We study the problem of solving one-sided, zero-sum, partially observable stochastic games (POSGs). These games model sequential interactions between two adversaries, where one player has partial observability of the game state. They are applicable to many important domains, such as robotics and cybersecurity. Solving such games is computationally challenging since the solution depends on the first player's belief about the game state, which belongs to a continuous (and often high-dimensional) belief space. In the literature, only a single method has demonstrated reliable performance for solving these types of games, namely Heuristic Search Value Iteration (HSVI). However, this method is restricted to small games. We address this limitation by presenting a new method with similar approximation and convergence guarantees but improved scalability and flexibility, which we call SAB: Shapley iteration with Aggregated Beliefs. Our method aggregates the belief space into a finite set of representative beliefs and computes their values through Shapley iteration. It then approximates the value function of the POSG through interpolation from these values. We prove that SAB converges and provide a bound on its approximation error. Experiments across several benchmark games show that SAB matches the performance of HSVI on small game instances while also scaling to larger games. Moreover, we find that SAB is up to 79% faster than HSVI at obtaining a near-optimal approximation.

Evolving security vulnerabilities and shifting operational conditions require frequent updates to network security policies. These updates include adjustments to incident response procedures and modifications to access controls, among others. Reinforcement learning methods have been proposed for automating such policy adaptations, but most methods in the research literature lack performance guarantees and adapt slowly to changes. In this paper, we address these limitations and present a method for computing security policies that is scalable, offers theoretical guarantees, and adapts quickly to changes. The method uses a model or simulator of the system, which is updated when changes occur, and combines three components: belief estimation through particle filtering, offline policy computation through feature-based aggregation, and online policy adaptation through rollout. In particular, feature-based aggregation enables scalable offline optimization of a policy, while rollout adapts the policy online to changes in the system model without repeating the offline optimization. We analyze the approximation error of the aggregation and show that the rollout efficiently adapts policies to changes under certain conditions. Simulations and testbed results demonstrate that our method outperforms state-of-the-art methods on several benchmarks, including CAGE-2.

Time-series anomaly detection is critical for ensuring safety in high-stakes applications, where robustness is a fundamental requirement rather than a mere performance metric. Addressing the vulnerability of these systems to adversarial manipulation is therefore essential. Existing defenses are largely heuristic or provide certified robustness only under $\ell_p$-norm constraints, which are incompatible with time-series data. In particular, $\ell_p$-norm fails to capture the intrinsic temporal structure in time series, causing small temporal distortions to significantly alter the $\ell_p$-norm measures. Instead, the similarity metric Dynamic Time Warping (DTW) is more suitable and widely adopted in the time-series domain, as DTW accounts for temporal alignment and remains robust to temporal variations. To date, however, there has been no certifiable robustness result in this metric that provides guarantees. In this work, we introduce the first DTW-certified robust defense in time-series anomaly detection by adapting the randomized smoothing paradigm. We develop this certificate by bridging the $\ell_p$-norm to DTW distance through a lower-bound transformation. Extensive experiments across various datasets and models validate the effectiveness and practicality of our theoretical approach. Results demonstrate significantly improved performance, e.g., up to 18.7\% in F1-score under DTW-based adversarial attacks compared to traditional certified models.

The practical applicability of multiagent reinforcement learning is hindered by its low sample efficiency and slow learning speed. While reward shaping and expert guidance can partially mitigate these challenges, their efficiency is offset by the need for substantial manual effort. To address these constraints, we introduce Multiagent Environment-aware semi-Automated Guide (MEAG), a novel framework that leverages widely known, highly efficient, and low-resolution single-agent pathfinding algorithms for shaping rewards to guide multiagent reinforcement learning agents. MEAG uses these single-agent solvers over a coarse-grid surrogate that requires minimal manual intervention, and guides agents away from random exploration in a manner that significantly reduces computational costs. When tested across a range of densely and sparsely connected multiagent navigation environments, MEAG consistently outperforms state-of-the-art algorithms, achieving up to faster convergence and higher rewards. These improvements enable the consideration of MARL for more complex real-world pathfinding applications ranging from warehouse automation to search and rescue operations, and swarm robotics.

Anomaly detection in decision-making sequences is a challenging problem due to the complexity of normality representation learning and the sequential nature of the task. Most existing methods based on Reinforcement Learning (RL) are difficult to implement in the real world due to unrealistic assumptions, such as having access to environment dynamics, reward signals, and online interactions with the environment. To address these limitations, we propose an unsupervised method named Offline Imitation Learning based Anomaly Detection (OIL-AD), which detects anomalies in decision-making sequences using two extracted behaviour features: action optimality and sequential association . Our offline learning model is an adaptation of behavioural cloning with a transformer policy network, where we modify the training process to learn a Q function and a state value function from normal trajectories. We propose that the Q function and the state value function can provide sufficient information about agents’ behavioural data, from which we derive two features for anomaly detection. The intuition behind our method is that the action optimality feature derived from the Q function can differentiate the optimal action from others at each local state, and the sequential association feature derived from the state value function has the potential to maintain the temporal correlations between decisions (state–action pairs). Our experiments show that OIL-AD can achieve outstanding online anomaly detection performance with up to 34.8% improvement in F 1 score over comparable baselines. The source code is available on https://github.com/chenwang4/OILAD . • An offline-imitation-learning-based anomaly detection framework (OIL-AD) is proposed. • OIL-AD is an unsupervised, online detection method, designed for practical implementation. • This paper introduces two behaviour features, action optimality and sequential association. • Action optimality measures whether the agent selects the normal action in each state. • Sequential association characterizes if the agent is making expected sequential decisions.

Online trajectory anomaly detection has become a critical task in many real-world applications. However, most existing works assume anomalies are significantly different from normal patterns or require knowing the destinations in advance. In this work, we focus on the problem of detecting anomalous subtrajectories in an online manner without knowing their destinations. This task presents a significant challenge as anomalous subtrajectories may largely overlap with normal trajectories, and we only have limited information on an ongoing trajectory during online detection. To overcome the limitations of current methods, we propose a novel diffusion-based conditional representation learning for online trajectory anomaly detection (DeCoRTAD), that aims to detect anomalies at the representation level, thereby improving computational efficiency. Our framework integrates a diffusion model with two encoders: one for capturing current information and another for encoding historical context. By conditioning the current encoder on the history encoder, we leverage past information as prior knowledge to achieve a more meaningful and compact latent space representation. Our method excels in capturing the normal representation in highly diverse trajectory data, therefore achieving great performance in online detection of fine-grained anomalies. Our experiments show that DeCoRTAD can achieve outstanding online anomaly detection performance in F1 score with an average improvement of 9.31%, and a maximum improvement of 22% over comparable baselines.

Vehicle platooning systems are vulnerable to malicious attacks that exploit vehicle-to-vehicle (V2V) communication, causing potential instability and increased collision risks. Conventional machine learning (ML) detection methods show promise but can be circumvented by intelligent adversaries. In this paper, we present a novel, end-to-end attack detection and mitigation approach that uniquely incorporates advancements in (adversarial) machine learning, control theory, and game theory. We employ a non-cooperative security game with imperfect information to model complex attack/defense interactions. This aids in making informed decisions regarding detector deployment and attack mitigation, even amidst possibly misleading attack detection reports. We model our control system reconfiguration attack mitigation approach as a switched system and provide a n in-depth stability analysis. The simulations conducted in a sophisticated simulator demonstrate our approach’s potential for real-world online deployment. Our game-based defense formulation significantly improves inter-vehicle distance and defense utilities against both cyber-physical and adversarially-masked attacks while reducing the distance disturbance caused by the ambient traffic by up to 87% compared to baseline defense approaches.

