• Formulates malware classification as a Markov Decision Process with episodic feature acquisition, achieving superior performance across diverse datasets: 99.20% F1-score on Microsoft Big2015, 98.64% on BODMAS, and 85.07% on EMBER 2018 using reinforcement learning. • Demonstrates systematic superiority over traditional approaches through comprehensive ablation studies, where static feature selection methods exhibit severe performance degradation (up to 10.40% F1-score reduction) while D3QN maintains consistent improvements across all evaluation scenarios. • Validates robust transferability with 76.08% average recall on unseen EMBER 2024 malware variants across six diverse file formats, demonstrating 27.55% relative improvement over traditional methods and effective zero-day threat detection capabilities. • Introduces quantitative intelligence assessment framework proving strategic learning behavior with 62.5% categorical preference deviation from random baselines, 57.7% feature specialization, and autonomous discovery of domain-aligned cybersecurity patterns without explicit supervision. Traditional malware detection methods exhibit computational inefficiency due to exhaustive feature extraction requirements, creating accuracy-efficiency trade-offs that limit real-time deployment. We formulate malware classification as a Markov Decision Process with episodic feature acquisition and propose a Dueling Double Deep Q-Network (D3QN) framework for adaptive sequential feature selection. The agent learns to dynamically explore informative features per sample before terminating with classification decisions, optimizing both detection accuracy and computational cost through reinforcement learning. We evaluate our approach on Microsoft Big2015 (9-class, 1795 features), BODMAS and EMBER 2018 (binary, 2381 features) datasets. D3QN achieves 99.20%, 98.64%, and 85.07% F1-scores respectively while utilizing approximately 60 features on average, representing 96.6% and 97.5% dimensionality reduction compared to full feature sets. Comprehensive ablation studies across six feature selection methods demonstrate that traditional approaches suffer severe performance degradation (averaging 1.85-10.40% F1-score reduction) when constrained to comparable feature subsets, while D3QN maintains consistent improvements (+1.38% to +5.08%) across all evaluation scenarios. Cross-dataset transferability validation on EMBER 2024 demonstrates superior zero-day detection capabilities, achieving 76.08% average recall on unseen malware variants across diverse file formats–representing 27.55% relative improvement over traditional methods. Quantitative intelligence assessment reveals strategic learning behavior with 62.5% categorical preference deviation from random baselines and 57.7% feature specialization. The learned policies exhibit autonomous discovery of domain-aligned patterns, identifying structural anomaly indicators and behavioral signatures characteristic of cybersecurity expertise. Our results validate reinforcement learning-based sequential feature selection for malware classification, achieving superior accuracy with substantial computational reduction through learned adaptive policies that outperform static dimensionality reduction techniques across diverse threat landscapes.

Federated learning (FL) enables collaborative model training using decentralized private data from multiple clients. While FL has shown robustness against poisoning attacks with basic defenses, our research reveals new vulnerabilities stemming from non-independent and identically distributed (non-IID) data among clients. These vulnerabilities pose a substantial risk of model poisoning in real-world FL scenarios.To demonstrate such vulnerabilities, we develop a novel collaborative backdoor poisoning attack called CollaPois. In this attack, we distribute a single pre-trained model infected with a Trojan to a group of compromised clients. These clients then work together to produce malicious gradients, causing the FL model to consistently converge towards a low-loss region centered around the Trojan-infected model. Consequently, the impact of the Trojan is amplified, especially when the benign clients have diverse local data distributions and scattered local gradients. CollaPois stands out by achieving its goals while involving only a limited number of compromised clients, setting it apart from existing attacks. Also, CollaPois effectively avoids noticeable shifts or degradation in the FL model’s performance on legitimate data samples, allowing it to operate stealthily and evade detection by advanced robust FL algorithms.Thorough theoretical analysis and experiments conducted on various benchmark datasets demonstrate the superiority of CollaPois compared to state-of-the-art backdoor attacks. Notably, CollaPois bypasses existing backdoor defenses, especially in scenarios where clients possess diverse data distributions. Moreover, the results show that CollaPois remains effective even when involving a small number of compromised clients. Notably, clients whose local data is closely aligned with compromised clients experience higher risks of backdoor infections.

Cross-domain generalization remains a critical challenge in deepfake detection, with existing methods exhibiting severe performance degradation across unseen generative architectures. We propose CAMME (Cross-domain Adaptive Multi-Modal Embeddings), which dynamically integrates visual, textual, and frequency-domain features via embedding-level multi-modal self-attention. Treating each modality as a distinct sequence element enables cross-modal interactions that adaptively weight discriminative features based on input characteristics. Unlike static fusion approaches, CAMME learns input-specific contributions, dynamically emphasizing optimal signals across visual semantics, textual consistency, and spectral artifacts. Evaluation across twelve generative architectures demonstrates superior cross-domain performance: 77.34% average F1-score on natural scenes (7.30% improvement) and 66.46% on facial datasets (13.25% improvement). CAMME exhibits exceptional robustness with 14.7% Attack Success Rate against seven prominent adversarial attacks (4-6 × improvement) and 96.63% accuracy under natural perturbations. Ablation results confirm the importance of each modality and the effectiveness of our inter-modal attention over standard fusion methods.

State-of-the-art methods have demonstrated exceptional performance in detecting deepfakes, particularly when the entire image content is generated by text-to-image generation models. However, significant challenges remain in the cases of instructional image editing, where AI models conditionally generate content based on both a real image and an edit prompt. In such scenarios, the generated content closely resembles the original image, making it more difficult for detection systems to identify deepfakes. In this paper, we propose a novel multimodal capsule network designed to address the detection of deepfake image editing. Specifically, low-level capsules from multiple modalities are integrated to predict capsules in subsequent layers. High-level capsules compete to select relevant low-level capsules, effectively aggregating local features to detect manipulated entities. The proposed approach is evaluated on diverse datasets, including natural images from real-world scenarios. Experimental results demonstrate that our model significantly outperforms state-of-the-art methods, achieving a substantial 10% improvement in Three-turn Edits or 20% improvement in Open Images Edits. Ablation studies further validate the robustness of the network, achieving detection rates exceeding 94% against natural perturbations and over 91% recall against white-box and black-box attacks. Additionally, the model effectively adapts to unseen image editing datasets, highlighting its ability to generalize across diverse and previously unencountered editing scenarios.

Internet miscreants increasingly utilize short-lived disposable domains to launch various attacks. Existing detection mechanisms are either too late to catch such malicious domains due to limited information and their short life spans or unable to catch them due to evasive techniques such as cloaking and captcha. In this work, we investigate the possibility of detecting malicious domains early in their life cycle using a content-agnostic approach. We observe that attackers often reuse or rotate hosting infrastructures to host multiple malicious domains due to increased utilization of automation and economies of scale. Thus, it gives defenders the opportunity to monitor such infrastructure to identify newly hosted malicious domains. However, such infrastructures are often shared hosting environments where benign domains are also hosted, which could result in a prohibitive number of false positives. Therefore, one needs innovative mechanisms to better distinguish malicious domains from benign ones even when they share hosting infrastructures. In this work, we build MANTIS, a highly accurate practical system that not only generates daily blocklists of malicious domains but also is able to predict malicious domains on-demand. We design a network graph based on the hosting infrastructure that is accurate and generalizable over time. Consistently, our models achieve a precision of 99.7%, a recall of 86.9% with a very low false positive rate (FPR) of 0.1 % and on average detects 19K new malicious domains per day, which is over 5 times the new malicious domains flagged daily in VirusTotal. Further, MANTIS predicts malicious domains days to weeks before they appear in popular blocklists.

