Cloud AI services have experienced rapid development, raising concerns to privacy protection of cloud tenants. Many proposals have been made to use the NPU Trusted Execution Environment (TEE) to protect AI workloads on the cloud. However, existing designs typically bind NPUs exclusively to a single tenant, preventing multiple tenants from sharing the computing power of the TEE-NPUs.As such, we have designed a novel TEE for discrete NPUs, named NXT (NPU eXtension for Trust), which breaks the exclusive binding architecture and allows multiple tenants to securely share the TEE-NPU cluster. Firstly, we introduced an NPU Trusted Agent (NTA) to the most privileged level of the TEE system to assist in the global scheduling of the TEE-NPU cluster. Secondly, we implemented a flexible isolation mechanism to provide security for multi-tenant fine-grained sharing of NPU resources. Thirdly, we support efficient communication between workloads within TEE-NPUs and with legacy NPUs, accelerating multi-workload collaborative computing. We evaluated NXT by extending gem5 and a cycle-accurate NPU simulator to build a prototype. Results show that NXT improves overall utilization by up to 3.49× and ANTT by up to 7.24×, with only 6.38% average overhead for scheduling and isolation. It also boosts parallel inference performance by 58.3% for GPT-2(XL) and 63.6% for LLaMA-13B compared to static protection schemes.

Federated edge learning (FEEL) is an advanced paradigm in edge artificial intelligence, enabling privacy-preserving collaborative model training through periodic communication between edge devices and a central server. FEEL involves three key processes: 1) sensing; 2) computation; and 3) communication for data acquisition, processing, and exchange, respectively. Due to limited system resources, optimizing each process individually may lead to suboptimal learning performance. This challenge has sparked research into integrated sensing-computation–communication (ISCC) design for enhanced FEEL. While previous work has optimized general learning parameters, such as batch size and computing frequency, there is a lack of customized designs considering the neural network architecture as an optimizable variable in ISCC for FEEL. To close this gap, we introduce a novel design where each device generates a submodel through controllable weight dropout, adding flexibility by directly manipulating the learning process and reducing computation and communication overhead. To guide ISCC resource allocation in this new setting, we present a comprehensive convergence analysis, revealing the tight coupling of sensing, computation, and communication across devices and their impact on FEEL convergence. Building on these theoretical insights, we formulate an ISCC problem aiming to maximize the FEEL convergence rate through joint optimization of variables, such as batch size, sensing power, dropout rate, and communication power. This nonconvex problem is decomposed into two subproblems via alternating optimization: one controls batch size using a sorting algorithm, while the other focuses on ISCC device parameters, transformable into a convex problem solved by successive convex approximation. Extensive experiments using human motion recognition datasets demonstrate the superiority of the proposed design over baseline schemes.

With the widely application of cloud, a series of privacy challenges arise. Generally, encryption methods are used to ensure privacy, which may result in high computation and communication overheads. Access control is another fundamental and important measure to protect resources. Usually cloud computing systems are managed through RESTful web services and users can conduct access control measures like role-based access control (RBAC) to manage the permissions to RESTful resources. By running integration test, test cases and the corresponding RESTful permissions can be parsed out automatically. We are the first to define the role engineering problem based on integration test and summarize three metrics for role engineering. Then we propose a novel role engineering method based on spectral clustering analysis which supporting more feature set such as permission weight, role hierarchy and customized number of roles. Finally, we conduct experiments using real integration test on three cloud computing systems to demonstrate the effectiveness and performance, outperforming prior works.

Sensing, computation, and communication (SC2) are highly coupled processes in federated edge learning (FEEL) and need to be jointly designed in a task-oriented manner for pursuing the best FEEL performance under the stringent resource constraints at edge devices. However, this remains an open problem as there is a lack of theoretical understanding on how the SC2resources jointly affect the FEEL performance. In this paper, we address the problem of joint SC2resource allocation for FEEL via a concrete case study of human motion recognition based on wireless sensing. Specifically, the joint SC2resource allocation problem is cast to maximize the convergence speed of FEEL, under the constraints on training time and energy supply of each edge device. Solving this problem entails solving two subproblems in order: the first one reduces to determining a joint sensing and communication resource allocation that maximizes the total number of samples sensed during the entire training process; the second one concerns the partition of the total number of sensed samples over communication rounds to determine the batch size at each round for convergence speed maximization. Finally, extensive simulation results are provided to validate the superiority of the proposed scheme over several baseline schemes.