We propose a novel supervised dictionary learning framework for text classification, integrating the Lempel-Ziv-Welch (LZW) algorithm for data compression and dictionary construction. This two-phase approach refines dictionaries by optimizing dictionary atoms for discriminative power using mutual information and class distribution. Our method facilitates classifier training, such as SVMs and neural networks. We introduce the information plane area rank (IPAR) to evaluate the information-theoretic performance of our algorithm. Tested on six benchmark text datasets, our model performs nearly as well as top models in limited-vocabulary settings, lagging by only about 2% while using just 10% of the parameters. However, its performance drops in diverse-vocabulary contexts due to the LZW algorithm's limitations with low-repetition data. This contrast highlights its efficiency and limitations across different dataset types.

Anomaly detection is often deployed in centralised systems, for which critical failure points exist. However, the rising availability of low-cost, wireless-connected devices introduces opportunities for new anomaly detection techniques that leverage more robust topologies. In this paper, we propose a novel collaborative training scheme for anomaly detection models that involves sharing machine learning models amongst devices for incremental training. Using the Adversarial Autoencoder architecture, pseudo-rehearsal, and gossip-based communication, our framework provides all participating devices with a structured representation of other devices' data, so that training can continue even in the event of a device failure, with a 43 % smaller performance degradation than state of the art alternatives. Under both optimal conditions and those with device failure, our model consistently exhibits better anomaly detection performance.

Trajectory anomaly detection has become increasingly important in many fields. However, most existing methods do not support online detection, and have limited performance in multi-source/multi-destination environments. To address these challenges, we propose an unsupervised method named intention orientation based trajectory anomaly detection (IO-TAD) which detects anomalies by inferring their underlying intentions in an online manner. IO-TAD is also robust in multi-source/multi-destination environments. To achieve this, we leverage Inverse Reinforcement Learning and recover the reward functions as a representation of the agents' intentions. Then we use the inferred value functions to quantify the agents' intention orientation pattern. The intuition behind our method is that different normal agents take different routes, or have different intentions, but the ways they achieve their intentions are similar. In other words, the values of normal trajectories tend to be monotonically increasing, while anomalous behaviour is non-monotonic. Our experiments show IO-TAD can achieve good online anomaly detection performance with up to 18.1% improvement in$F_{1}$score over the comparable state-of-the-art methods.

While several wireless network simulators exist, the absence of modern, standardised network datasets may adversely affect the application of machine learning methods to problems involving wireless networks. Due to the difficulty of acquiring and sharing more modern network datasets, many of the rapidly evolving techniques in machine learning are only ever ported to network analysis through archaic network datasets such as KDD'99-creating a divide between communication networks and machine learning. To address this divide, this paper presents a new network simulation framework that brings existing network and machine learning tools together to conveniently generate data consisting of PCAP or raw physical layer data and derived statistics in a format that is directly consumable by machine learning algorithms. The proposed simulation frame-work allows the user to design custom networks through a simple configuration file-based scheme instead of learning a sophisticated network simulator. Measuring our framework's performance on consumer-grade devices, raw data is efficiently generated at up to 1, 200 Ethernet frames per second, and processed into feature vectors of 5.6 samples per second (where each sample uses 1 second of simulation time) or 0.29 samples per second if physical layer signals are used.

Despite their many successful applications, Deep Neural Networks (DNNs) are vulnerable to intentionally designed adversarial examples. Adversarial robustness describes the ability of a machine learning model, e.g., a neural network, to defend against such adversarial attacks. In coding theory, codebooks are designed to minimize the impact of errors occurring with transmission through a noisy channel. Motivated by the similarities between passing a codeword through a noisy channel and defending against adversarial attacks, Error-Correcting Output Codes (ECOCs) are used to achieve state-of-the-art adversarial robustness. Research on codebook designs and the association of codewords to classification labels (assignment) is still at the very early stages, with great room for improvement. In this work, we present novel codebook design and assignment procedures in two stages due to the complexity (NP-hardness) of the underlying problem. A rule-based heuristic codebook design method is proposed in the first stage and an optimization problem to assign the codewords to labels is proposed in the second stage. Since this optimization is NP-hard, a greedy algorithm is proposed to provide a sub-optimal solution. We demonstrate the effectiveness of our framework on three benchmark datasets, under different types of adversarial attacks. The experimental results show that our error-correcting output code framework can effectively improve the adversarial robustness of machine learning models, with up to a 10% increase in accuracy.

By offering unprecedented flexibility for wireless communications, Software-Defined Radio (SDR) is an enabling technology for distributed platforms such as Internet-of-Things (IoT) and cognitive radio. We propose a network of SDR devices with sensing and computation capabilities without relying on a centralized architecture to monitor and develop a detailed representation of network behavior in the radio frequency spectrum. This network representation provides a foundation for the task of anomaly detection for security purposes, in which SDRs collect the data needed to identify adversarial actors in a selected environment. We present and evaluate a novel distributed anomaly detection scheme for SDRs that applies Round Robin Learning (RRL) and Random Exchange Learning (REL) strategies to a type of machine learning model known as an autoencoder. In doing so, participating devices fill in gaps in a single model’s data representation, in order to build a model closer to that achieved through centralized training, while maintaining the advantages of distributed systems. The distributed architecture developed improves network robustness, decreases the amount of intranetwork communication, and preserves privacy of individual SDR subnetworks. Furthermore, it was found to consistently raise the average anomaly detection performance for groups of participating devices without exchanging any training data.

Nonlinear regression, although widely used in engineering, financial and security applications for automated decision making, is known to be vulnerable to training data poisoning. Targeted poisoning attacks may cause learning algorithms to fit decision functions with poor predictive performance. This paper presents a new analysis of local intrinsic dimensionality (LID) of nonlinear regression under such poisoning attacks within a Stackelberg game, leading to a practical defense. After adapting a gradient-based attack on linear regression that significantly impairs prediction capabilities to nonlinear settings, we consider a multi-step unsupervised black-box defense. The first step identifies samples that have the greatest influence on the learner's validation error; we then use the theory of local intrinsic dimensionality, which reveals the degree of being an outlier of data samples, to iteratively identify poisoned samples via a generative probabilistic model, and suppress their influence on the prediction function. Empirical validation demonstrates superior performance compared to a range of recent defenses.

Multiagent reinforcement learning has shown success in guiding the agents' behaviour in systems that have realworld significance. In these frameworks, agents learn how to interact with the environment and other agents while satisfying their objectives. Unfortunately, the level of complexity of realworld problems requires a significant investment of computational resources before multiagent reinforcement learning methods are able to deliver results. However, by incorporating a priori domain knowledge, more computationally-efficient algorithms can be developed. In this paper, for the first time, we present a Domain-Aware Multiagent Actor-Critic (DAMAC) algorithm, which integrates domain knowledge with the centralised learning and decentralised execution multiagent reinforcement learning approach using domain-specific solvers. Our experiments show that our algorithm achieves substantial high reward and reduces the training time by two orders of magnitude as compared to other multiagent reinforcement learning algorithms. This enables the adoption of this powerful framework in more resource-constrained scenarios.

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

We propose a novel supervised dictionary learning framework, which is based on discriminative power maximization and the classic Lempel-Ziv-Welch (LZW) source coding algorithm that can handle variable-length inputs. In the first stage, the input data is serialized and a dictionary is generated by the LZW algorithm. In the second stage, the dictionary is updated by discarding or selecting a certain amount of atoms in order to increase discriminative power measured by information bottleneck as well as linear similarity metrics. This dictionary learning framework is then analyzed using information bottleneck principles. Experiments on real life datasets from information security and social networking illustrate the effectiveness and accuracy of the proposed algorithms, especially for variable length data. This novel framework also identifies interpretable dictionary atoms, which provide insights to decision processes of the learning system.