The capability of generating high-quality source code using large language models (LLMs) reduces software development time and costs. However, they often introduce security vulnerabilities due to training on insecure open-source data. This highlights the need for ensuring secure and functional code generation. This paper introduces PromSec, an algorithm for prom optimization for secure and functioning code generation using LLMs. In PromSec, we combine 1) code vulnerability clearing using a generative adversarial graph neural network, dubbed as gGAN, to fix and reduce security vulnerabilities in generated codes and 2) code generation using an LLM into an interactive loop, such that the outcome of the gGAN drives the LLM with enhanced prompts to generate secure codes while preserving their functionality. Introducing a new contrastive learning approach in gGAN, we formulate code-clearing and generation as a dual-objective optimization problem, enabling PromSec to notably reduce the number of LLM inferences. PromSec offers a cost-effective and practical solution for generating secure, functional code. Extensive experiments conducted on Python and Java code datasets confirm that PromSec effectively enhances code security while upholding its intended functionality. Our experiments show that while a state-of-the-art approach fails to address all code vulnerabilities, PromSec effectively resolves them. Moreover, PromSec achieves more than an order-of-magnitude reduction in operation time, number of LLM queries, and security analysis costs. Furthermore, prompts optimized with PromSec for a certain LLM are transferable to other LLMs across programming languages and generalizable to unseen vulnerabilities in training. This study is a step in enhancing the trustworthiness of LLMs for secure and functional code generation, supporting their integration into real-world software development.

This paper introduces SGCode, a flexible prompt-optimizing system to generate secure code with large language models (LLMs). SGCode integrates recent prompt-optimization approaches with LLMs in a unified system accessible through front-end and back-end APIs, enabling users to 1) generate secure code, which is free of vulnerabilities, 2) review and share security analysis, and 3) easily switch from one prompt optimization approach to another, while providing insights on model and system performance. We populated SGCode on an AWS server with PromSec, an approach that optimizes prompts by combining an LLM and security tools with a lightweight generative adversarial graph neural network to detect and fix security vulnerabilities in the generated code. Extensive experiments show that SGCode is practical as a public tool to gain insights into the trade-offs between model utility, secure code generation, and system cost. SGCode has only a marginal cost compared with prompting LLMs. SGCode is available at: http://3.131.141.63:8501/.

Malicious domain detection (MDD) is an open security challenge that aims to detect if an Internet domain name is associated with cyber attacks. Many techniques have been applied to tackle this problem, among which graph neural networks (GNNs) are deemed one of the most effective approaches. GNN-based MDD employs domain name system (DNS) logs to represent Internet domains as nodes in a graph, dubbed domain maliciousness graph (DMG) and trains a GNN model to infer the maliciousness of Internet domains by leveraging the maliciousness of already identified ones. As this method heavily relies on the "publicly" accessible DNS logs to build DMGs, it creates a vulnerability for adversaries to manipulate the features and edges of their domain nodes within these graphs. The current body of literature primarily focuses on threat models that involve manipulating individual adversary (attacker) nodes. Nonetheless, adversaries usually create numerous domains to accomplish their attack objectives, aiming to reduce costs and evade detection. Hence, they aim to remain undetected across as many domains as possible. In this work, we call the attack that manipulates several nodes in the DMG concurrently a multi-instance evasion attack. To the best of our knowledge, this type of attack has not been explored in the prior art. We present both theoretical and empirical evidence to show that the existing single-instance evasion techniques for GNN-based MDDs are inadequate to launch multi-instance evasion attacks. Therefore, we propose an inference-time, multi-instance adversarial attack, dubbed MintA, against GNN-based MDD. MintA optimizes node perturbations to enhance the evasiveness of a node and its neighborhood. MintA only requires black-box access to the target model to launch the attack successfully. In other words, MintA does not require any knowledge of the MDD model’s parameters, architecture, or information on non-adversary nodes. We formulate an optimization problem that satisfies the attack objectives of MintA and devise an approximate solution for it. We evaluate MintA on a state-of-the-art GNN-based MDD technique using real-world data, and our experiments demonstrate an attack success rate of over 80%. The findings of this study serve as a cautionary note for security experts, highlighting the vulnerability of GNN-based MDD to practical attacks that can impede the effectiveness and advantages of this approach.

Trojan backdoor is a poisoning attack against neural network (NN) classifiers in which adversaries try to exploit the (highly desirable) model reuse property to implant Trojans into model parameters for backdoor breaches through a poisoned training process. To misclassify an input to a target class, the attacker activates the backdoor by augmenting the input with a predefined trigger that is only known to her/him. Most of the proposed defenses against Trojan attacks assume a white-box setup, in which the defender either has access to the inner state of NN or is able to run backpropagation through it. In this work, we propose a more practical black-box defense, dubbed TrojDef. In a black-box setup, the defender can only run forward-pass of the NN. TrojDef is motivated by the Trojan poisoned training, in which the model is trained on both benign and Trojan inputs. TrojDef tries to identify and filter out Trojan inputs (i.e., inputs augmented with the Trojan trigger) by monitoring the changes in the prediction confidence when the input is repeatedly perturbed by random noise. We derive a function based on the prediction outputs which is called the prediction confidence bound to decide whether the input example is Trojan or not. The intuition is that Trojan inputs are more stable as the misclassification only depends on the trigger, while benign inputs will suffer when augmented with noise due to the perturbation of the classification features. Through mathematical analysis, we show that if the attacker is perfect in injecting the backdoor, the Trojan infected model will be trained to learn the appropriate prediction confidence bound, which is used to distinguish Trojan and benign inputs under arbitrary perturbations. However, because the attacker might not be perfect in injecting the backdoor, we introduce a nonlinear transform to the prediction confidence bound to improve the detection accuracy in practical settings. Extensive empirical evaluations show that TrojDef significantly outperforms the-state-of-the-art defenses and is highly stable under different settings, even when the classifier architecture, the training process, or the hyperparameters change.

We propose to identify compromised mobile devices from a network administrator’s point of view. Intuitively, inadvertent users (and thus their devices) who download apps through untrustworthy markets are often lured to install malicious apps through in-app advertisements or phishing. We thus hypothesize that devices sharing similar apps would have a similar likelihood of being compromised, resulting in an association between a compromised device and its apps. We propose to leverage such associations to identify unknown compromised devices using the guilt-by-association principle. Admittedly, such associations could be relatively weak as it is hard, if not impossible, for an app to automatically download and install other apps without explicit user initiation. We describe how we can magnify such associations by carefully choosing parameters when applying graph-based inferences. We empirically evaluate the effectiveness of our approach on real datasets provided by a major mobile service provider. Specifically, we show that our approach achieves nearly 98% AUC (area under the ROC curve) and further detects as many as 6 ~ 7 times of new compromised devices not covered by the ground truth by expanding the limited knowledge on known devices. We show that the newly detected devices indeed present undesirable behavior in terms of leaking private information and accessing risky IPs and domains. We further conduct in-depth analysis of the effectiveness of graph inferences to understand the unique structure of the associations between mobile devices and their apps, and its impact on graph inferences, based on which we propose how to choose key parameters.

Graph neural networks (GNNs) are susceptible to privacy inference attacks (PIAS) given their ability to learn joint representation from features and edges among nodes in graph data. To prevent privacy leakages in GNNs, we propose a novel heterogeneous randomized response (HeteroRR) mechanism to protect nodes’ features and edges against PIAS under differential privacy (DP) guarantees, without an undue cost of data and model utility in training GNNs. Our idea is to balance the importance and sensitivity of nodes’ features and edges in redistributing the privacy budgets since some features and edges are more sensitive or important to the model utility than others. As a result, we derive significantly better randomization probabilities and tighter error bounds at both levels of nodes’ features and edges departing from existing approaches, thus enabling us to maintain high data utility for training GNNs. An extensive theoretical and empirical analysis using benchmark datasets shows that HeteroRR significantly outperforms various baselines in terms of model utility under rigorous privacy protection for both nodes’ features and edges. That enables us to defend PIAs in DP-preserving GNNs effectively.

