Deployment Of Networks Research Articles

Efficient filter pruning: Reducing model complexity through redundancy graph decomposition

Filter pruning has emerged as a crucial technique facilitating the efficient deployment of Convolutional Neural Networks (CNNs) on edge devices. The resultant lightweight models often present a challenging trade-off between performance and compression rate. Moreover, the excessive granularity of filter selection, coupled with intricate fine-tuning procedures, hampers the practical applicability of numerous pruning methodologies. To tackle these limitations, our paper introduces a novel pruning approach that initially constructs an undirected graph, effectively capturing convolutional layer redundancy by quantifying the similarity among output feature maps. Subsequently, our method iteratively eliminates highly redundant and low-value filters, employing both k-core decomposition and importance ranking criteria. In practical implementation, we adopt a one-shot pruning strategy, thereby minimizing the requirement for extensive fine-tuning. Comprehensive comparative experiments demonstrate the efficacy of our proposed approach in achieving remarkable improvements in model compression without compromising performance. For instance, with CIFAR-10, our method achieves state-of-the-art results on DenseNet-40 with a mere 70 fine-tuning epochs, while effectively removing 49.39% of FLOPs, resulting in a mere 0.01% accuracy loss. Furthermore, in the context of ImageNet, our approach excels by removing 48.81% of FLOPs and 23.71% of parameters from ResNet-50, exhibiting only 1.47% and 0.83% decline in Top-1 and Top-5 Accuracy respectively. The code is publicly available at https://github.com/lijiang99/redundancy-graph-decomposition.

Optimizing Virtual Functions Deployment in Multi-UAV IoT Networks

Optimizing Virtual Functions Deployment in Multi-UAV IoT Networks

Optimizing high availability multi-controller placement in SDN/NFV 5G networks: a survey

<p>In meeting the diverse and occasionally conflicting quality of service (QoS) requirements associated with modern communication networks, 5G technology has emerged as a pivotal player. In its architecture, 5G has adopted network function virtualization (NFV) and cloud-based approaches, aiming to simplify network and service deployment, operational processes, and management. The convergence of software defined networking (SDN) and NFV offers an effective solution, enabling scalable and high-performance 5G networks. However, this integration poses critical challenges, with the placement of SDN controllers being a central concern due to its significant impact on network performance, covering aspects such as latency, costs, and energy efficiency. This challenge is known as the controller placement problem (CPP). The central theme of this paper revolves around the intricate relationship between 5G core networks, virtualization technology, and the pressing concern of SDN controller placement, underscoring its significance in the modern networking landscape. We provide a survey of recent methodologies aimed at solving the CPP within the realm of SDN, with a particular focus on resiliency and high availability.</p>

Energy-Aware Microservice-Based Application Deployment in UAV-Based Networks for Rural Scenarios

Yearly, the rates of Internet penetration are on the rise, surpassing 80% in developed nations. Despite this progress, over two billion individuals in rural and low-income regions face a complete absence of Internet access. This lack of connectivity hinders the implementation of vital services like remote healthcare, emergency assistance, distance learning, and personal communications. To bridge this gap and bring essential services to rural populations, this paper leverages Unmanned Aerial Vehicles (UAVs). The proposal introduces a UAV-based network architecture and an energy-efficient algorithm to deploy Internet of Things (IoT) applications. These applications are broken down into microservices, strategically distributed among a subset of UAVs. This approach addresses the limitations associated with running an entire IoT application on a single UAV, which could lead to suboptimal outcomes due to battery and computational constraints. Simulation results conducted in a realistic scenario underscore the effectiveness of the proposed solution. The evaluation includes assessing the percentage of IoT requests successfully served to users in the designated area and reducing the energy consumption required by UAVs during the handling of such requests.

5G Wireless Technology: Spectrum Sharing and Heterogeneous Networks

Abstract: This paper addresses the critical challenge of 5G network cell planning, emphasizing the essential role of estimating the number of cells or base stations required for a given area and user bandwidth. The proposed model explores four scenarios with varying network parameters to determine the optimal base station deployment. The increasing demand for 5G networks necessitates a focus on small cells or femtocells, offering advantages such as consistent coverage, power adjustment, energy efficiency, and higher data rates. However, deploying excessive femtocells may lead to unnecessary handovers, requiring optimization strategies. The study introduces an analytical model for heterogeneous cellular networks, integrating fourth and fifth-generation systems. The model, represented by a two dimensional Markov Chain, employs a novel decomposition approach for analyzing system performance measures. Validation is conducted through simulation and simultaneous equation systems. The proliferation of mobile devices and data traffic necessitates dense 5G network deployments, with small cells gaining traction due to their versatile coverage, power adaptability, energy efficiency, and higher data rates. However, an excessive deployment of small cells can lead to frequent handovers, prompting the need for enhanced handover strategies and coexistence with other technologies in the 5G landscape. The paper concludes by highlighting the impending need for practical implementation testing of 5G systems, emphasizing the significant role of small cell deployments. It discusses an initiative focusing on testing small cellenabled operator business models in real-world scenarios, anticipating the pivotal role of small cells in future 5G systems.