Recent studies have demonstrated that reinforcement learning (RL) agents are susceptible to adversarial manipulation, similar to vulnerabilities previously demonstrated in the supervised learning setting. While most existing work studies the problem in the context of computer vision or console games, this paper focuses on reinforcement learning in autonomous cyber defence under partial observability. We demonstrate that under the black-box setting, where the attacker has no direct access to the target RL model, causative attacks-attacks that target the training process-can poison RL agents even if the attacker only has partial observability of the environment. In addition, we propose an inversion defence method that aims to apply the opposite perturbation to that which an attacker might use to generate their adversarial samples. Our experimental results illustrate that the countermeasure can effectively reduce the impact of the causative attack, while not significantly affecting the training process in non-attack scenarios.

This paper studies the complexity of solving two classes of non-cooperative games in a distributed manner, in which the players communicate with a set of system nodes over noisy communication channels. The complexity of solving each game class is defined as the minimum number of iterations required to find a Nash equilibrium (NE) of any game in that class with ∈ accuracy. First, we consider the class G of all N-player non-cooperative games with a continuous action space that admit at least one NE. Using information-theoretic inequalities, a lower bound on the complexity of solving G is derived which depends on the Kolmogorov 2∈-capacity of the constraint set and the total capacity of the communication channels. Our results indicate that the game class G can be solved at most exponentially fast. We next consider the class of all N-player non-cooperative games with at least one NE such that the players' utility functions satisfy a certain (differential) constraint. We derive lower bounds on the complexity of solving this game class under both Gaussian and non-Gaussian noise models. Finally, we derive upper and lower bounds on the sample complexity of a class of quadratic games. It is shown that the complexity of solving this game class scales according to Θ (1/∈2) where € is the accuracy parameter.

Directional wireless power transfer (WPT) technology provides a promising energy solution to remotely recharge the Internet of things sensors using directional antennas. Under a harvest-then-transmit protocol, the access point can adaptively allocate the transmit power among multiple energy directions to maximize the social welfare of the sensors, i.e., downlink sum received energy or uplink sum rate, based on full or quantized channel gains reported from the sensors. However, such power allocation can be challenged if each sensor belongs to a different agent and works in a competitive way. In order to maximize their own utilities, the sensors have the incentives to falsely report their channel gains, which unfortunately reduces the social welfare. To tackle this problem, we design the strategy-proof mechanisms to ensure that each sensor’s dominant strategy is to truthfully reveal its channel gain regardless of other sensors’ strategies. Under the benchmark full channel gain reporting (CGR) scheme, we adopt the Vickrey-Clarke-Groves (VCG) mechanism to derive the price functions for both downlink and uplink, where the truthfulness is guaranteed by asking each sensor to pay the social welfare loss of all other sensors attributable to its presence. For the 1-bit CGR scheme, the problem is more challenging due to the severe information asymmetry, where each sensor has true valuation of full channel gain but may report the false information of quantized channel gain. We prove that the classic VCG mechanism is no longer truthful and then propose two threshold-based price functions for both downlink and uplink, where the truthfulness is ensured by letting each sensor pay its own achievable utility improvement due to its participation. The numerical results validate the truthfulness of the proposed mechanism designs.

Real-time Internet of Things (IoT) applications have stringent delay requirements when implemented over distributed sensing and communication networks in smart traffic control. They require the system to reach a permissible neighbourhood of an optimum solution with a tolerable delay. The performance of such applications mostly depends on the delay introduced by the underlying optimization algorithms, with the localized computational capability. In this paper, we study a smart traffic control scenario-a real-time IoT application, where a group of autonomous vehicles independently decide on their lane velocities, in collaboration with road-side units to efficiently utilize intersections with minimal environmental impact. We decompose this problem as an unconstrained network utility maximization problem. A consensus-based, constant step-size gradient descent algorithm is proposed to obtain a near-optimal solution. We analyze the delay-accuracy tradeoff in reaching a near-optimal velocity. Delay is measured in terms of the number of iterations required before the scheduling operation can be done for a particular tolerance. The operation of the algorithm under quantized message passing is also studied. On contrary to the existing methods to intersection management problems, our approach studies the limit at which an optimization algorithm fails to cater for the requirements of a real-time application and must fall back for a pareto-optimal solution, due to the communication constraints. We used simulation of urban mobility to incorporate the microscopic behavior of traffic flows to our simulations and compared our solution with traditional and state-of-the-art intersection management techniques.

Software-defined radios (SDRs) with substantial cognitive (computing) and networking capabilities provide an opportunity for observing radio communications in an area and potentially identifying malicious rogue agents. Assuming a prevalence of encryption methods, a cognitive network of such SDRs can be used as a low-cost and flexible scanner/sensor array for distributed detection of anomalous communications by focusing on their statistical characteristics. Identifying rogue agents based on their wireless communications patterns is not a trivial task, especially when they deliberately try to mask their activities. We address this problem using a novel framework that utilizes adversarial learning, non-linear data transformations to minimize the rogue agent's attempts at masking their activities, and game theory to predict the behavior of rogue agents and take the necessary countermeasures.

This paper studies the stability and dynamic control of underlay mobile edge networks. First, the stability region for a multiuser edge network is obtained under the assumption of full channel state information. This result provides a benchmark figure for comparing performance of the proposed algorithms. Second, a centralized joint flow control and scheduling algorithm is proposed to stabilize the queues of edge devices while respecting the average and instantaneous interference power constraints at the core access point. This algorithm is proven to converge to a utility point arbitrarily close to the maximum achievable utility within the stability region. Finally, more practical implementation issues such as distributed scheduling are examined by designing efficient scheduling algorithms taking advantage of communication diversity. The proposed distributed solutions utilize mini-slots for contention resolution and achieve a certain fraction of the utility optimal point. The performance lower bounds for distributed algorithms are determined analytically. The detailed simulation study is performed to pinpoint the cost of distributed control for mobile edge networks with respect to centralized control.

Next-generation networks will be characterized by three key features:heterogeneity, in terms of technologies and services,dynamics, in terms of rapidly varying environments and uncertainty, andsize, in terms of the numbers of users, nodes, and services. The emergence of such large-scale and decentralized heterogeneous networks operating under dynamic and uncertain environments imposes new challenges in the design, analysis, and optimization of networks. The past decade has witnessed a confluence among the disciplines of networks, games, and economics, which has necessitated novel mathematical tools and designs that can truly remove the boundaries between these disciplines. In this context, advancing game-theoretic models and tailoring them towards the optimization and operation of future networked systems become pressing needs for our research community. The main goal of this IEEE JSAC Special Issue on “Game Theory for Networks” is to collect cutting-edge contributions that address and show the latest developments in game-theoretic models for emerging networking applications. The response of the community to the call has been overwhelming. We received a total of 120 submissions. We want to thank all the authors who submitted their works to this Special Issue. After a strict and selective review process, we accepted 40 papers and decided to publish two issues. Papers were selected based on their appropriateness for and relevance to the Special Issue as well as their technical merits. Unfortunately, a number of interesting papers did not make the cut because of the criteria set forth above and also due to the constraints on the total page count in a JSAC Special Issue. We hope that such interesting papers will find other venues for publication.

This is the second part of the IEEE JSAC Special Issue on “Game Theory for Networks.” The response of the community to the call has been overwhelming. We received a total of 120 submissions. We want to thank all the authors who submitted their works to this Special Issue. After a strict and selective review process, we accepted 40 papers and decided to publish two issues, each of 20 papers. The first one was published in February 2017. The papers of this second issue cover a wide selection of topics as follows.

Cloud computing enables users to consume various IT resources in an on-demand manner, and with low management overhead. However, customers can face new security risks when they use cloud computing platforms. In this paper, we focus on one such threat-the co-resident attack, where malicious users build side channels and extract private information from virtual machines co-located on the same server. Previous works mainly attempt to address the problem by eliminating side channels. However, most of these methods are not suitable for immediate deployment due to the required modifications to current cloud platforms. We choose to solve the problem from a different perspective, by studying how to improve the virtual machine allocation policy, so that it is difficult for attackers to co-locate with their targets. Specifically, we (1) define security metrics for assessing the attack; (2) model these metrics, and compare the difficulty of achieving co-residence under three commonly used policies; (3) design a new policy that not only mitigates the threat of attack, but also satisfies the requirements for workload balance and low power consumption; and (4) implement, test, and prove the effectiveness of the policy on the popular open-source platform OpenStack.