We consider interdependent systems managed by multiple defenders that are under the threat of stepping-stone attacks. We model such systems via game-theoretic models and incorporate the effect of behavioral probability weighting that is used to model biases in human decision-making, as descended from the field of behavioral economics. We then incorporate into our framework called TASHAROK, two types of tax-based mechanisms for such interdependent security games where the central regulator incentivizes defenders to invest well in securing their assets so as to achieve the socially optimal outcome. We first show that due to the nature of our interdependent security game, no reliable tax-based mechanism can incentivize the socially optimal investment profile while maintaining a weakly balanced budget. We then show the effect of behavioral probability weighting bias on the amount of taxes paid by defenders, and prove that higher biases make defenders pay more taxes under the two mechanisms. We then explore voluntary participation in tax-based mechanisms. To evaluate our mechanisms, we use four representative real-world interdependent systems where we compare the game-theoretic optimal investments to the socially optimal investments under the two mechanisms. We show that the mechanisms yield higher decrease in the social cost for behavioral decision-makers compared to rational decision-makers.

High-quality intelligence of Internet threat (e.g., malware files, malicious domains, phishing URLs and malicious IPs) are important for both security practitioners and the research community. Given the agility of attackers, the scale of the Internet, and the fast-evolving landscape of threats, one could not rely solely on a single source (such as an anti-malware engine or an IP blacklist) for obtaining accurate, up-to-date, and comprehensive threat analysis. Instead, we need to aggregate the analysis from multiple sources. However, it is non-trivial to do such aggregation effectively. A common practice is to label an indicator (malware, domains, URLs, etc.) as malicious if it is marked by a number of sources above an ad-hoc certain threshold. Often, this results in sub-optimal performance as it assumes that all sources are of similar quality/expertise, independent, and temporally stable, which unfortunately are often not true in practice. A natural alternative is to train a supervised machine learning model. However, this approach needs a sufficiently large amount of manually labeled ground truth, which is time-consuming to collect and has to be updated frequently, resulting in substantial recurring costs. In this paper, we propose SIRAJ, a novel framework for aggregating the detection output of various intelligence sources such as anti-malware engines. SIRAJ is based on the pretrain and fine-tune paradigm. Specifically, we use self-supervised learning-based approaches to learn a pre-trained embedding model that converts multi-source inputs into a high-dimensional embedding. The embeddings are learned through three carefully designed pretext tasks that imbue them with knowledge about dependencies between scanners and their temporal dynamics. The learned embeddings could be used for diverse downstream machine learning tasks. SIRAJ is designed to be general and can be used for diverse domains such as URLs, malware, and IPs. Further, SIRAJ works well even when there is limited to no labeled data available. Through extensive experiments, we show that our learned representations can produce results comparable to supervised methods while only requiring as little as 100 labeled samples. Importantly, the results show that SIRAJ accurately detects threat indicators much earlier than the baseline algorithms, a feat that is critical against short-lived indicators like Phishing URLs.

We model the behavioral biases of human decision-making in securing interdependent systems and show that such behavioral decision-making leads to a suboptimal pattern of resource allocation compared to non-behavioral (rational) decision-making. We provide empirical evidence for the existence of such behavioral bias model through a controlled subject study with 145 participants. We then propose three learning techniques for enhancing decision-making in multi-round setups. We illustrate the benefits of our decision-making model through multiple interdependent real-world systems and quantify the level of gain compared to the case in which the defenders are behavioral. We also show the benefit of our learning techniques against different attack models. We identify the effects of different system parameters (e.g., the defenders' security budget availability and distribution, the degree of interdependency among defenders, and collaborative defense strategies) on the degree of suboptimality of security outcomes due to behavioral decision-making.

The pervasiveness of neural networks (NNs) in critical computer vision and image processing applications makes them very attractive for adversarial manipulation. A large body of existing research thoroughly investigates two broad categories of attacks targeting the integrity of NN models. The first category of attacks, commonly called Adversarial Examples, perturbs the model’s inference by carefully adding noise into input examples. In the second category of attacks, adversaries try to manipulate the model during the training process by implanting Trojan backdoors. Researchers show that such attacks pose severe threats to the growing applications of NNs and propose several defenses against each attack type individually. However, such one-sided defense approaches leave potentially unknown risks in real-world scenarios when an adversary can unify different attacks to create new and more lethal ones bypassing existing defenses.In this work, we show how to jointly exploit adversarial perturbation and model poisoning vulnerabilities to practically launch a new stealthy attack, dubbed AdvTrojan. AdvTrojan is stealthy because it can be activated only when: 1) a carefully crafted adversarial perturbation is injected into the input examples during inference, and 2) a Trojan backdoor is implanted during the training process of the model. We leverage adversarial noise in the input space to move Trojan-infected examples across the model decision boundary, making it difficult to detect. The stealthiness behavior of AdvTrojan fools the users into accidentally trusting the infected model as a robust classifier against adversarial examples. AdvTrojan can be implemented by only poisoning the training data similar to conventional Trojan backdoor attacks. Our thorough analysis and extensive experiments on several benchmark datasets show that AdvTrojan can bypass existing defenses with a success rate close to 100% in most of our experimental scenarios and can be extended to attack federated learning as well as high-resolution images.

Adversarial examples are among the biggest challenges for machine learning models, especially neural network classifiers. Adversarial examples are inputs manipulated with perturbations insignificant to humans while being able to fool machine learning models. Researchers achieve great progress in utilizing adversarial training as a defense. However, the overwhelming computational cost degrades its applicability, and little has been done to overcome this issue. Single-Step adversarial training methods have been proposed as computationally viable solutions; however, they still fail to defend against iterative adversarial examples. In this work, we first experimentally analyze several different state-of-the-art (SOTA) defenses against adversarial examples. Then, based on observations from experiments, we propose a novel single-step adversarial training method that can defend against both single-step and iterative adversarial examples. Through extensive evaluations, we demonstrate that our proposed method successfully combines the advantages of both single-step (low training overhead) and iterative (high robustness) adversarial training defenses. Compared with ATDA on the CIFAR-10 dataset, for example, our proposed method achieves a 35.67% enhancement in test accuracy and a 19.14% reduction in training time. When compared with methods that use BIM or Madry examples (iterative methods) on the CIFAR-10 dataset, our proposed method saves up to 76.03% in training time, with less than 3.78% degeneration in test accuracy. Finally, our experiments with the ImageNet dataset clearly show the scalability of our approach and its performance advantages over SOTA single-step approaches.

Ever since the beginning of the outbreak of the COVID-19 pandemic, attackers acted quickly to exploit the confusion, uncertainty and anxiety caused by the pandemic and launched various attacks through COVID-19 themed malicious domains. Malicious domains are rarely deployed independently, but rather almost always belong to much bigger and coordinated attack campaigns. Thus, analyzing COVID-themed malicious domains from the angle of attack campaigns would help us gain a deeper understanding of the scale, scope and sophistication of the threats imposed by such malicious domains. In this paper, we collect data from multiple sources, and identify and characterize COVID-themed malicious domain campaigns, including the evolution of such campaigns, their underlying infrastructures and the different strategies taken by attackers behind these campaigns. Our exploration suggests that some malicious domains have strong correlations, which can guide us to identify new malicious domains and raise alarms at the early stage of their deployment. The results shed light on the emergency for detecting and mitigating public event related cyber attacks.