As cloud services have become popular, and their adoption is growing, consumers are becoming more concerned about the cost of cloud services. Cloud Service Providers (CSPs) generally use a pay-per-use billing scheme in the cloud services model: consumers use resources as they needed and are billed for their resource usage. However, CSPs are untrusted and privileged; they have full control of the entire operating system (OS) and may tamper with bills to cheat consumers. So, how to provide a trusted solution that can keep track of and verify the consumers’ resource usage has been a challenging problem. In this article, we propose a T-Counter framework based on Intel SGX. The T-Counter allows applications to construct a trusted solution to measure its CPU usage by itself in cloud computing. These constructed applications are instrumented with counters in basic blocks and added three components in trusted parts to count instructions and defend against malicious CSPs’ manipulations. We propose two algorithms which selectively instrument counters in the CFG. T-Counter is implemented as an extension of the LLVM framework and integrated with the SGX SDK. Theoretical analyses and evaluations show that T-Counter can effectively measure CPU usage and defend against malicious CSPs’ manipulations.

Nowadays, data privacy is one of the most critical concerns in cloud computing, and many privacy-preserving distributed computing systems based on the trusted execution environment (e.g., Intel SGX) have been proposed to protect the user's privacy during cloud-outsourced computation. However, these SGX-based solutions are vulnerable to some traffic analyses, and loading all tasks into the enclave introduces much overhead for frequent EPC-paging. In this paper, we propose a T-SGX framework, which keeps the confidentiality of a distributed job and guarantees the system efficiency by allowing dynamically loading an enclave shared object for the task under processing. In T-SGX, all these objects are secretly shared and stored in a verifiably distributed share management system (SMS) outside the TCB. To mitigate the exposure of sensitive information, we present an efficient oblivious transfer (OT) protocol under the Decisional Diffie-Hellman (DDH) assumption for obliviously transmitting desired shares. Detailed security analysis demonstrates that the proposed T-SGX achieves the goal of secure distributed computing without privacy leakage to unauthorized parties. Finally, we benchmark the framework in six real-world applications, and the experimental results show that T-SGX significantly outperforms a state-of-the-art solution, with 11.9%-29.7% less overhead performing an SGX-based application.

In this paper, the training time minimization problem is investigated in a quantized FEEL system, where the heterogeneous edge devices send quantized gradients to the edge server via orthogonal channels. In particular, a stochastic quantization scheme is adopted for compression of uploaded gradients, which can reduce the burden of per-round communication but may come at the cost of increasing number of communication rounds. The intrinsic trade-off between the number of communication rounds and per-round latency is characterized. Specifically, we analyze the convergence behavior of the quantized FEEL in terms of the optimality gap. Constrained by total bandwidth, the training time minimization problem is formulated as a joint quantization level and bandwidth allocation optimization problem. To this end, an algorithm based on alternating optimization is proposed, which alternatively solves the subproblem of quantization optimization via successive convex approximation and the subproblem of bandwidth allocation via bisection search. With different learning tasks and models, the validation of our analysis and the near-optimal performance of the proposed algorithm are demonstrated by the experimental results.