Tiny computers located in end-user premises are becoming popular as local servers for Internet of Things (IoT) and Fog computing services. These highly distributed servers that can host and distribute content and applications in a peer-to-peer (P2P) fashion are known as nano data centers (nDCs). Despite the growing popularity of nano servers, their energy consumption is not well-investigated. To study energy consumption of nDCs, we propose and use flow-based and time-based energy consumption models for shared and unshared network equipment, respectively. To apply and validate these models, a set of measurements and experiments are performed to compare energy consumption of a service provided by nDCs and centralized data centers (DCs). A number of findings emerge from our study, including the factors in the system design that allow nDCs to consume less energy than its centralized counterpart. These include the type of access network attached to nano servers and nano server's time utilization (the ratio of the idle time to active time). Additionally, the type of applications running on nDCs and factors such as number of downloads, number of updates, and amount of preloaded copies of data influence the energy cost. Our results reveal that number of hops between a user and content has little impact on the total energy consumption compared to the above-mentioned factors. We show that nano servers in Fog computing can complement centralized DCs to serve certain applications, mostly IoT applications for which the source of data is in end-user premises, and lead to energy saving if the applications (or a part of them) are off-loadable from centralized DCs and run on nDCs.

While cloud computing has facilitated easy and affordable access to IT resources, it has also introduced a wide range of security risks from almost every layer and component of cloud systems. In this paper, we focus on one risk at the virtual machine level and the co-resident attack, where by constructing various types of side channels, malicious users can obtain sensitive information from other virtual machines that co-locate on the same physical server. Most previous work has focused on the elimination of side channels, or more generally speaking, the possible countermeasures after attackers co-locate with their targets. In contrast, we provide a different perspective, and propose a defence mechanism that makes it difficult and expensive for attackers to achieve co-residence in the first place. Specifically, we first identify the potential differences between the behaviors of attackers and legal users. Second, we apply clustering analysis and semi-supervised learning techniques to classify users. Third, we model the problem as a two-player security game, and give a detailed analysis of the optimum strategies for both players. Finally, we demonstrate that the attacker's overall cost is increased dramatically by one-to-two orders of magnitude as a result of our defence mechanism.

Strategic users in a wireless network cannot be assumed to follow the network algorithms blindly. Moreover, some of these users aim to use their knowledge about network algorithms to maliciously gain more resources and also to create interference to other users. We consider a scenario, in which the network and legitimate users gather probabilistic information about the presence of malicious users by observing the network over a long time period. The network (mechanism designer) and legitimate users modify their actions according to this Bayesian information. We consider Bayesian mechanisms, both pricing schemes and auctions, and obtain the Bayesian Nash Equilibrium (BNE) points. The BNE points provide conditions under which, the uncertainty about user's nature (type) is better for regular (legitimate) users. To derive these conditions, we compare the Bayesian case to the complete information case. We obtain the optimal prices and allocations, which counter the malicious users. We also provide detection methods based on machine learning algorithms for the detection of malicious users, by observing the prices and rate allocations. In addition, we provide detection using regression learning by observing the anomalies in the utility functions of malicious users from prices, which is implemented along with the pricing mechanism itself. For the designer and the regular users, in a complementary fashion, the results of the detections provide a better estimate of the statistics of malicious users to implement the pricing mechanisms. We have also proposed a truthful Bayesian mechanism in the presence of malicious users. The numerical studies for malicious user detection are carried out with the model proposed in the paper as well as using real Botnet dataset.

The liberalization of wireless communication services markets and the subsequent competition among network operators, is expected to foster optimal utilization of the scarce wireless spectrum and ensure the provision of cost-efficient services to users. However, such markets may function inefficiently due to collusion of operators which yields a de-facto monopoly. Although it is illegal and detrimental to the users, creation of such cartels arises often in the form of implicit price fixing. In this paper, we consider a general such market where a set of operators sell communication services to a large population of users. We use an evolutionary game model to capture the user dynamics in selecting operators, under limited information about the actual service quality, and we analyze the anticipated interaction of the operators using coalitional game theory. We define a coalition formation game in order to rigorously study the conditions that render monopolistic or oligopolistic markets stable under different notions of coalition stability. We also provide direct and indirect regulation methods, such as setting price upper bounds or allocating different amounts of spectrum, in order to discourage undesirable equilibriums. Our approach provides intuitions about collusion strategies, as well as on directions for identifying and preventing them.

When there are strategic and malicious users in a wireless network, the resource allocation is complicated due to the information limitation about the nature of users and network parameters. Bayesian games are appropriate tools to analyze the network resource allocation with heterogeneous users. We consider a scenario with arbitrary number of malicious users in the network, in which individual users gather probabilistic information about the density of malicious users. Users and the base station observe the network over a long time period and modify their actions accordingly. The power allocation in wireless networks which we consider in this paper, is subject to Quality of Service (QoS) requirements. We consider Bayesian pricing mechanisms where the prices are modified using the Bayesian information about types of the users to satisfy the QoS requirements. We also give detection methods based on regression learning algorithms which are used for forming the probability of a user being malicious. The utilities of the users are formed by observing the power strategies of the users and the anomalies are detected. We obtain numerically, the Bayesian Nash Equilibrium (BNE) points of the Bayesian games. We also evaluate the effect of incomplete information on the satisfaction of the QoS requirements of the users in the mechanisms. These mechanisms are with prices which were originally developed for networks with complete information.

Interactive cloud computing and cloud-based applications are a rapidly growing sector of the expanding digital economy because they provide access to advanced computing and storage services via simple, compact personal devices. Recent studies have suggested that processing a task in the cloud is more energy-efficient than processing the same task locally. However, these studies have generally ignored the power consumption of the network and end-user devices when accessing the cloud. In this paper, we develop a power consumption model for interactive cloud applications that includes the power consumption of end-user devices and the influence of the applications on the power consumption of the various network elements along the path between the user and the cloud data centre. As examples, we apply our model to Google Drive and Microsoft Skydrive's word processing, presentation and spreadsheet interactive applications. We demonstrate via extensive packet-level traffic measurements that the volume of traffic generated by a session of the application vastly exceeds the amount of data keyed in by the user. This has important implications on the overall power consumption of the service. We show that using the cloud to perform certain tasks consumes more power (by a watt to 10 watts depending on the scenario) than performing the same tasks locally on a low-power consuming computer and a tablet.

Dual control aims to concurrently learn and control an unknown system. However, actively learning the system conflicts directly with any given control objective for it will disturb the system during exploration. This paper presents a receding horizon approach to dual control, where a multiobjective optimization problem is solved repeatedly and subject to constraints representing system dynamics. Balancing a standard finite-horizon control objective, a knowledge gain objective is defined to explicitly quantify the information acquired when learning the system dynamics. Measures from information theory, such as entropy-based uncertainty, Fisher information, and relative entropy, are studied and used to quantify the knowledge gained as a result of the control actions. The resulting iterative framework is applied to Markov decision processes and discrete-time nonlinear systems. Thus, the broad applicability and usefulness of the presented approach is demonstrated in diverse problem settings. The framework is illustrated with multiple numerical examples.

Online social networks (OSNs) with their huge number of active users consume significant amount energy both in the data centers and in the transport network. Existing studies focus mainly on the energy consumption in the data centers and do not take into account the energy consumption during the transport of data between end-users and data centers. To indicate the amount of the neglected energy, this paper provides a comprehensive framework and a set of measurements for understanding the energy consumption of cloud applications such as photo sharing in social networks. A new energy model is developed to estimate the energy consumption of cloud applications and applied to sharing photos on Facebook, as an example. Our results indicate that the energy consumption involved in the network and end-user devices for photo sharing is approximately equal to 60% of the energy consumption of all Facebook data enters. Therefore, achieving an energy-efficient cloud service requires energy efficiency improvement in the transport network and end-user devices along with the related data centers.