Autoencoder-based anomaly detection methods have been used in identifying anomalous users from large-scale enterprise logs with the assumption that adversarial activities do not follow past habitual patterns. Most existing approaches typically build models by reconstructing single-day and individual-user behaviors. However, without capturing long-term signals and group-correlation signals, the models cannot identify low-signal yet long-lasting threats, and will wrongly report many normal users as anomalies on busy days, which, in turn, lead to high false positive rate. In this paper, we propose ACOBE, an Anomaly detection method based on COmpound BEhavior, which takes into consideration long-term patterns and group behaviors. ACOBE leverages a novel behavior representation and an ensemble of deep autoencoders and produces an ordered investigation list. Our evaluation shows that ACOBE outperforms prior work by a large margin in terms of precision and recall, and our case study demonstrates that ACOBE is applicable in practice for cyberattack detection.

End-to-end email encryption (E2EE) ensures that an email could only be decrypted and read by its intended recipients. E2EE’s strong security guarantee is particularly desirable for the enterprises in the event of breaches: even if attackers break into an email server, under E2EE no contents of emails are leaked. Meanwhile, E2EE brings significant challenges for an enterprise to detect and filter unwanted emails (spams and phishing emails). Most existing solutions rely heavily on email contents (i.e., email body and attachments), which would be difficult when email contents are encrypted. In this paper, we investigate how to detect unwanted emails in a content-agnostic manner, that is, without access to the contents of emails at all.

Malicious domains, including phishing websites, spam servers, and command and control servers, are the reason for many of the cyber attacks nowadays. Thus, detecting them in a timely manner is important to not only identify cyber attacks but also take preventive measures. There has been a plethora of techniques proposed to detect malicious domains by analyzing Domain Name System (DNS) traffic data. Traditionally, DNS acts as an Internet miscreant’s best friend, but we observe that the subtle traces in DNS logs left by such miscreants can be used against them to detect malicious domains. Our approach is to build a set of domain graphs by connecting “related” domains together and injecting known malicious and benign domains into these graphs so that we can make inferences about the other domains in the domain graphs. A key challenge in building these graphs is how to accurately identify related domains so that incorrect associations are minimized and the number of domains connected from the dataset is maximized. Based on our observations, we first train two classifiers and then devise a set of association rules that assist in linking domains together. We perform an in-depth empirical analysis of the graphs built using these association rules on passive DNS data and show that our techniques can detect many more malicious domains than the state-of-the-art.

Visible light communication (VLC)-based indoor localization approaches enjoy many advantages, such as utilizing ubiquitous lighting infrastructure, high location accuracy, and no interruption to RF-based devices. However, existing VLC-based localization methods lack a real-time backward channel from the device to landmarks and necessitate computation at the device, which make them unsuitable for real-time tracking of small IoT devices. In this paper, we propose and prototype RETRO, that establishes an almost zero-delay backward channel by retroreflection. RETRO localizes passive IoT devices without requiring computation and heavy sensing (e.g., camera) at the devices. Multiple photodiodes (i.e., landmarks) are mounted on any single unmodified light source to sense the retroreflected optical signal (i.e., location signature). We derive a closed-form expression, which is validated by experiments and ray tracing simulations, for the reflected optical power relative to the location and the orientation of the retroreflector. The expression is applied to a received signal strength indicator and trilateration based localization algorithm. Extensive experiments demonstrate centimeter-level location accuracy and single-digit angular error. For practicality concern, to mitigate the thickness problem of a single retroreflector, the capabilities of different retroreflector arrays are studied. The range of the localization system is theoretically evaluated for different light emission patterns.

Many real-world social networks are decentralized in nature, and the only way to analyze such a network is to collect local views of the social graph from individual participants. Since local views may contain sensitive information, it is often desirable to apply differential privacy in the data collection process, which provides strong and rigorous privacy guarantees. In many practical situations, the local view of a participant contains not only her own connections, but also those of her neighbors, which are private and sensitive for the neighbors, but not directly so for the participant herself. We call such information beyond direct connections an extended local view (ELV)</>, and study two fundamental problems related to ELVs: first, how do we correctly enforce differential privacy for all participants in the presence of ELVs? Second, how can the data collector utilize ELVs to obtain accurate estimates of global graph properties?

Neural Network classifiers have been used successfully in a wide range of applications. However, their underlying assumption of attack free environment has been defied by adversarial examples. Researchers tried to develop defenses; however, existing approaches are still far from providing effective solutions to this evolving problem. In this paper, we design a generative adversarial net (GAN) based zero knowledge adversarial training defense, dubbed ZK-GanDef, which does not consume adversarial examples during training. Therefore, ZK-GanDef is not only efficient in training but also adaptive to new adversarial examples. This advantage comes at the cost of small degradation in test accuracy compared to full knowledge approaches. Our experiments show that ZK-GanDef enhances test accuracy on adversarial examples by up-to 49.17% compared to zero knowledge approaches. More importantly, its test accuracy is close to that of the state-of-the-art full knowledge approaches (maximum degradation of 8.46%), while taking much less training time.

In vehicular ad-hoc networks, due to high mobility, vehicles usually communicate for short periods of time with several neighboring vehicles and are required to process data fast; sometimes in the order of few milliseconds. This urgency of data processing is further heightened in safety-critical scenarios that involve many vehicles. Such scenarios require data to be prioritized and processed with minimum delay. While packet scheduling has been extensively studied, these studies focus on channel scheduling, our work focuses on processing received packets by a vehicle in dense scenarios. In this paper, we formulate the prioritized data processing problem as an integer linear program given a prior knowledge of the request sequence and prove that it is NP-complete. Due to the difficulty of predicting the traffic patterns and obtaining the request sequence in advance, we propose an online algorithm that does not require the prior knowledge of the request sequence and achieves an O(1) competitive ratio. The proposed online algorithm strives to accept higher severity packets for processing in order to maximize the cumulative severity given vehicular communications/computation capacity constraints. Using real traffic traces, we evaluate the performance of the online algorithm against three online algorithms, in which two of them use an exponentially weighted moving average-based threshold while the other one accepts requests as capacity permits. Our evaluation shows that our algorithm achieves up to 492% more cumulative severity compared to the three other baseline algorithms.

Inference based techniques are one of the major approaches to analyze DNS data and detect malicious domains. The key idea of inference techniques is to first define associations between domains based on features extracted from DNS data. Then, an inference algorithm is deployed to infer potential malicious domains based on their direct/indirect associations with known malicious ones. The way associations are defined is key to the effectiveness of an inference technique. It is desirable to be both accurate (i.e., avoid falsely associating domains with no meaningful connections) and with good coverage (i.e., identify all associations between domains with meaningful connections). Due to the limited scope of information provided by DNS data, it becomes a challenge to design an association scheme that achieves both high accuracy and good coverage.

Indoor localization is very important to enable Internet-of-things (IoT) applications. Visible light communication (VLC)-based indoor localization approaches enjoy many advantages, such as utilization of existing ubiquitous lighting infrastructure, high location and orientation accuracy, and no interruption to RF -based devices. However, existing VLC-based localization methods lack a real-time backward channel from the device to landmarks and necessitate computation at the device, which make them unsuitable for real-time tracking of small IoT devices. In this paper, we propose and prototype a retroreflector-based visible light localization system (RETRO), that establishes an almost zero-delay backward channel using a retroreflector to reflect light back to its source. RETRO localizes passive IoT devices without requiring computation and heavy sensing (e.g., camera) at the devices. Multiple photodiodes (i.e., landmarks) are mounted on any single unmodified light source to sense the retroreflected optical signal (i.e., location signature). We theoretically derive a closed-form expression for the reflected optical power related to the location and orientation of the retroreflector, and validate the theory by experiments. The characterization of received optical power is applied to a received signal strength indicator and trilateration based localization algorithm. Extensive experiments demonstrate centimeter-level location accuracy and single-digit angular error.