Training a machine learning model with federated edge learning (FEEL) is typically time consuming due to the constrained computation power of edge devices and the limited wireless resources in edge networks. In this study, the training time minimization problem is investigated in a quantized FEEL system, where heterogeneous edge devices send quantized gradients to the edge server via orthogonal channels. In particular, a stochastic quantization scheme is adopted for compression of uploaded gradients, which can reduce the burden of per-round communication but may come at the cost of increasing the number of communication rounds. The training time is modeled by taking into account the communication time, computation time, and the number of communication rounds. Based on the proposed training time model, the intrinsic trade-off between the number of communication rounds and per-round latency is characterized. Specifically, we analyze the convergence behavior of the quantized FEEL in terms of the optimality gap. Furthermore, a joint data-and-model-driven fitting method is proposed to obtain the exact optimality gap, based on which the closed-form expressions for the number of communication rounds and the total training time are obtained. Constrained by the total bandwidth, the training time minimization problem is formulated as a joint quantization level and bandwidth allocation optimization problem. To this end, an algorithm based on alternating optimization is proposed, which alternatively solves the subproblem of quantization optimization through successive convex approximation and the subproblem of bandwidth allocation by bisection search. With different learning tasks and models, the validation of our analysis and the near-optimal performance of the proposed optimization algorithm are demonstrated by the simulation results.

We investigate the downlink transmission for multi-user multi-input multi-out (MU-MIMO) system, in which the regularized zero forcing (RZF) precoder is adopted and the power allocation and regularization factor are optimized. Our aim is to find a power allocation and regularization factor control policy that can minimize the long-term average power consumption subject to long-term delay constraint for each user. The induced optimization problem is formulated as a constrained Markov decision process (CMDP), which is efficiently solved by the proposed constrained deep reinforcement learning algorithm, called successive convex approximation policy optimization (SCAPO). The SCAPO is based on solving a sequence of convex objective/feasibility optimization problems obtained by replacing the objective and constraint functions in the original problems with convex surrogate functions. At each iteration, the SCAPO merely needs to estimate the first-order information and solve a convex surrogate problem that can be efficiently parallel tackled. Moreover, the SCAPO enables to reuse old experiences from previous updates, thereby significantly reducing the implementation cost. Numerical results have shown that the novel SCAPO can achieve the state-of-the-art performance over advanced baselines.

Container technology has been used for running multiple isolated operating system distros on a host or deploying large scale microservice-based applications. In most cases, containers share the same kernel with the host and other containers on the same host, and the application in the container can make system calls of the host kernel like a normal process on the host. Seccomp is a security mechanism for the Linux kernel, through which we can prohibit certain system calls from being executed by the program. Docker began to support the seccomp mechanism from version 1.10 and disables around 44 system calls out of 300+ by default. However, for a particular container, there are still many system calls that are unnecessary for running it allowed to be executed, and the abuse of system calls by a compromised container can trigger the security vulnerabilities of a host kernel. Unfortunately, Docker does not provide a way to get the necessary system calls for a particular container. In this paper, we propose RSDS, a method combining dynamic analysis and static analysis to get the necessary system calls for a particular container. Our experiments show that our solution can reduce system calls by 69.27%-85.89% compared to the default configuration on an x86-64 PC with Ubuntu 16.04 host OS and does not affect the functionalities of these containers.

Recent human action recognition models mainly focus on exploiting human features, such as pose or skeleton features. However, most of these methods do not pay enough attention to action-related backgrounds. In this work we propose a novel multi-stream features fusion framework based on discriminative patch exploiting. Unlike existing part-based or attention-based multi-stream methods, our work improves the recognition accuracy by 1) Paying more attention to exploiting of global and local discriminative patches, which include not only the acting human but also the interactive scenes. 2) Proposing an effective multi-stream feature pooling and fusion mechanism: 2D and 3D features from RGB frames and discriminative patches are combined to enhance spatial-temporal feature representation ability. Our framework is evaluated on two widely used video action benchmarks, where it outperforms other state-of-the-art methods: the accuracy up to 87.8% at HMDB51, and 98.8% at UCF101.

Real-world anomalous events are complicated, diverse, and rarely occurred. The main challenge to anomaly detection is to learn normal and anomalous patterns accurately. In this work, we propose discriminative clip mining for anomaly detection and classification: firstly, by introducing clip-level class activation mapping, an efficient Discriminative Anomalous Clip Miner (DACM) is developed to mine discriminative anomalous clips from a large number of normal ones; secondly, with the mined discriminative clips, an attentive ranking loss is designed to increase the anomalous instances hit rate of the traditional Multiple Instance Learning (MIL) model. Furthermore, by integrating the DACM and attentive MIL, one novel anomaly detection framework is proposed to learn more contrastive anomalous and normal patterns, and thus higher recognition performance can be achieved. Our experimental results on the widely-used UCF-Crime dataset show that, as compared to the state-of-the-art approaches, the proposed method achieves competitive performance both in anomaly detection and anomalous activity classification.