Strategic users in a wireless network cannot be assumed to follow the network algorithms blindly. Moreover, some of these users could be controlled by powerful Botnets, which aim to use their knowledge about network algorithms to maliciously gain more resources and also to create interference to other users. We consider a scenario; in which a mechanism designer and legitimate users together, in a wireless network, gather probabilistic information about the presence of malicious users and modify their actions accordingly. The probabilistic information is gathered by observing the network over a long time period. We study Bayesian mechanisms, both pricing schemes and auctions, and obtain the Nash Equilibrium (NE) points of the underlying Bayesian games. The NE points provide conditions indicating when it is better for users to to hide or reveal their nature (types). The prices and allocations in the mechanisms are later modified using the Bayesian information about the type of the users. The numerical studies show the NE points and illustrate the results.

While the services-based model of cloud computing makes more and more IT resources available to a wider range of customers, the massive amount of data in cloud platforms is becoming a target for malicious users. Previous studies show that attackers can co-locate their virtual machines (VMs) with target VMs on the same server, and obtain sensitive information from the victims using side channels. This paper investigates VM allocation policies and practical countermeasures against this novel kind of co-resident attack by developing a set of security metrics and a quantitative model. A security analysis of three VM allocation policies commonly used in existing cloud computing platforms reveals that the server's configuration, oversubscription and background traffic have a large impact on the ability to prevent attackers from co-locating with the targets. If the servers are properly configured, and oversubscription is enabled, the best policy is to allocate new VMs to the server with the most VMs. Based on these results, a new strategy is introduced that effectively decreases the probability of attackers achieving co-residence. The proposed solution only requires minor changes to current allocation policies, and hence can be easily integrated into existing cloud platforms to mitigate the threat of co-resident attacks.

Can we measure the difficulty of an optimization problem? Although optimization plays a crucial role in modern science and technology, a formal framework that puts problems and solution algorithms into a broader context has not been established. This paper presents a conceptual approach which gives a positive answer to the question for a broad class of optimization problems. Adopting an information and computational perspective, the proposed framework builds upon Shannon and algorithmic information theories. As a starting point, a concrete model and definition of optimization problems is provided. Then, a formal definition of optimization difficulty is introduced which builds upon algorithmic information theory. Following an initial analysis, lower and upper bounds on optimization difficulty are established. One of the upper-bounds is closely related to Shannon information theory and black-box optimization. Finally, various computational issues and future research directions are discussed.

Cloud computing and cloud-based services are a rapidly growing sector of the expanding digital economy. Recent studies have suggested that processing a task in the cloud is more energy-efficient than processing the same task locally. However, these studies have generally ignored the network transport energy and the additional power consumed by end-user devices when accessing the cloud. In this paper, we develop a simple model to estimate the incremental power consumption involved in using interactive cloud services. We then apply our model to a representative cloud-based word processing application and observe from our measurements that the volume of traffic generated by a session of the application typically exceeds the amount of data keyed in by the user by more than a factor of 1000. This has important implications on the overall power consumption of the service. We provide insights into the reasons behind the observed traffic levels. Finally, we compare our estimates of the power consumption with performing the same task on a low-power consuming computer. Our study reveals that it is not always energy-wise to use the cloud. Performing certain tasks locally can be more energy-efficient than using the cloud.

It is well known that Max-Weight type scheduling algorithms are throughput optimal since they achieve the maximum throughput while maintaining the network stability. However, the majority of existing works employing Max-Weight algorithm require the complete channel state information (CSI) at the scheduler without taking into account the associated overhead. In this work, we design a Scheduling and Selective Feedback algorithm (SSF) taking into account the overhead due to acquisition of CSI. SSF algorithm collects CSI from only those users with sufficiently good channel quality so that it always schedules the user with the highest queue backlog and channel rate product at every slot. We characterize the achievable rate region of SSF algorithm by showing that SSF supports 1 + ϵ fraction of the rate region when CSI from all users are collected. We also show that the value of ϵ depends on the expected number of users which do not send back their CSI to the base station. For homogenous and heterogeneous channel conditions, we determine the minimum number of users that must be present in the network so that the rate region is expanded, i.e., ϵ > 0. We also demonstrate numerically in a realistic simulation setting that this rate region can be achieved by collecting CSI from only less than 50% of all users in a CDMA based cellular network utilizing high data rate (HDR) protocol.

It is well known that opportunistic scheduling algorithms are throughput optimal under dynamic channel and network conditions. However, these algorithms achieve a hypothetical rate region which does not take into account the overhead associated with channel probing and feedback required to obtain the full channel state information at every slot. In this work, we design a joint scheduling and channel probing algorithm by considering the overhead of obtaining the channel state information. We adopt a correlated and non-stationary channel model, which is more realistic than those used in the literature. We use concepts from learning and information theory to accurately track channel variations to minimize the number of channels probed at every slot, while scheduling users to maximize the achievable rate region of the network. Simulation results show that with the proposed algorithm, the network can carry higher user traffic.

It is well known that Max-Weight scheduling provides queue stability whenever this is possible. However, Max-Weight scheduling requires the complete channel state information (CSI) to make the best transmission decision at every time slot. The common assumption in this line of research assumes that the network controller has full CSI at every decision time without taking into account the overhead associated with channel probing. In practice, however, acquiring CSI is not cost-free and requires certain amount of resources. In this work, we design a Scheduling and Dynamic Feedback algorithm, named SDF, by considering the overhead of obtaining the channel state information. We first establish a bound on the achievable rate region of SDF algorithm by proving that SDF supports 1+ ϵ fraction of of the full rate region (the rate region when all users are probed) where ϵ only depends on the expected number of users which are not probed. Then, for homogenous channel, we show that when the number of users in the network is greater than 3, ϵ >;0, i.e., we guarantee to expand the rate region. We also demonstrate numerically in a realistic simulation setting that this rate region can be achieved by probing only less than 50% of all channels in a CDMA based cellular network utilizing high data rate protocol under normal channel conditions.

This paper presents a game-theoretic and learning approach to security risk management based on a model that captures the diffusion of risk in an organization with multiple technical and business processes. Of particular interest is the way the interdependencies between processes affect the evolution of the organization's risk profile as time progresses, which is first developed as a probabilistic risk framework and then studied within a discrete Markov model. Using zero-sum dynamic Markov games, we analyze the interaction between a malicious adversary whose actions increases the risk level of the organization and a defender agent, e.g. security and risk management division of the organization, which aims to mitigate risks. We derive min-max (saddle point) solutions of this game to obtain the optimal risk management strategies for the organization to achieve a certain level of performance. This methodology also applies to worst-case scenario analysis where the adversary can be interpreted as a nature player in the game. In practice, the parameters of the Markov game may not be known due to the costly nature of collecting and processing information about the adversary as well an organization with many components itself. We apply ideas from Q-learning to analyze the behavior of the agents when little information is known about the environment in which the attacker and defender interact. The framework developed and results obtained are illustrated with a small example scenario and numerical analysis.

Network mechanism design aims to achieve system level goals such as efficiency or social welfare maximization through resource allocation mechanisms on networks, where individual users are selfish and independent decision-makers. This paper focuses on mechanisms for energy efficiency in uplink of multi-carrier wireless systems and rate control in wireless networks. The problem is modeled as strategic (nonco-operative) game with a resource pricing scheme, where the prices are imposed by a mechanism designer. The users decide on their actions according to both own preferences and given prices. The rules and incentives of the mechanism are designed in such a way that the system objective which is a combination of social welfare and energy efficiency is maximized. A weighting parameter in the objective function allows to smoothly vary the emphasis from social welfare to energy efficiency according to the preference of the designer. The users are assumed initially to be concerned only about their throughput. However, the designer modifies their incentives using pricing so that they become more energy-aware. A distributed optimization framework is developed accordingly in which users have general concave utilities (in terms of throughput) that are unknown to the designer and the energy-efficiency objective is expressed as a convex function of user power levels. An iterative pricing mechanism is obtained as a result, and illustrated through simulations.