A large amount of valuable information resides in decentralized social graphs, where no entity has access to the complete graph structure. Instead, each user maintains locally a limited view of the graph. For example, in a phone network, each user keeps a contact list locally in her phone, and does not have access to other users' contacts. The contact lists of all users form an implicit social graph that could be very useful to study the interaction patterns among different populations. However, due to privacy concerns, one could not simply collect the unfettered local views from users and reconstruct a decentralized social network.

Internet of Things (IoT) applications can benefit greatly from cloud-hosted message broker services that utilize publish-subscribe communications. The operators of IoT cloud-hosted services are often interested in delivering services that maximize their revenue given quality of service guarantees. In this paper, we formulate the problem of maximizing the profit of the service provider given the prior knowledge of the request sequence as an integer linear program and prove that it is strongly NP-complete, and thus there is no fully polynomialtime approximation scheme for the problem, unless P = NP. Due to the above-mentioned problem and the difficulty of obtaining the request sequence in advance in real-world scenarios, we propose an auction-based online algorithm that does not require the prior knowledge of the request sequence. We prove that the competitive ratio of the online algorithm is (9(log(N)), where N is the number of cloud zones that host the publish- subscribe services. Moreover, we show that no online algorithm can achieve a competitive ratio better than Ω(log(N)). Therefore, our online algorithm achieves the optimal competitive ratio in the asymptotic sense. Our simulations, based on real data traces, show that our algorithm achieves up to 83% more profit compared to a heuristic approach, while consuming 60% less resources.

We design and implement a novel communications latency based authentication scheme, dubbed CLAS, that strengthens the security of state-of-the-art web authentication approaches by leveraging the round trip network communications latency (RTL) between clients and authenticators. In addition to the traditional credentials, CLAS profiles RTL values of clients and uses them to defend against password compromise. The key challenges are (i) to prevent RTL manipulation, (ii) to alleviate network instabilities, and (iii) to address mobile clients. CLAS addresses the first challenge by introducing a novel network architecture, which makes it extremely difficult for attackers to simulate legitimate RTL values. The second challenge is addressed by outlier removal and multiple temporal profiling, while the last challenge is addressed by augmenting CLAS with out-of-band-channels or other authentication schemes. CLAS restricts login to profiled locations while demanding additional information for nonprofiled ones, which highly reduces the attack surface even when the legitimate credentials are compromised. Additionally, unlike many state-of-the-art authentication mechanisms, CLAS is resilient to phishing, pharming, man-in-the-middle, and social engineering attacks. Furthermore, CLAS is transparent to users and incurs negligible overhead. The experimental results show that CLAS can achieve very low false positive and false negative rates.

In this paper, we leverage the characteristics of round-trip communications latency (RTL) to design and implement a novel highly secure and usable web authentication scheme, dubbed CLAS. CLAS uses, in addition to the traditional credentials, round trip network communications latency to uniquely identify users. CLAS introduces a novel network architecture which turns RTL into a robust authentication feature that is extremely difficult to forge. CLAS offers robust defense against password compromise because, unlike many traditional authentication mechanisms, it is resilient to phishing/pharming, man-in-the-middle, and social engineering attacks. Most importantly, CLAS is transparent to users and incurs negligible overhead. Our experimental results show that CLAS can achieve 0.0017 false positive rate while maintaining false negative rate below 0.007.

Malicious domains are key components to a variety of cyber attacks. Several recent techniques are proposed to identify malicious domains through analysis of DNS data. The general approach is to build classifiers based on DNS-related local domain features. One potential problem is that many local features, e.g., domain name patterns and temporal patterns, tend to be not robust. Attackers could easily alter these features to evade detection without affecting much their attack capabilities. In this paper, we take a complementary approach. Instead of focusing on local features, we propose to discover and analyze global associations among domains. The key challenges are (1) to build meaningful associations among domains; and (2) to use these associations to reason about the potential maliciousness of domains. For the first challenge, we take advantage of the modus operandi of attackers. To avoid detection, malicious domains exhibit dynamic behavior by, for example, frequently changing the malicious domain-IP resolutions and creating new domains. This makes it very likely for attackers to reuse resources. It is indeed commonly observed that over a period of time multiple malicious domains are hosted on the same IPs and multiple IPs host the same malicious domains, which creates intrinsic association among them. For the second challenge, we develop a graph-based inference technique over associated domains. Our approach is based on the intuition that a domain having strong associations with known malicious domains is likely to be malicious. Carefully established associations enable the discovery of a large set of new malicious domains using a very small set of previously known malicious ones. Our experiments over a public passive DNS database show that the proposed technique can achieve high true positive rates (over 95%) while maintaining low false positive rates (less than 0.5%). Further, even with a small set of known malicious domains (a couple of hundreds), our technique can discover a large set of potential malicious domains (in the scale of up to tens of thousands).

In local differential privacy (LDP), each user perturbs her data locally before sending the noisy data to a data collector. The latter then analyzes the data to obtain useful statistics. Unlike the setting of centralized differential privacy, in LDP the data collector never gains access to the exact values of sensitive data, which protects not only the privacy of data contributors but also the collector itself against the risk of potential data leakage. Existing LDP solutions in the literature are mostly limited to the case that each user possesses a tuple of numeric or categorical values, and the data collector computes basic statistics such as counts or mean values. To the best of our knowledge, no existing work tackles more complex data mining tasks such as heavy hitter discovery over set-valued data. In this paper, we present a systematic study of heavy hitter mining under LDP. We first review existing solutions, extend them to the heavy hitter estimation, and explain why their effectiveness is limited. We then propose LDPMiner, a two-phase mechanism for obtaining accurate heavy hitters with LDP. The main idea is to first gather a candidate set of heavy hitters using a portion of the privacy budget, and focus the remaining budget on refining the candidate set in a second phase, which is much more efficient budget-wise than obtaining the heavy hitters directly from the whole dataset. We provide both in-depth theoretical analysis and extensive experiments to compare LDPMiner against adaptations of previous solutions. The results show that LDPMiner significantly improves over existing methods. More importantly, LDPMiner successfully identifies the majority true heavy hitters in practical settings.

Since the emergence of Content Centric Networking (CCN) as a new paradigm for content delivery in the Internet, copious of research targeted the evaluation or the enhancement of CCN caching schemes. Motivated by providing the Internet Service Providers with incentives to perform caching, the increasing deployment of in-network cloudlets, and the low cost of storage devices, we study caching in CCN from an economical point of view, where the content providers pay the Internet Service Providers in exchange for caching their content items. We propose an online caching algorithm for CCN that does not require the exact knowledge of content items' popularities to minimize the total cost paid by the content providers. The total cost here is the sum of the caching costs and the retrieval costs. Our analysis shows that the proposed algorithm achieves an O(1) competitive ratio when compared to the optimal offline caching scheme that possesses the exact knowledge of content items' popularities. We also show through simulations that the proposed algorithm can cut the cost incurred by widely used caching schemes such as Leave Copy Down (LCD) and Leave Copy Everywhere (LCE) by up to 65%.

Content caching at intermediate nodes is an effective way to optimize the operations of Computer networks, so that future requests can be served without going back to the origin of the content. Several caching techniques have been proposed in literature, including techniques that require major changes to the Internet architecture. In this work, we present a low complexity, distributed, and online caching algorithm based on content popularity. Our algorithm performs en-route caching using a simple cost-reward comparison. Therefore, it can be integrated with the current TCP/IP model. We use the concept of competitive ratio to measure the performance of any online caching algorithm, in terms of traffic savings, with respect to the performance of the optimal offline algorithm that has a complete knowledge of the future. We show that under our settings, no online algorithm can achieve a better competitive ratio than Ω(logn), where n is the number of nodes in the network. Furthermore, we show that under realistic scenarios, our algorithm has an asymptotically optimal competitive ratio in terms of the number of nodes in the network. We also study several extensions to the basic algorithm and show their effectiveness through extensive simulations.