This paper investigates a X-duplex decode-and-forward relay system in the presence of direct link from the source to destination. X-duplex relays can adaptively switch between half-duplex mode and full-duplex mode according to the instantaneous channel conditions. Unlike previous work on X-duplex relay, we treat the direct link as an additional information path and propose a new transmission scheme based on dirty paper coding (DPC). By using DPC as the precoding scheme at the source, the messages can be divided into two parts which are sent from the source to destination through the relay link and direct link, respectively. In addition, the destination performs successive interference cancellation as the decoding strategy. The optimal transmit powers at the source and relay by maximizing the end-to-end achievable rate are obtained. The numerical results show that the new transmission scheme achieves a better performance over the reference scheme.

We propose a novel randomized channel sparsifying hybrid precoding (RCSHP) design to reduce the signaling overhead of channel estimation and the hardware cost and power consumption at the base station (BS), in order to fully harvest benefits of frequency division duplex (FDD) massive multiple-input multiple-output (MIMO) systems. RCSHP allows time-sharing among multiple analog precoders, each serving a compatible user group. The analog precoder is adapted to the channel statistics to properly sparsify the channel for the associated user group, such that the resulting effective channel (product of channel and analog precoder) not only has enough spatial degrees of freedom (DoF) to serve this group of users, but also can be accurately estimated under the limited pilot budget. The digital precoder is adapted to the effective channel based on the duality theory to facilitate the power allocation and exploit the spatial multiplexing gain. We formulate the joint optimization of the time-sharing factors and the associated sets of analog precoders and power allocations as a general utility optimization problem, which considers the impact of effective channel estimation error on the system performance. Then we propose an efficient stochastic successive convex approximation algorithm to provably obtain Karush-Kuhn-Tucker (KKT) points of this problem.

In this paper, we investigate the hard combining scheme of cooperative spectrum sensing in a practical scenario where secondary users (SUs) may use different local detectors and have different signal-to-noise ratios (SNRs). For simplicity, energy detector and pilot-based detector are considered for local quantization. The suboptimal AND and OR rules are implemented in the fusion center (FC). We point out that the necessary condition of the optimal local threshold in distributed detection can be extended to the heterogeneous cooperative sensing network by some modification. Subsequently, two iterative threshold optimization algorithms are extended from methods in conventional cooperative sensing scenarios: the ascending order of Bayesian risk optimization (AOBO) and the person-by-person optimization (PBPO). Simulation results show that both the AOBO and PBPO outperform the uniform-threshold approach, when the sensing abilities at SUs are much different. When all SUs have an identical SNR and only use energy detectors, the PBPO method has a same near optimal performance compared to the uniform-threshold approach.

Recent human action recognition methods mainly model a two-stream or 3D convolution deep learning network, with which humans spatial-temporal features can be exploited and utilized effectively. However, due to the ignoring of interaction exploiting, most of these methods cannot get good enough performance. In this paper, we propose a novel action recognition framework with Graph Convolutional Network (GCN) based Interaction Reasoning: Objects and discriminative scene patches are detected using an object detector and class active mapping (CAM), respectively; and then a GCN is introduced to model the interaction among the detected objects and scene patches. Evaluation of two widely used video action benchmarks shows that the proposed work can achieve comparable performance: the accuracy up to 43.6% at EPIC Kitchen, and 47.0% at VLOG benchmark without using optical flow, respectively.