In mobile commerce, a company provides location based services to a set of mobile users. The users report to the company their location with a level of granularity to maintain a degree of anonymity, depending on their perceived risk, and receive in return monetary benefits or better services from the company. This paper formulates a quantitative model in which information theoretic metrics such as entropy, quantify the anonymity level of the users. The individual perceived risks of users and the benefits they obtain are considered to be linear functions of their chosen location information granularity. The interaction between the mobile commerce company and its users are investigated using mechanism design techniques as a privacy game. The user best responses and optimal strategies for the company are derived under budgetary constraints on incentives, which are provided to users in order to convince them to share their private information at the desired level of granularity.

Vehicular networks (VANETs) can be used to improve transportation security, reliability, and management. This paper investigates security aspects of VANETs within a game-theoretic framework where defensive measures are optimized with respect to threats posed by malicious attackers. The formulations are chosen to be abstract on purpose in order to maximize applicability of the models and solutions to future systems. The security games proposed for vehicular networks take as an input centrality measures computed by mapping the centrality values of the car networks to the underlying road topology. The resulting strategies help locating most valuable or vulnerable points (e.g., against jamming) in vehicular networks. Thus, optimal deployment of traffic control and security infrastructure is investigated both in the static (e.g., fixed roadside units) and dynamic cases (e.g., mobile law enforcement units). Multiple types of security games are studied under varying information availability assumptions for the players, leading to fuzzy game and fictitious play formulations in addition to classical zero-sum games. The effectiveness of the security game solutions is evaluated numerically using realistic simulation data obtained from traffic engineering systems.

Security risk management is becoming increasingly important in a variety of areas related to information technology (IT), such as telecommunications, cloud computing, banking information systems, etc. In this paper, we develop a systematic quantitative framework for security risk management in IT-intensive organizations. This framework provides a unified viewpoint for considering a wide array of security risk factors which can disrupt business continuity. Our approach integrates the three phases of security risk management, namely risk modeling, assessment, and control/mitigation, through a formulation based on directed graphs, cascades of failures, and mathematical optimization. We consider how security events can propagate through an organization and how resource allocation decisions can be made in order to mitigate the amount of damage they cause. The applicability and effectiveness of our framework is demonstrated through a numerical study which shows significant cost reductions when compared to heuristic methods.

Digital identity and trust management mechanisms play an important role on the Internet. They help users make decisions on trustworthiness of digital identities in online communities or e-commerce environments, which have significant security consequences. This work aims to contribute to construction of an analytical foundation for digital identity and trust by adopting a quantitative approach. A game theoretic model is developed to quantify community effects and other factors in trust decisions. The model captures factors such as peer pressure and personality traits. The existence and uniqueness of a Nash equilibrium solution is studied and shown for the trust game defined. In addition, synchronous and asynchronous update algorithms are shown to converge to the Nash equilibrium solution. A numerical analysis is provided for a number of scenarios that illustrate the interplay between user behavior and community effects.

The paper characterizes the class of all concave resource allocation problems in interference coupled wireless systems. An axiomatic framework for interference functions proposed by Yates in 1995 is used to model interference coupling in our paper. The paper shows that there exists no transformation, which ensures concavity for all linear interference functions for all functions of SINR. The paper then characterizes the largest class of utility functions under a certain requirement, such that the corresponding class of utility functions functions, which are a function of SINR in the s-domain are concave. The paper shows that such a class of utility functions is a restricted class due to a requirement, which ensures concavity. Furthermore, the paper shows that the largest class of interference functions, which ensures concavity for resource allocation problems are the log-convex interference functions. These results differ from the convex case, where we are interested in minimizing utility functions of inverse SINR.

This paper investigates the possibility of having convex formulations of optimization problems for interference coupled wireless systems. An axiomatic framework for interference functions proposed by Yates in 1995 is used to model interference coupling in our paper. The paper shows, that under certain very natural assumptions -- the exponential mapping is the unique transformation (up to a constant), for ``convexification'' of resource allocation problems for linear interference functions. The paper shows that it is sufficient to check for the joint convexity of the sum of weighted utility functions of inverse signal-to-interference (plus noise)-ratio, if we would like the resulting resource allocation problem to be convex. The paper characterizes the largest class of interference functions, which allow a convex formulation of a problem for interference coupled wireless systems. It extends previous literature on log--convex interference functions and provides boundaries on the class of problems in wireless systems, which are jointly convex and hence can be efficiently solved at least from a numerical perspective.

Security risk management for information technology-based organizations has become increasingly important in recent years. However, the risk assessment and mitigation strategies that these organizations employ have remained relatively ad hoc and qualitative. In this paper, we extend a quantitative framework for risk assessment called Risk-Rank to include risk mitigation through Markov Decision Processes. By doing so, we provide an analysis-to-action quantitative approach to security risk management, enabling IT managers to perform more comprehensive evaluations of their risk exposures. We demonstrate the effectiveness of this approach through an example related to the patching of computers in a corporate network.

This paper studies the properties of Nash equilibrium for noncooperative games in interference coupled wireless systems, where it serves as an incentive-compatible solution concept and an operating point. A broad class of noncooperative power control games played among the users of the wireless system is defined based on a general interference function framework, which models the interference coupling through a set of axioms. Both cases of coupling with and without self-interference are considered. The special properties of the underlying interference functions as well as relevant sufficient conditions are investigated to establish the existence and uniqueness of a Nash equilibrium solution. These properties play an important role in developing incentive-compatible distributed algorithms for a variety of wireless networks with interference coupling.

This paper presents a distributed Support Vector Machine (SVM) algorithm in order to detect malicious software (malware) on a network of mobile devices. The light-weight system monitors mobile user activity in a distributed and privacy-preserving way using a statistical classification model which is evolved by training with examples of both normal usage patterns and unusual behavior. The system is evaluated using the MIT reality mining data set. The results indicate that the distributed learning system trains quickly and performs reliably. Moreover, it is robust against failures of individual components.

This paper investigates the properties of social choice functions that represent resource allocation strategies in interference coupled wireless systems. The allocated resources can be physical layer parameters such as power vectors or antenna weights. Strategy proofness and efficiency of social choice functions are used to capture the respective properties of resource allocation strategy outcomes being non-manipulable and Pareto optimal. In addition, this paper introduces and investigates the concepts of (strict) intuitive fairness and non-participation in interference coupled systems. The analysis indicates certain inherent limitations when designing strategy proof and efficient resource allocation strategies, if the intuitive fairness and non-participation are imposed. These restrictions are investigated in an analytical social choice function framework for interference coupled wireless systems. Among other results, it is shown that a strategy proof and efficient resource allocation strategy for interference coupled wireless systems cannot simultaneously satisfy continuity and the frequently encountered property of non-participation.

This paper studies a constrained optical signal-to-noise ratio (OSNR) optimization problem in optical networks from the perspective of system performance. A system optimization problem is formulated with the objective of achieving an OSNR target for each channel while satisfying the total power constraint. In order to establish existence of a unique optimal solution, the conditions are derived, which can be used as a basis for an admission control scheme. The original problem is then converted to a relaxed system problem by using a barrier function and solved by a distributed iterative algorithm. Next, the system optimization framework developed is compared to the game theoretic one in [1]. The effects of parameters in both formulations are investigated to study efficiency of Nash equilibria in the OSNR game and pricing mechanisms affecting overall system performance. The theoretical analysis is supported by numerical simulations and experiments conducted on an optical fiber link.