Network coding-based opportunistic routing has emerged as an elegant way to optimize the capacity of lossy wireless multihop networks by reducing the amount of required feedback messages. Most of the works on network coding-based opportunistic routing in the literature assume that the links are independent. This assumption has been invalidated by the recent empirical studies that showed that the correlation among the links can be arbitrary. In this work, we show that the performance of network coding-based opportunistic routing is greatly impacted by the correlation among the links. We formulate the problem of maximizing the throughput while achieving fairness under arbitrary channel conditions, and we identify the structure of its optimal solution. As is typical in the literature, the optimal solution requires a large amount of immediate feedback messages, which is unrealistic. We propose the idea of performing network coding on the feedback messages and show that if the intermediate node waits until receiving only one feedback message from each next-hop node, the optimal level of network coding redundancy can be computed in a distributed manner. The coded feedback messages require a small amount of overhead, as they can be integrated with the packets. Our approach is also oblivious to losses and correlations among the links, as it optimizes the performance without the explicit knowledge of these two factors.

Many web computing systems are running real time database services where their information change continuously and expand incrementally. In this context, web data services have a major role and draw significant improvements in monitoring and controlling the information truthfulness and data propagation. Currently, web telemedicine database services are of central importance to distributed systems. However, the increasing complexity and the rapid growth of the real world healthcare challenging applications make it hard to induce the database administrative staff. In this paper, we build an integrated web data services that satisfy fast response time for large scale Tele-health database management systems. Our focus will be on database management with application scenarios in dynamic telemedicine systems to increase care admissions and decrease care difficulties such as distance, travel, and time limitations. We propose three-fold approach based on data fragmentation, database websites clustering and intelligent data distribution. This approach reduces the amount of data migrated between websites during applications' execution; achieves cost-effective communications during applications' processing and improves applications' response time and throughput. The proposed approach is validated internally by measuring the impact of using our computing services' techniques on various performance features like communications cost, response time, and throughput. The external validation is achieved by comparing the performance of our approach to that of other techniques in the literature. The results show that our integrated approach significantly improves the performance of web database systems and outperforms its counterparts.

Content caching at intermediate nodes is a very effective way to optimize the operations of Computer networks, so that future requests can be served without going back to the origin of the content. Several caching techniques have been proposed since the emergence of the concept, including techniques that require major changes to the Internet architecture such as Content Centric Networking. Few of these techniques consider providing caching incentives for the nodes or quality of service guarantees for content owners. In this work, we present a low complexity, distributed, and online algorithm for making caching decisions based on content popularity, while taking into account the aforementioned issues. Our algorithm performs en-route caching. Therefore, it can be integrated with the current TCP/IP model. In order to measure the performance of any online caching algorithm, we define the competitive ratio as the ratio of the performance of the online algorithm in terms of traffic savings to the performance of the optimal offline algorithm that has a complete knowledge of the future. We show that under our settings, no online algorithm can achieve a better competitive ratio than O(log n), where n is the number of nodes in the network. Furthermore, we show that under realistic scenarios, our algorithm has an asymptotically optimal competitive ratio in terms of the number of nodes in the network.

Abstract Local monitoring has been demonstrated as a powerful technique for mitigating security attacks in multi‐hopad hocnetworks. In local monitoring, nodes overhear partial neighborhood communication to detect misbehavior such as packet drop or delay. However, local monitoring as presented in the literature is vulnerable to a class of attacks that we introduce here calledstealthy packet dropping. Stealthy packet dropping disrupts the packet from reaching the destination by malicious behavior at an intermediate node. However, the malicious node gives the impression to its neighbors that it performed the legitimate forwarding action. Moreover, a legitimate node comes under suspicion. We introduce four ways of achieving stealthy packet dropping, none of which is currently detectable. We provide a protocol called DISA, based on local monitoring, to remedy each attack. DISA incorporates two techniques—having the neighbors maintain additional information about the routing path, and adding some checking responsibility to each neighbor. We show through analysis and simulation that basic local monitoring (BLM) fails to efficiently mitigate any of the presented attacks while DISA successfully mitigates them. Copyright © 2009 John Wiley & Sons, Ltd.

The performance of wireless networks can be enhanced by performing network coding on the intermediate relay nodes. To enhance the throughput of large wireless networks, we can decompose them into a superposition of simple relay networks called two-hop relay networks. Previously, the capacity region of two-hop relay networks with multiple unicast sessions and limited feedback was characterized where packet erasure channels are used. A near-optimal coding scheme that exploits the broadcast nature and the diversity of the wireless links was proposed. However, the complexity of the scheme is exponential in terms of the number of sessions, as it requires the knowledge of the packets that are received by any subset of the receivers. In this paper, we provide a polynomial time coding scheme and characterize its performance using linear equations. The coding scheme uses random network coding to carefully mix intra and intersession network coding and makes a linear, not exponential, number of decisions. For two-hop relay networks with two sessions, we provide an optimal coding scheme that does not require the knowledge of the channel conditions. We also provide a linear programming formulation that uses our two-hop relay network results as a building block in large lossy multihop networks.

In this paper, we tackle the network coding-based opportunistic routing problem for multicast. We present the factors that affect the performance of the multicast protocols. Then, we formulate the problem as an optimization problem. Using the duality approach, we show that a distributed solution can be used to achieve the optimal solution. The distributed solution consists of two phases. In the first phase, the most reliable broadcasting tree is formed based on the ETX metric. In the second phase, a credit assignment algorithm is run at each node to determine the number of coded packets that the node has to send. The distributed algorithm adapts to the changes in the channel conditions and does not require explicit knowledge of the properties of the network. To reduce the number of feedback messages, and to resolve the problem of delayed feedback, we also perform network coding on the feedback messages. We evaluate our algorithm using simulations which show that in some realistic cases the throughput achieved by our algorithm can be double or triple that of the state-of-the-art.

Recent research has shown that the performance of opportunistic routing and network coding in wireless networks is greatly impacted by the correlation among the links. However, it is difficult to measure the correlation among the links, especially because of the time-varying behavior of the wireless links. Therefore, it is crucial to design a distributed algorithm that does not require the explicit knowledge of the channels' states and can adapt to the varying channel conditions. In this paper, we formulate the problem of maximizing the throughput while achieving fairness under arbitrary channel conditions, and we identify the structure of its optimal solution. As is typical in the literature, the optimal solution requires a large amount of immediate feedback messages, which is unrealistic. We propose the idea of performing network coding on the feedback messages and show that if the intermediate node waits until receiving only one feedback message from each next-hop node, the optimal level of network coding redundancy can be computed in a distributed manner. The coded feedback messages require a small amount of overhead as they can be integrated with the packets. Our approach is also oblivious to losses and correlations among the links as it optimizes the performance without the explicit knowledge of these two factors.

Abstract In a wireless network, mutual secure neighbor verification (MSN) is defined as the capability of a node (verifier) to verify the claim by another node (claimer) that it exists within a certain physical distance from the verifier. This problem has received great attention because it has numerous practical applications. Current state‐of‐the‐art approaches to solve this problem, such as the use of time of flight, signal strength, and angle of arrival, suffer from impracticality in terms of application and computation cost. In this work, we propose two algorithms to mitigate the MSN problem during the incremental deployment phase of static senor networks. Each node should announce its location and the power level it uses for transmission. Cooperative and base station verifications are used to detect nodes that lie about their locations. The simulation results show that we can achieve high detection (>90%) of nodes that forge their location information using either high power transmission or by colluding with other malicious nodes. Copyright © 2011 John Wiley & Sons, Ltd.