Trajectories and spatiotemporal attention model have been successfully used in skeleton-based action recognition. Most existing methods focus more attention on temporal structure mining. However, only a few local joints and their position features (e.g., critical position changes of hand, head, leg etc.) are responsible for the action label. In this work, we introduce a novel action recognition framework using Key Joints Selection and Spatiotemporal Mining, which can identify both key joints and their position & velocity histogram as well as trajectory features for action classification. First, histogram of human joints position and velocity are developed to enhance the spatiotemporal structure representation of existing trajectory-based methods. Second, the key joints are selected according to their information gains, and then their position & velocity histograms are weighted and composed with trajectory features to form one richer representation for final action classification. Experiments on two widely-tested benchmark datasets show that by combining the strength of both richer features and key joints selecting, our method can achieve state-of-the-art or competitive performance compared with existing results using sophisticated models such as deep learning, with advantages regarding the recognition accuracy and robustness.

The joint multicast beamforming and user scheduling problem is studied, with the objective of minimising total transmitting power across multiple channels by jointly assigning each user to appropriate channel and designing multicast beamformer for each channel. The problem of interest is formulated in two different optimisation problems, a mixed binary quadratically constrained quadratic program and a highly‐structured non‐smooth program. Two different algorithms, based on convex relaxation and convex restriction, respectively, are proposed to solve the problem. The performance ratio between the approximate solution provided by the convex‐relaxation‐based algorithm and optimal solution is proven to be upper bounded by a constant independent of problem data. The convex‐restriction‐based algorithm is guaranteed to converge to a critical point to the non‐smooth formulation problem. Finally, extensive simulation results verify the theoretical analysis and demonstrate the advantage of the proposed co‐design scheme over conventional fixed scheduling and random scheduling in terms of power consumption.

Failing to promote the least privilege principle in administration can lead to substantial vulnerabilities in cloud computing. A malicious insider like a compromised cloud administrator can affect security of data and workloads belonging to cloud customers. Enforcing the least privilege principle in cloud administration can fairly restrict the permissions of administrators and reduce the attack surface. However, writing a least privilege policy can be hard and error prone for cloud service providers. In this paper, we propose a framework called Least Privilege for Cloud (LPCloud) to address these concerns. LPCloud automatically produces policies for minimization of administrators' privileges at the granularity of representational state transfer (REST) application program interfaces (API), and enforces the policies without affecting current systems. Specifically, we introduce a novel algorithm to partition privileges based on dependencies between API calls. This paper presents design of LPCloud, including a service called Policy Generator which produces partitioned policies and a component named Policy Enforcer to enforce the policies. We implement a prototype of our framework in OpenStack Mitaka. Experiments indicate that LPCloud can produce proper policies to enforce the least privilege principle. Meantime, the average performance overhead is 10.1% which is in acceptable level.

The access control mechanisms of existing cloud systems, mainly OpenStack, fail to provide two key factors: i) centralized access mediation and ii) flexible policy customization. This situation prevents cloud administrators and end customers from enhancing their security. Furthermore, a variety of clouds have implemented their access control systems and policies in separated ways. This might confuse the customers whose businesses are built on multiple clouds, as they have to take efforts to accommodate their policies for different platforms. The OpenStack Security Modules (OSM) project has developed a least-invasive access control framework for OpenStack to enable different access control models to be implemented as loadable modules. This framework can be a good replacement of the existing permission checks in OpenStack and other platforms. We also propose an integration mechanism for multiple policies to form a single decision. This paper presents the design and implementation of OSM, including a new service called patron and an attachment module called access endpoint middleware (AEM). Experiments on the tempest benchmark indicate that OSM has improved the flexibility and security of policy management without affecting other services. Meantime, the average performance overhead remains as low as 7.3%, which is acceptable for practical use.

This paper discusses a typical cooperative sensing model where the local observations are conditionally dependent. The primary user sends a binary phase-shifted keying modulated signal and the hard combining strategy is implemented among second users. Two low-complexity local quantization schemes corresponding to the choice of quantizing either the amplitude or the sign as the one-bit reported message are compared. We show that the optimal fusion rule for quantizing the sign is non-monotone, which is different from the design for quantizing amplitude scheme. Based on large deviation theory, we finish the parameter optimization for both schemes and use the error exponent to evaluate the asymptotical performance. Several closed-form expressions are derived. Numerical results show that to quantize the sign is asymptotically optimal in low-SNR regime, while to quantize the amplitude performs well in high-SNR regime. This work shows that the dependence among local observations may greatly help the final decision in some case, and the non-monotone fusion rule may achieve better performance than the monotone one for dependent observations.