This paper develops a framework for analyzing security risk dependencies in organizations and ranking the risks. The framework captures how risk `diffuses' via complex interactions and reaches an equilibrium by introducing a risk-rank algorithm. A conceptual structure of an organization-comprised of business units, security threats/vulnerabilities, and people-is leveraged for modeling risk dependencies and cascades. The risk-rank algorithm captures risk diffusion over time and ranks various risks based on a balancing of the immediate risk versus the future one emerging via cascading across system dependencies. Thus, the presented framework facilitates a systematic prioritization of risks in organizations.

We study two-player security games which can be viewed as sequences of nonzero-sum matrix games played by an attacker and a defender. At each stage of the game iterations, the players make imperfect observations of each other's previous actions. The underlying decision process can be viewed as a fictitious play (FP) game, but what differentiates this class from the standard one is that the communication channels that carry action information from one player to the other, or the sensor systems, are error prone. Two possible scenarios are addressed in the paper: (i) if the error probabilities associated with the sensor systems are known to the players, then our analysis provides guidelines for each player to reach a Nash equilibrium (NE), which is related to the NE of the underlying static game; (ii) if the error probabilities are not known to the players, then we study the effect of observation errors on the convergence to the NE and the final outcome of the game. We discuss both the classical FP and the stochastic FP, where for the latter the payoff function of each player includes an entropy term to randomize its own strategy, which can be interpreted as a way of concealing its true strategy.

High variability of access resources in heterogeneous wireless networks and limited computing power and battery life of mobile computing devices such as smartphones call for novel approaches to satisfy the quality-of-service requirements of emerging wireless services and applications. Towards this end, we first investigate a Markov-based stochastic scheme for modeling and estimation of bandwidth and delay on heterogeneous wireless networks. Borrowing clustering techniques from machine learning literature for intelligent state quantization, we demonstrate that the performance of the Markov model is enhanced significantly. We implement a measurement tool Zeus on smartphones and collect real-world data on 802.11g, 2.5G, and 3G wireless networks. The accuracy of the developed model is evaluated through simulation studies based on the collected data. Furthermore, a distributed rate-control scheme leveraging the predictions of our model is developed and observed to be much more efficient than a baseline additive-increase multiplicative- decrease scheme.

This paper presents a system that combines mobile services, ubiquitous computing, and augmented reality concepts in order to bring the kind of information-rich environment, which is currently limited to computer screens, to the physical world. This objective is achieved by constructing an architectu

We investigate a novel robust flow control framework for heterogeneous network access by devices with multi-homing capabilities. Towards this end, we develop an H-infinity-optimal control formulation for allocating rates to devices on multiple access networks with heterogeneous time-varying characteristics. H-infinity analysis and design allow for the coupling between different devices to be relaxed by treating the dynamics for each device as independent of the others. Thus, the distributed end-to-end rate control scheme proposed in this work relies on minimum information and achieves fair and robust rate allocation for the devices. An efficient utilization of the access networks is established through an equilibrium analysis in the static case. We perform measurement tests to collect traces of the available bandwidth on various WLANs and Ethernet. Through simulations, our approach is compared with AIMD and LQG schemes. In addition, the efficiency, fairness, and robustness of the H-infinity-optimal rate controller developed are demonstrated via simulations using the measured real world network characteristics. Its favorable characteristics and general nature indicate applicability of this framework to a variety of networked systems for flow control.

We consider the problem of rate allocation among multiple simultaneous video streams sharing multiple heterogeneous access networks. We develop and evaluate an analytical framework for optimal rate allocation based on observed available bit rate (ABR) and round-trip time (RTT) over each access network and video distortion-rate (DR) characteristics. The rate allocation is formulated as a convex optimization problem that minimizes the total expected distortion of all video streams. We present a distributed approximation of its solution and compare its performance against Hinfin-optimal control and two heuristic schemes based on TCP-style additive-increase-multiplicative-decrease (AIMD) principles. The various rate allocation schemes are evaluated in simulations of multiple high-definition (HD) video streams sharing multiple access networks. Our results demonstrate that, in comparison with heuristic AIMD-based schemes, both media-aware allocation and Hinfin-optimal control benefit from proactive congestion avoidance and reduce the average packet loss rate from 45% to below 2%. Improvement in average received video quality ranges between 1.5 to 10.7 dB in PSNR for various background traffic loads and video playout deadlines. Media-aware allocation further exploits its knowledge of the video DR characteristics to achieve a more balanced video quality among all streams.

Compared to only a few years ago, today there is an abundance of annotated image data available on the Internet. For researchers on image retrieval, this is an unforseen but welcome consequence of the rise of Web 2.0 technologies. Popular social networking and content sharing services seem to hold the key to the integration of context and semantics into retrieval. However, at least for now, it appears that this promise has to be taken with a grain of salt. In this paper, we present preliminary empirical results on the tagging behavior of power users of content sharing and social bookmarking services. Our findings suggest different promising research directions for image retrieval and we briefly discuss some of them.

We present a distributed machine learning framework based on support vector machines that allows classification problems to be solved iteratively through parallel update algorithms with minimal communication overhead. Decomposing the main problem into multiple relaxed subproblems allows them to be simultaneously solved by individual computing units operating in parallel and having access to only a subset of the data. A sufficient condition is derived under which a synchronous, discrete-time gradient update algorithm converges to the approximate solution. We apply the proposed distributed learning framework in the context of automatic image tagging as a first processing layer. Initial results from corresponding experiments indicate that he proposed framework has favorable properties including efficiency, configurability, robustness, suitability for online learning, and low communication overhead.

We propose a method for the detection of trends in social bookmarking systems. Compared to other work in this emerging field, our approach has a more sound statistical basis. In order to cope with the problem of vanishing probabilities due to data sparsity, we apply smoothing and show that it allows for an easy calibration of our trend detector resulting in better generalization and scalability. We test our approach on a collection of 105, 000, 000 bookmarks collected from the del.icio.us bookmarking service. To our knowledge, this is the largest corpus of a real world bookmarking service analyzed in this context. The results show that our method outperforms previously proposed methods and successfully detects trends in the data.

We introduce a lightweight biometric solution for user authentication over networks using online handwritten signatures. The algorithm proposed is based on a modified Hausdorff distance and has favorable characteristics such as low computational cost and minimal training requirements. Furthermore, we investigate an information theoretic model for capacity and performance analysis for biometric authentication which brings additional theoretical insights to the problem. A fully functional proof-of-concept prototype that relies on commonly available off-the-shelf hardware is developed as a client-server system that supports Web services. Initial experimental results show that the algorithm performs well despite its low computational requirements and is resilient against over-the-shoulder attacks.

The World Wide Web, which has started as a document repository, is rapidly transforming to a full fledged virtual environment that facilitates services, interaction, and communication. Under this light, the Semantic Web and Web 2.0 movements can be seen as intermediate steps of a natural evolution towards a new paradigm, the 3D Internet. We provide an overview of the concept 3D Internet and discuss why it is a goal worth pursuing, what it does entail, and how one can realize it. Our goal in this paper is to discuss a research agenda and raise interest in networking, security, distributed computing, and machine learning communities. We explore first the motivation for the 3D Internet and the possibilities it brings. Subsequently, we investigate the specific system level and research challenges that need to be addressed in order to make the 3D Internet a reality.

We present a cooperative intrusion detection approach inspired by biological immune system principles and P2P communication techniques to develop a distributed anomaly detection scheme. We utilize dynamic collaboration between individual artificial immune system (AIS) agents to address the well-known false positive problem in anomaly detection. The AIS agents use a set of detectors obtained through negative selection during a training phase and exchange status information and detectors on a periodical and event-driven basis, respectively. This cooperation scheme follows peer-to-peer communication principles in order to avoid a single point of failure and increase the robustness of the system. We illustrate our approach by means of two specific example scenarios in a novel network security simulator.