Web prediction is a classification problem in which we attempt to predict the next set of Web pages that a user may visit based on the knowledge of the previously visited pages. Predicting user's behavior while serving the Internet can be applied effectively in various critical applications. Such application has traditional tradeoffs between modeling complexity and prediction accuracy. In this paper, we analyze and study Markov model and all- Kth Markov model in Web prediction. We propose a new modified Markov model to alleviate the issue of scalability in the number of paths. In addition, we present a new two-tier prediction framework that creates an example classifier EC, based on the training examples and the generated classifiers. We show that such framework can improve the prediction time without compromising prediction accuracy. We have used standard benchmark data sets to analyze, compare, and demonstrate the effectiveness of our techniques using variations of Markov models and association rule mining. Our experiments show the effectiveness of our modified Markov model in reducing the number of paths without compromising accuracy. Additionally, the results support our analysis conclusions that accuracy improves with higher orders of all- Kth model.

A practical way for maximizing the throughput of a wireless network is to decompose the network into a superposition of small two-hop networks such that network coding can be performed inside these small networks to resolve bottlenecks. We call these networks 2-hop relay networks. Therefore, studying the capacity of 2-hop relay networks is very important. Most practical network coding protocols that perform the superposition ignore the diversity among the links by turning off coding when the channels are lossy. Other protocols deal with the packets separately - not as members of flows - which makes the network coding problem with lossy links intractable. In this paper, we use a different approach by looking at flows or batches instead of individual packets. We characterize the capacity region of the 2-hop relay network with packet erasure channels when the coding operations are limited to XOR. We derive our results by constructing an upper bound on the capacity region and then providing a coding scheme that can achieve the upper bound. The capacity characterization is in terms of linear equations. We also extend our 2-hop relay networks results to multihop wireless networks by providing a linear program that can perform the superposition optimally. We perform extensive simulations for both the 2-hop relay and large wireless networks and show the superiority of our protocols over the network coding protocols that deal with the packets separately.

The broadcast nature of wireless links makes wireless networks an attractive environment for intersession network coding. Most intersession network coding protocols exploit this property, but ignore the diversity among the links by turning off coding when the channels are lossy. Other protocols deal with the packets separately - not as members of flows - which makes the intersession network coding problem with lossy links untractable. In this paper, we use a different approach by looking at flows or batches instead of individual packets. We characterize the capacity region of the 2-hop relay network when the coding operations are limited to XOR. The 2-hop relay network represents all of the local intersession network coding opportunities in large multihop networks. The characterization is in terms of linear equations. We also provide a coding scheme that can achieve the capacity with almost zero feedback overhead. Simulation results show that our scheme enhances the throughput by 82% while maintaining fairness among the flows compared to the intersession network coding protocols that deal with the packets separately.

The network coding problem across multiple unicasts is an open problem. Previously, the capacity region of 2-hop relay networks with multiple unicast sessions and limited feedback was characterized where the coding and decoding nodes are neighbors and packet erasure channels are used. A near-optimal coding scheme that exploits the broadcast nature and the diversity of the wireless links was proposed. However, the complexity of the scheme is hyper exponential as it requires the knowledge of the packets that are received by any subset of the receivers. In this paper, we provide a polynomial time coding scheme and characterize its performance using linear equations. The coding scheme uses random network coding to carefully mix intra and intersession network coding and makes a linear, not exponential, number of decisions. We also provide a linear programming formulation that uses our 2-hop relay network results as a building block in large lossy multihop networks. Through simulations, we verify the superiority of our proposed schemes over state-of-the art.

Over the past decade, local monitoring has been shown to be a powerful technique for improving security in multihop wireless sensor networks (WSNs). Indeed, local monitoring-based security algorithms are becoming the most popular tool for providing security in WSNs. However, local monitoring as it is currently practiced is costly in terms of energy consumption, a major drawback for energy-constrained systems such as WSNs. In WSN environments, the scarce power resources are typically addressed through sleep-wake scheduling of the nodes. However, sleep-wake scheduling techniques in WSNs are vulnerable even to simple attacks. In this paper, a new technique is proposed that promises to allow operation of WSNs in a manner that is both energy-efficient and secure. The proposed technique combines local monitoring with a novel, more secure form of sleep-wake scheduling. The latter is a new methodology dubbed Elmo (Energy Aware Local MOnitoring in Sensor Networks), which enables sleep-wake management in a secure manner even in the face of adversarial nodes that choose not to awaken nodes responsible for monitoring their traffic. An analytical proof is given showing that security coverage is not weakened under Elmo. Moreover, ns-2 simulation results show that the performance of local monitoring is practically unchanged, while energy savings of 20 to 100 times are achieved, depending on the scenario.

Stealthy packet dropping is a suite of four attacks-misrouting, power control, identity delegation, and colluding collision-that can be easily launched against multihop wireless ad hoc networks. Stealthy packet dropping disrupts the packet from reaching the destination through malicious behavior at an intermediate node. However, the malicious node gives the impression to its neighbors that it performs the legitimate forwarding action. Moreover, a legitimate node comes under suspicion. A popular method for detecting attacks in wireless networks is behavior-based detection performed by normal network nodes through overhearing the communication in their neighborhood. This leverages the open broadcast nature of wireless communication. An instantiation of this technology is local monitoring. We show that local monitoring, and the wider class of overhearing-based detection, cannot detect stealthy packet dropping attacks. Additionally, it mistakenly detects and isolates a legitimate node. We present a protocol called Sadec that can detect and isolate stealthy packet dropping attack efficiently. Sadec presents two techniques that can be overlaid on baseline local monitoring: having the neighbors maintain additional information about the routing path, and adding some checking responsibility to each neighbor. Additionally, Sadec provides an innovative mechanism to better utilize local monitoring by considerably increasing the number of nodes in a neighborhood that can do monitoring. We show through analysis and simulation experiments that baseline local monitoring fails to efficiently mitigate most of the presented attacks while SADEC successfully mitigates them.

Stealthy packet dropping in multihop wireless sensor networks can be realized by the colluding collision attack. Colluding collision attack disrupts a packet from reaching its destination by malicious collusion at intermediate nodes. Moreover, the malicious nodes give the impression to their neighbors that they performed the legitimate forwarding action. Therefore, a legitimate node comes under suspicion. We provide a protocol called MCC to remedy the attack. MCC builds forwarding counters at each neighbor of the suspicious node and collaboratively compare the values of these counters to detect any discrepancy that could be caused by malicious packet dropping. We show through analysis and simulation that the basic local monitoring (BLM) fails to efficiently mitigate stealthy colluding collision attacks while MCC successfully mitigates them.

Local monitoring has been demonstrated as a powerful technique for mitigating security attacks in multi-hop ad-hoc networks. In local monitoring, nodes overhear partial neighborhood communication to detect misbehavior such as packet drop or delay. However, local monitoring as presented in the literature is vulnerable to stealthy packet dropping through transmission power control. Stealthy packet dropping disrupts the packet from reaching the destination by malicious behavior at an intermediate node. However, the malicious node gives the impression to its neighbors that it performed the legitimate forwarding action. Moreover, a legitimate node comes under suspicion. We provide a protocol called MPC to remedy the attack. MPC expands the set of possible monitors of a node by redefining the monitor role and functionality. We show through analysis and simulation that the basic local monitoring (BLM) fails to efficiently mitigate stealthy power control attacks while MPC successfully mitigates them.

We propose an uplink non-cooperative game-theoretic power control algorithm for wireless data in direct sequence spread spectrum code division multiple access (DS-CDMA) single cell communication systems. We put forward a target function composed of the difference between a utility function and a pricing function. In this proposed game-theoretic algorithm, each user unilaterally and based on local information decides its transmit power level that maximizes its target function. The pricing function used is a linear function of the transmit power with a slope (pricing factor) chosen in a market-based mechanism, where the base station (BS) decides the pricing factor and communicates it to the users in the cell. Finally, an S-modular games framework is exploited to establish the existence and uniqueness of a Nash equilibrium (NE) operating point of the proposed game. This NE operating game exhibits Pareto dominance over that of NE operating point of other existing Algorithms in the literature as we show in the simulations.