In this paper, downlink transmission schemes that are able to improve the energy efficiency of multi-cell massive distributed antenna systems (DAS) are investigated. We employ a power consumption model by considering the transmit power, the backhaul power, the uplink pilot transmit power and the circuit power, in contrast to existing works which focus on co-located antenna systems where the backhaul power is negligible. By applying random matrix theory and given the power consumption model, we derive the asymptotic energy efficiency achieved by two transmission schemes, first a full transmission scheme where all remote radio heads (RRHs) in a cell jointly transmit, and second single-RRH user association transmission scheme where each user is associated with only one RRH. A new algorithm to associate user with RRH is proposed based on the asymptotic energy efficiency. The proposed algorithm takes into consideration the power attached to antennas as well as the backhauling power of RRH. Simulations show that the proposed user association algorithm for DAS achieves higher energy efficiency than full transmission scheme and the baseline nearest RRH association scheme, especially when the number of antennas is large.

Massive multiple-input multiple-output (MIMO) is an attractive solution to achieve high data rate and tackle interference. In this paper, we study the downlink of a two-cell massive MIMO system employing coordinated zero-forcing beamforming with limited feedback, where each base station (BS) acquires both the direct link channel state information (DCSI) and the cross link CSI (CCSI) from the users through a limited feedback link. The DCSI is used to support multi-user MIMO transmission (control intra-cell interference) at each BS and the CCSI is used to control the inter- cell interference. Since the total number of feedback bits at each user is fixed, there is a tradeoff between the quantization errors of the DCSI and CCSI. We consider the allocation of the feedback bits over the DCSI and CCSI to maximize the sum-rate of the system. We obtain closed-form solution for the asymptotically optimal feedback bits allocation as the number of antennas per BS goes to infinity. The solution reveals insights on how the key system parameters such as path gains, transmit powers and the number of antennas/users affect the optimal feedback bits allocation. Numerical results validate our theoretical analysis.

ABSTRACT With the reformation of spectrum policy and the development of cognitive radio, secondary users will be allowed to access spectrums licensed to primary users. Spectrum auctions can facilitate this secondary spectrum access in a market‐driven way. To design an efficient auction framework, we first study the supply and demand pressures and the competitive equilibrium of the secondary spectrum market, considering the spectrum reusability. In well‐designed auctions, competition among participants should lead to the competitive equilibrium according to the traditional economic point of view. Then, a discriminatory price spectrum double auction framework is proposed for this market. In this framework, rational participants compete with each other by using bidding prices, and their profits are guaranteed to be non‐negative. A near‐optimal heuristic algorithm is also proposed to solve the auction clearing problem of the proposed framework efficiently. Experimental results verify the efficiency of the proposed auction clearing algorithm and demonstrate that competition among secondary users and primary users can lead to the competitive equilibrium during auction iterations using the proposed auction framework. Copyright © 2011 John Wiley & Sons, Ltd.

It is often impractical to obtain global channel knowledge and conduct centralized optimization for wireless networks. We study distributed weighted sum-rate maximization (WSRM) in general MIMO interference networks, named B-MAC interference networks. It is desirable to exploit the structure of the problem to design distributed algorithms with high performance and low signaling overhead. We recently unveiled a polite water-filling (PWF) structure satisfied by all Pareto optimal inputs of important achievable regions of the B-MAC interference networks. The PWF offers an elegant method to decompose a network into multiple equivalent single user channels and thus, facilitates the design of distributed algorithms. Based on the PWF, we design efficient distributed algorithms which only need local channel knowledge and converge to a stationary point of the WSRM problem. For TDD networks, the duality inherited in PWF and the channel reciprocity are further exploited to reduce the signaling overhead. The proposed algorithms are shown by simulations to outperform the state-of-the-art.

