As edge-based deep learning applications become more complex, optimizing performance on heterogeneous System-on-Chips (SoCs) presents unique challenges. Traditional pipelining techniques distributing the computation across different on-chip processing units, while effective for throughput, do not address the latency demands posed by modern neural networks with complex interdependencies and extensive operator parallelism. There is a potential in leveraging operator parallelism to enable concurrent execution across multiple processing units, thereby reducing inference latency. However, prioritizing pipelining or parallel execution often necessitates a compromise, where optimizing one performance metric adversely impacts the other. This paper introduces Para-Pipe, a hierarchical mapping framework that integrates intra-and inter-stage operator parallelism within a pipelined architecture. Para-Pipe navigates the trade-off between throughput and latency by selectively fine-tuning parallelism levels within and across pipeline stages. This strategy can significantly reduce inter-processor communication overhead, significantly improving energy efficiency. Our evaluation demonstrates that Para-Pipe generates multiple Pareto-optimal configurations, achieving a balance between throughput and latency on an Amlogic SoC equipped with ARM big.LITTLE CPUs and GPU, as well as the Black Sesame Technology SoC featuring a deep learning accelerator and two DSPs. More importantly, throughput-optimized configurations under Para-Pipe on Amlogic SoC show an average energy efficiency improvement of 11.0% over purely pipelined strategies and 23.3% relative to non-pipelined parallel execution.

Quantum key distribution (QKD)-enabled optical networks utilize quantum mechanics to secure communications by establishing secure quantum channels. A critical criterion for QKD-enabled optical networks is ensuring network availability, which requires each quantum key service to adhere to a maximum unavailability constraint to maintain the network availability. This paper addresses the challenge of offering dedicated path protection for quantum key services in QKD-enabled optical networks. It ensures that the unavailability gap between the working and protection paths remains within the allowable limit for each quantum key service, adhering to specific unavailability constraints. Considering the constraints of the limited timeslot resources and network availability, we proposed quantum key service approach named a maximum availability (MA) algorithm. Simulation results indicate that the MA algorithm surpasses both the traditional dedicated-path protection (TDP) and fixed dedicated-path routing (FDR) algorithms in reducing total timeslot consumption and enhancing average availability.

This paper addresses the problems of core and spectrum assignment (CSA) in space-division-multiplexing (SDM) elastic optical networks (EONs) with multi-core fiber (MCF). We first present spectrum sensitivity (SS) of the crosstalk-sensitive core matrix to evaluate inter-core crosstalk (XT) in trench-assisted (TA) MCF. We then propose the XT-sensitive spectrum assignment to suppress XT. We consider both static and dynamic network planning based on the spectrum sensitivity. In the static scenario, an integer linear programming (ILP) model along with effective heuristic CSA algorithms are developed to suppress XT and improve the spectrum efficiency. In the dynamic scenario, two heuristic CSA algorithms are proposed aiming to initiate a graceful tradeoff between average crosstalk and spectrum efficiency. Simulation results demonstrate the superiority of the proposed CSA algorithms compared with the existing CSA algorithm in MCF-based SDM-EONs.

The spectrum efficiency is a greatly important issue when we establish connection requests in elastic optical data center networks (EODCNs). In this article, we address the spectrum efficiency problems of the shared-path protection with spectrum-sharing-maximized approaches for the optical network survivability. Two integer linear program (ILP) models, i.e., flow-based and path-based ILP models, named FB-ILP and PB-ILP, are developed to minimize the frequency slots (FSs) occupied with the spectrum-sharing-maximized protection, and two heuristic approaches with the spectrum-sharing-maximized protection (HA-SSMP) and with the general spectrum-sharing protection (HA-GSSP) are also proposed to improve spectrum efficiency in EODCNs. For comparison, we introduce a heuristic approach with the existing shared-path protection (HA-ESPP) in EODCNs. On the one hand, simulation results show that FB-ILP can minimize the number of FSs occupied, but leads to the higher average number of hops and longer running time compared to PB-ILP, HA-SSMP, HA-GSSP, and HA-ESPP in static traffic scenario in EODCNs. On the other hand, the simulation results of our proposed HA-SSMP are very close to the solutions of FB-ILP. In dynamic traffic scenario, simulation results show that HA-SSMP significantly improves the spectrum efficiency and effectively suppresses the blocking probability, but leads to the higher average number of hops compared to HA-GSSP and HA-ESPP in EODCNs.

In this article, the network performance of equilibrium allocation of quantum key resources was investigated in quantum key distribution (QKD)-enabled optical data center networks. To effectively use quantum key resources, we propose three novel efficient load balancing routing, wavelength, and time-slot assignment (LB-RWTA) approaches, including LB-RWTA with flexible security level (LB-RWTA-FSL), LB-RWTA with specific security level (LB-RWTA-SSL), and LB-RWTA without security level (LB-RWTA-NSL). Particularly, the proposed LB-RWTA-FSL approach is uniquely featured by adaptive security level classification (ASLC) and load balancing (LB), aiming to demarcate the security level (SL) and reduce the overall network congestion. We introduce an existing routing, wavelength, and time-slot assignment (E-RWTA) approach without SL, called E-RWTA, for comparison. Simulation results show the effectiveness of our proposed approaches in terms of much higher quantum key resource efficiency and thus much higher network security performance than the state-of-the-art quantum key resource allocation approaches compared with the E-RWTA approach in QKD-enabled optical data center networks.

