helping, except for social reward, are associated with the use for contribution. Our results also show that among perceived feedback, feedback on posted knowledge affects motivation and use for contribution while feedback on knowledge request postings directly influences the use for contribution.