Contemporary wireless devices integrate multiple networking technologies, such as cellular, WiMax and IEEE 802.11a/b/g, as alternative means of accessing the Internet. Efficient utilization of available bandwidth over heterogeneous access networks is important, especially for media streaming applications with high data rates and stringent delay requirements. In this work we consider the problem of rate allocation among multiple video streaming sessions sharing multiple access networks. We develop and evaluate an analytical framework for optimal video rate allocation, based on observed available bit rate (ABR) and round trip time (RTT) over each access network, as well as the video distortion-rate (DR) characteristics. The rate allocation is formulated as a convex optimization problem that minimizes the sum of expected distortion of all video streams. We then present a distributed approximation of the optimization, which enables autonomous rate allocation at each device in a media- and network-aware fashion. Performance of the proposed allocation scheme is compared against robust rate control based on H∞ optimal control and two heuristic schemes employing TCP style additive-increase-multiplicative-decrease (AIMD) principles. Wesimulate in NS-2 [1] simultaneous streaming of multiple high-definition(HD) video streams over multiple access networks, using ABR and RTT traces collected on Ethernet, IEEE 802.11g, and IEEE 802.11b networks deployed in a corporate environment. In comparison with heuristic AIMD-based schemes, rate allocation from both the media-aware convex optimization scheme and H∞ optimal control benefit from proactive avoidance of network congestion, and can reduce the average packet loss ratio from 27% to below 2%, while improving the average received video quality by 3.3 - 4.5 dB in PSNR.

This paper presents an expert peering system for information exchange in the knowledge society. Our system realizes an intelligent, real-time search engine for enterprise Intranets or online communities that automatically relays user queries to knowledgable specialists. According to its very nature, the system requires a sound integration of concepts drawn from various areas of Computer Science. In addition to our solutions to problems in scalable processing, data transfer, and networking, we also address issues of interface design and usability, as well as aspects of machine intelligence. Results obtained from extensive experiments demonstrate the efficiency and robustness of our system and convey its potential for next generation web services.

We propose a hierarchical approach to document categorization that requires no pre-configuration and maps the semantic document space to a predefined taxonomy. The utilization of search engines to train a hierarchical classifier makes our approach more flexible than existing solutions which rely on (human) labeled data and are bound to a specific domain. We show that the structural information given by the taxonomy allows for a context aware construction of search queries and leads to higher tagging accuracy. We test our approach on different benchmark datasets and evaluate its performance on the single- and multi-tag assignment tasks. The experimental results show that our solution is as accurate as supervised classifiers for web page classification and still performs well when categorizing domain specific documents.

We consider a scenario where devices with multiple networking capabilities access networks with heterogeneous characteristics. In such a setting, we address the problem of efficient utilization of multiple access networks (wireless and/or wireline) by devices via optimal assignment of traffic flows with given utilities to different networks. We develop and analyze a device middleware functionality that monitors network characteristics and employs a Markov Decision Process (MDP) based control scheme that in conjunction with stochastic characterization of the available bit rate and delay of the networks generates an optimal policy for allocation of flows to different networks. The optimal policy maximizes, under available bit rate and delay constraints on the access networks, a discounted reward which is a function of the flow utilities. The flow assignment policy is periodically updated and is consulted by the flows to dynamically perform network selection during their lifetimes. We perform measurement tests to collect traces of available bit rate and delay characteristics on Ethernet and WLAN networks on a work day in a corporate work environment. We implement our flow assignment framework in ns-2 and simulate the system performance for a set of elastic video-like flows using the collected traces. We demonstrate that the MDP based flow assignment policy leads to significant enhancement in the QoS provisioning (lower packet delays and packet loss rates) for the flows, as compared to policies which do not perform dynamic flow assignment but statically allocate flows to different networks using heuristics like average available bit rate on the networks.

We present a game-theoretic treatment of distributed power control in CDMA wireless systems using outage probabilities. We first prove that the noncooperative power control game considered admits a unique Nash equilibrium (NE) for uniformly strictly convex pricing functions and under some technical assumptions on the SIR threshold levels. We then analyze global convergence of continuous-time as well as discrete-time synchronous and asynchronous iterative power update algorithms to the unique NE of the game. Furthermore, we show that a stochastic version of the discrete-time update scheme, which models the uncertainty due to quantization and estimation errors, converges almost surely to the unique NE point. We finally investigate and demonstrate the convergence and robustness properties of these update schemes through simulation studies.

We study power control in multicell CDMA wireless networks as a team optimization problem where each mobile attains its individual fixed target SIR level by transmitting with minimum possible power level. We derive conditions under which the power control problem admits a unique feasible solution. Using a Lagrangian relaxation approach similar to F. Kelly et al. (1998) we obtain two decentralized dynamic power control algorithms: primal and dual power update, and establish their global stability utilizing both classical Lyapunov theory and the passivity framework [J.T. Wen and M. Arcak, February 2004]. We show that the robustness results of passivity studies [(X. Fan et al., July 2004), (X. Fan et al., 2004)] as well as most of the stability and robustness analyses of F. Kelly et al. (1998) in the literature are applicable to the power control problem considered. In addition, some of the basic principles of call admission control are investigated from the perspective of the model adopted in this paper. We illustrate the proposed power control schemes through simulations.

This paper initiates a study toward developing and applying randomized algorithms for stability of high-speed communication networks. The focus is on congestion and delay-based flow controllers for sources, which are "utility maximizers" for individual users. First, we introduce a nonlinear algorithm for such source flow controllers, which uses as feedback aggregate congestion and delay information from bottleneck nodes of the network, and depends on a number of parameters, among which are link capacities, user preference for utility, and pricing. We then linearize this nonlinear model around its unique equilibrium point and perform a robustness analysis for a special symmetric case with a single bottleneck node. The "symmetry" here captures the scenario when certain utility and pricing parameters are the same across all active users, for which we derive closed-form necessary and sufficient conditions for stability and robustness under parameter variations. In addition, the ranges of values for the utility and pricing parameters for which stability is guaranteed are computed exactly. These results also admit counterparts for the case when the pricing parameters vary across users, but the utility parameter values are still the same. In the general nonsymmetric case, when closed-form derivation is not possible, we construct specific randomized algorithms which provide a probabilistic estimate of the local stability of the network. In particular, we use Monte Carlo as well as quasi-Monte Carlo techniques for the linearized model. The results obtained provide a complete analysis of congestion control algorithms for internet style networks with a single bottleneck node as well as for networks with general random topologies.

In this paper, we develop, analyze and implement a congestion control scheme in a noncooperative game framework, where each user's cost function is composed of a pricing function proportional to the queueing delay experienced by the user, and a fairly general utility function which captures the user demand for bandwidth. Using a network model based on fluid approximations and through a realistic modeling of queues, we establish the existence of a unique equilibrium as well as its global asymptotic stability for a general network topology, where boundary effects are also taken into account. We also provide sufficient conditions for system stability when there is a bottleneck link shared by multiple users experiencing nonnegligible communication delays. In addition, we study an adaptive pricing scheme using hybrid systems concepts. Based on these theoretical foundations, we implement a window-based, end-to-end congestion control scheme, and simulate it in ns-2 network simulator on various network topologies with sizable propagation delays.

In this paper, we develop, analyze and implement a congestion control scheme obtained in a noncooperative game framework where each user's cost function is composed of a pricing function, proportional to the queueing delay experienced by the user, and a fairly general utility function which captures the user demand for bandwidth. Using a network model based on fluid approximations and through a realistic modeling of queues, we establish the existence of a unique equilibrium as well as its global asymptotic stability for a general network topology. We also provide sufficient conditions for system stability when there is a bottleneck link shared by multiple users experiencing nonnegligible communication delays. Based on these theoretical foundations, we implement a window-based, end-to-end congestion control scheme, and simulate it in ns-2 network simulator on various network topologies with sizable propagation delays.