The algorithms in this paper exploit optimal input structure in interference networks and is a major advance from the state-of-the-art. Optimization under multiple linear constraints is important for interference networks with individual power constraints, per-antenna power constraints, and/or interference constraints as in cognitive radios. While for single-user MIMO channel transmitter optimization, no one uses general purpose optimization algorithms such as steepest ascent because water-filling is optimal and much simpler, this is not true for MIMO multiaccess channels (MAC), broadcast channels (BC), and the non-convex optimization of interference networks because the traditional water-filling is far from optimal for networks. We recently found the right form of water-filling, polite water-filling, for some capacity/achievable regions of the general MIMO interference networks, named B-MAC networks, which include BC, MAC, interference channels, X networks, and most practical wireless networks as special cases. In this paper, we use weighted sum-rate maximization under multiple linear constraints in interference tree networks, a natural extension of MAC and BC, as an example to show how to design highly efficiency and low complexity algorithms. Several times faster convergence speed and orders of magnitude higher accuracy than the state-of-the-art are demonstrated by numerical examples.

We found a network version of water-filling, named polite water-filling, that is optimal for all boundary points of the capacity regions of MAC and BC and for all boundary points of a set of achievable regions of a general class of interference networks, named MIMO B-MAC networks that include BC, MAC, interference channels, X networks, and most practical networks as special cases. It is polite because it strikes an optimal balance between reducing interference to others and maximizing a link's own rate. Unlike in single-user MIMO channels, where the optimal input covariance can be solved by the water-filling, the traditional water-filling is far from optimal in networks. Thus, general purpose optimization algorithms have been used for networks but have high complexity and do not work well for non-convex cases. Together with our duality result, the polite water-filling can be used to design highly efficient low-complexity iterative centralized/distributed algorithms for the optimization of input covariance matrices, including both power and beamforming matrices, because it takes the advantage of the structure of the problems. References to the resulting algorithms that outperform the state-of-the-art by a wide margin are provided.

It is well known that in general, the traditional water-filling is far from optimal in networks. We recently found the long-sought network version of water-filling named polite water-filling that is optimal for a large class of MIMO networks called B-MAC networks, of which interference Tree (iTree) networks is a subset whose interference graphs have no directional loop. iTree networks is a natural extension of both broadcast channel (BC) and multiaccess channel (MAC) and possesses many desirable properties for further information theoretic study. Given the optimality of the polite water-filling, general purpose optimization algorithms for networks are no longer needed because they do not exploit the structure of the problems. Here, we demonstrate it through the weighted sum-rate maximization. The significance of the results is that the algorithm can be easily modified for general B-MAC networks with interference loops. It illustrates the properties of iTree networks and for the special cases of MAC and BC, replaces the current steepest ascent algorithms for finding the capacity regions. The fast convergence and high accuracy of the proposed algorithms are verified by simulation.

In this paper, we establish a rate duality between the forward and reverse links of MIMO interference channel, where the reverse links are obtained by exchanging the roles of transmitters and receivers in the forward links, and the corresponding channel matrices are conjugate transpose of the forward channel matrices. Since the capacity region for general interference channel is unknown, we show that the forward and reverse links have the same achievable rate region by treating interference as noise under some sum power constraint. The explicit expression of the corresponding input covariance matrix transformation is provided. We discuss the connection between the proposed transformation and the MAC-BC transformations in the previous works. As an application, a duality based iterative algorithm is proposed to maximize the sum rate of MIMO interference channel under sum power constraint. We also extend the algorithm to individual power constraint. The proposed algorithms are shown to be effective by simulation.

It is difficult to implement optimal beamforming for multi-user multiple-input multiple-output (MIMO) downlink due to the high complexity. This paper proposes a low complexity generalized beamforming (GBF) scheme combined with an efficient user selection to maximize the weighted sum-rate. For each user, the outputs of the multiple antennas are combined with a receive GBF vector to create an equivalent multiple-input single-output (MISO) effective channel. First, user selection and receive GBF vectors are jointly optimized to construct a group of preferable effective channels. Then, transmit GBF vectors are obtained by zero-forcing over these effective channels. Simulation results show significant gain over currently known suboptimal schemes in various scenarios.