Secure verification of neighborhood membership (SVNM) is a relatively recently stated problem that has an increasing number of practical applications. The problem can be stated as the capability of a wireless network node (verifier) to verify the claim by another node (claimer) that it exists within a certain physical distance from the verifier. Several physical properties of the received signal are used for onehop location estimation - signal strength, time of flight, and angle of arrival, etc. However, many of these approaches are either not practically applicable or prohibitively expensive in many wireless network scenarios. In this work, we propose algorithms to detect and isolate nodes that lie about their location in the incremental addition phase of static sensor networks. Every node should announce its position and the power level it uses for transmission. Cooperative and base station verifications are used to detect nodes that lie about their locations. Our results show that we can achieve perfect detection (100%) of nodes that falsify their location information using either high power transmission or with the help of other nodes.

In this study, we introduce a novel OFDMA-based MAC protocol for underwater acoustic sensor networks. The design approach of this protocol exploits the multi-path characteristics of a fading acoustic channel. The goal is to convert the channel into parallel independent acoustic sub-channels that undergo flat fading. Communication between node pairs within the network is then done using subsets of these sub-channels thus fully utilizing the available limited bandwidth while avoiding interference completely. We derive the mathematical model for optimal sub-channel selection. We also present the results of the experiments that we conducted to evaluate the new protocol and we compare these results to those of a MAC protocol that is based on the CDMA scheme.

Abstract IEEE 802.16 standard lacks a Media Access Control (MAC) scheduling architecture to enforce Quality of Service (QoS) requirements of different services classes. In this paper, we propose an opportunistic and optimized scheduler that meets connections' QoS requirements while it pledges fairness among admitted connections. Our approach involves separating the scheduling problem into two sub‐problems. In the first sub‐problem, which addresses interclass time‐slots allocation, the proposed scheduler calculates the optimum number of time‐slots in each time frame corresponding to the service classes with the objective of minimizing the blocking probability of each class. In the second sub‐problem, the intraclass slot allocation problem, time‐slots for each class are allocated using an integrated cross‐layer priority functions (PFs) that provide proportional fairness among a class connections. The simulation results reveal that the proposed scheduler realizes our objectives, and provides efficient QoS scheduling without starving the connections of the BE class. Copyright © 2008 John Wiley & Sons, Ltd.

Local monitoring has been demonstrated as a powerful technique for mitigating security attacks in multi-hop ad-hoc networks. In local monitoring, nodes overhear partial neighborhood communication to detect misbehavior such as packet drop or delay. However, local monitoring as presented in the literature is vulnerable to an attack calledmisroutingattack. Packet misrouting disrupts the packet from reaching the destination by malicious behavior at an intermediate node. However, the malicious node gives the impression to its neighbors that it performed the legitimate forwarding action. Moreover, a legitimate node comes under suspicion. I provide a protocol called MlMl based on local monitoring to remedy the attack. It requires the neighbors to maintain additional information about the routing path. I show through performance analysis that the basic local monitoring fails to mitigate the misrouting attack while MlMl successfully does.

Local monitoring has been demonstrated as a powerful technique for mitigating security attacks in multi-hop ad-hoc networks. In local monitoring, nodes overhear partial neighborhood communication to detect misbehavior such as packet drop or delay. However, local monitoring as presented in the literature is vulnerable to a class of attacks that we introduce here called stealthy packet dropping. Stealthy packet dropping disrupts the packet from reaching the destination by malicious behavior at an intermediate node. However, the malicious node gives the impression to its neighbors that it performed the legitimate forwarding action. Moreover, a legitimate node comes under suspicion. We introduce four ways of achieving stealthy packet dropping, none of which is currently detectable. We provide a protocol called Mispar based on local monitoring to remedy each attack. It presents two techniques --- having the neighbors maintain additional information about the routing path, and adding some checking responsibility to each neighbor. We show through analysis and simulation that the basic local monitoring fails to mitigate any of the presented attacks while Mispar successfully mitigates them.

Sleep-wake protocols are critical in sensor networks to ensure long-lived operation. However, an open problem is how to develop efficient mechanisms that can be incorporated with sleep-wake protocols to ensure both long-lived operation and a high degree of security. Our contribution in this paper is to address this problem by using local monitoring, a powerful technique for detecting and mitigating control and data attacks in sensor networks. In local monitoring, each node oversees part of the traffic going in and out of its neighbors to determine if the behavior is suspicious, such as, unusually long delay in forwarding a packet. Here, we present a protocol called SLAM to make local monitoring parsimonious in its energy consumption and to integrate it with any extant sleep-wake protocol in the network. The challenge is to enable sleep-wake in a secure manner even in the face of nodes that may be adversarial and not wake up nodes responsible for monitoring its traffic. We prove analytically that the security coverage is not weakened by the protocol. We perform simulations in ns-2 to demonstrate that the performance of local monitoring is practically unchanged while listening energy saving of 30 to 129 times is achieved, depending on the network load.

Wireless reprogramming of a sensor network is useful for uploading new code or for changing the functionality of existing code. Through the process, a node should remain receptive to future code updates because reprogramming may be done multiple times during the node's lifetime. Existing reprogramming protocols, such as Deluge, achieve this by bundling the reprogramming protocol and the application as one program image, thereby increasing the overall size of the image which is transferred through the network. This increases both time and energy required for network reprogramming. We present a protocol called Stream that mitigates the problem by significantly reducing the size of the program image. Using the facility of having multiple code images on a node and switching between them, Stream pre-installs the reprogramming protocol as one image and the application program equipped with the ability to listen to new code updates as the second image. For a sample application, Stream reduces the size of the program image by 10 pages (48 packets/page) compared to Deluge. Stream is implemented on the Mica2 nodes and we conduct testbed and simulation experiments to show the reduction in energy and reprogramming time of Stream compared to Deluge.

In multihop wireless systems, such as ad-hoc and sensor networks, the need for cooperation among nodes to relay each other's packets exposes them to a wide range of security attacks. A particularly devastating attack is known as the wormhole attack, where a malicious node records control and data traffic at one location and tunnels it to a colluding node, which replays it locally. This can have an adverse effect in route establishment by preventing nodes from discovering routes that are more than two hops away. In this paper, we present a lightweight countermeasure for the wormhole attack, called LITEWORP, which does not require specialized hardware. LITEWORP is particularly suitable for resource-constrained multihop wireless networks, such as sensor networks. Our solution allows detection of the wormhole, followed by isolation of the malicious nodes. Simulation results show that every wormhole is detected and isolated within a very short period of time over a large range of scenarios. The results also show that the fraction of packets lost due to the wormhole when LITEWORP is applied is negligible compared to the loss encountered when the method is not applied.

Sensor networks enable a wide range of applications in both military and civilian domains. However, the deployment scenarios, the functionality requirements, and the limited capabilities of these networks expose them to a wide-range of attacks against control traffic (such as wormholes, Sybil attacks, rushing attacks, etc). In this paper we propose a lightweight protocol called DICAS that mitigates these attacks by detecting, diagnosing, and isolating the malicious nodes. DICAS uses as a fundamental building block the ability of a node to oversee its neighboring nodes’ communication. On top of DICAS, we build a secure routing protocol, LSR, which in addition supports multiple node-disjoint paths. We analyze the security guarantees of DICAS and use ns-2 simulations to show its effectiveness against three representative attacks. Overhead analysis is conducted to prove the lightweight nature of DICAS.