The application of non-orthogonal multiple access (NOMA) in millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) systems can enhance spectral efficiency. In this paper, we propose an angle-domain NOMA scheme over the multi-cell mmWave massive MIMO networks. This scheme is optimized through both user scheduling and precoders/decoders design to maximize the system sum rate, where the precoders are decomposed into outer and inner ones. We construct the outer precoders with the help of the users' spatial angle information, i.e., beam signatures, and propose two design strategies for both inner precoders and decoders, i.e., joint optimization of precoders/decoders (JOPD) and cooperative NOMA (C-NOMA). Specifically, in JOPD, the precoders/decoders are obtained through maximizing a nonconvex function subject to the users' quality-of-service (QoS) constraints, where an alternate optimization algorithm based on the constrained concave-convex procedure is proposed for its solutions. In C-NOMA, we adopt interference alignment to cooperatively serve the cell-edge users and achieve simplified yet effective precoders/decoders. Furthermore, we optimize C-NOMA through power allocation. Afterwards, user scheduling algorithms are proposed for both JOPD and C-NOMA. Extensive simulations verify that the proposed schemes exhibit improved performance in terms of both sum rate and users' QoS compared to that of existing mmWave NOMA schemes.

In this paper, we propose an angle-domain NOMA transmission scheme over multicell massive MIMO systems, where multiple users' signal can be superposed to be served by the same spatially angle-domain beams. Then, we carefully consider the performance degradation resulting from the severe inter-cell interference, and formulate an optimization problem in terms of jointly optimizing precoders and decoders (JOPD) to seek an optimal transmission policy with quality of service (QoS) requirements of both cell-edge users and cell-center users, which consequently is a maximization of a nonconvex function with nonconvex constraints. To solve this challenging problem, we invoke the constrained concave convex procedure (CCCP) method to optimize precoders with fixed decoders, while the decoders can be readily optimized with obtained precoders. Consequently, we propose an alternating optimization algorithm based on CCCP (AoCCCP) to jointly optimize precoders and decoders and then obtain a suboptimal solution of the prime problem. Simulation results verify that the proposed scheme exhibits significant performance gain in terms of sum rate as well as QoS guarantee.

In this paper, we investigate a spatially sparse code multiplexing (SCM) transmission scheme for the massive multiple-input multiple-output (MIMO) networks to enhance the access connectivity. We construct a non- orthogonal transmission policy over both power and angle domains to fully utilize the limited angle-domain degree of freedom (DoF). Firstly, the mapping structure in the angle domain and the detection method are presented. Then, we formulate an optimization problem to seek an optimal transmission policy for the proposed SCM framework, where both the design of the mapping matrix and the power allocation are concerned. To simplify the non-convex problem, we solve the problem with three steps. During the first step, we allocate different angle-domain beams for users to obtain a sparse mapping matrix; during the second step, the prime optimization is transformed as a convex power allocation problem. Finally, we pursue a suboptimal transmission strategy for the multiple clusters with iterative power allocation. Simulation results verify that the SCM scheme exhibits significant performance gain in terms of sum rate.

With the explosive expansion of the information and communication technology (ICT) section, its energy saving has become an important issue and is receiving wide interest. In this study, we propose an adaptive clock frequency strategy for router cards to minimize the total energy consumption of an IP over WDM network. Rather than always running at full speed, the clock frequency of a router card is adaptively adjusted according to its actual-carried traffic demand. Given forecast traffic demand matrixes between different node pairs in different time slots, we develop a mixed integer linear programming (MILP) model to optimally choose the clock frequencies for each router card in different time slots such that the total energy consumption of the router cards is minimized. For lower computational complexity, the optimization model is also decomposed into two models, which correspond to the two subproblems of the optimization problem. The first subproblem minimizes the total number of router cards at each network node based on the peak-hour traffic, and the second subproblem optimally chooses the clock frequencies for each router card in different time slots. Due to the high-computational complexity of the MILP models, we also develop an efficient heuristic algorithm, in which two key steps that tackle the two subproblems are specifically developed. The joint MILP model provides a lower bound on the energy consumption, which shows to save more than 40% energy compared to the case without adaptive router card clock frequency. It is also found that the heuristic algorithm is efficient and performs close to the MILP models. In addition, the results also show that a router card supporting a small number of discrete clock frequencies can perform close to a card with continuously changed clock frequencies, and the benefit of adaptive clock frequency becomes weak with increasing router card power consumption overhead.

DNS rebinding attacks subvert the same-origin policy of browsers, converting them into open network proxies. Using DNS rebinding, an attacker can circumvent organizational and personal firewalls, send spam email, and defraud pay-per-click advertisers. We evaluate the cost effectiveness of mounting DNS rebinding attacks, finding that an attacker requires less than $100 to hijack 100,000 IP addresses. We analyze defenses to DNS rebinding attacks, including improvements to the classic “DNS pinning,” and recommend changes to browser plug-ins, firewalls, and Web servers. Our defenses have been adopted by plug-in vendors and by a number of open-source firewall implementations.

DNS rebinding attacks subvert the same-origin policy of browsers and convert them into open network proxies. We survey new DNS rebinding attacks that exploit the interaction between browsers and their plug-ins, such as Flash and Java. These attacks can be used to circumvent firewalls and are highly cost-effective for sending spam e-mail and defrauding pay-per-click advertisers, requiring less than $100 to temporarily hijack 100,000 IP addresses. We show that the classic defense against these attacks, called "DNS pinning," is ineffective in modern browsers. The primary focus of this work, however, is the design of strong defenses against DNS rebinding attacks that protect modern browsers: we suggest easy-to-deploy patches for plug-ins that prevent large-scale exploitation, provide a defense tool, dnswall, that prevents firewall circumvention, and detail two defense options, policy-based pinning and host name authorization.