Cost dynamics of converged optical-wireless networks: enabling low-latency xRANs through a reconfigurable hybrid split

Future 6G and beyond wireless networks are anticipated to be highly versatile, accommodating a wide range of services, from ultra-low-latency applications like autonomous vehicles and extended reality to enhanced mobile broadband and massive connectivity for the Internet of Things. In tackling this, xRANs (cloud/virtualized/open radio access networks) encounter significant challenges, including automation, interoperability, scalability, reconfigurability, and standardization, within crosshaul (comprising fronthaul, midhaul, and backhaul) networks. Therefore, the development of programmable converged optical-wireless networks with exceptional flexibility is crucial. This study concentrates on the design of integrated optical and wireless networks to achieve the reconfigurability necessary for automation and to fulfill diverse latency requirements. Initially, we analyze the latency contributions from different network segments and traffic factors in the xRAN, followed by a comprehensive examination of the associated cost dynamics. Subsequently, we investigate the feasibility of integrating high-layer and low-layer splits within the same network to achieve different latency levels. Finally, our study delves into the relationship between latency and cost for converged optical-wireless networks with varying mixed split scenarios and throughput levels. Overall, this article aims to assist network planners in making well-informed decisions that balance throughput performance, cost, and latency requirements in upcoming network deployments.

Optimisation of spatiotemporal context-constrained full-view area coverage deployment in camera sensor networks via quantum annealing

Full-view coverage of a space area with a camera sensor network (CSN) is key to monitoring tasks like security monitoring. Unlike traditional CSN challenges that focus on mere target detection, the full-view area coverage problem (FVACP) demands recognition of targets irrespective of their locations or orientations. However, prior approaches often neglect real-world spatiotemporal context constraints like buildings and pedestrian dynamics, leading to inefficient CSN deployment. Moreover, FVACP’s complexity, being an NP-hard issue, underscores the need for effective optimisation strategies. Recently, quantum annealing (QA) has emerged as a promising solution, which potentially outperforms classical computing in optimisation tasks. Therefore, this study proposes a QA-based FVACP optimisation framework. It addresses spatial constraints by optimising candidate deployment points and tackles temporal constraints by optimising sensor orientations. These optimisation tasks are converted into quadratic unconstrained binary optimisation problems, which are suitable for QA techniques and benchmarking against classical methods. The effectiveness of the framework is validated through facial recognition-oriented experiments. Results demonstrate not only efficient CSN deployment with larger benefits and fewer cameras but also confirm the superiority of QA over classical computing given that it delivers approximate optimum outcomes across various scenarios. Consequently, CSN monitoring capabilities in real-world applications can be enhanced.

Information Dimension Matching in Memristive Computing System for Analog Deployment of Deep Neural Networks

AbstractMemristor, with the ability of analog computing, is widely investigated for improving the computing efficiency of deep neural networks (DNNs) deployment. However, how to fully take advantage of the analog computing ability of memristive computing system (MCS) for DNN deployment is still an open question. Here, a new neural network models deployment scheme, that is, an information dimension matching (IDM) scheme, is proposed to fully take advantage of the analog computing ability of MCS. Furthermore, the spatial and temporal DNN, that is convolutional neural network (CNN) and recurrent neural network (RNN) is used to verify the proposed deployment scheme, respectively. The experimental results indicate that, compared to the traditional deployment schemes, the proposed deployment scheme shows obvious inference accuracy and energy efficiency improvement (>4 × in four‐layer DNNs deployment), and the energy efficiency improvement increases dramatically with the layers increment of DNNs. This work paves the path for developing high computing efficiency analog MCS.

Fog-based deep learning framework for real-time pandemic screening in smart cities from multi-site tomographies

The quick proliferation of pandemic diseases has been imposing many concerns on the international health infrastructure. To combat pandemic diseases in smart cities, Artificial Intelligence of Things (AIoT) technology, based on the integration of artificial intelligence (AI) with the Internet of Things (IoT), is commonly used to promote efficient control and diagnosis during the outbreak, thereby minimizing possible losses. However, the presence of multi-source institutional data remains one of the major challenges hindering the practical usage of AIoT solutions for pandemic disease diagnosis. This paper presents a novel framework that utilizes multi-site data fusion to boost the accurateness of pandemic disease diagnosis. In particular, we focus on a case study of COVID-19 lesion segmentation, a crucial task for understanding disease progression and optimizing treatment strategies. In this study, we propose a novel multi-decoder segmentation network for efficient segmentation of infections from cross-domain CT scans in smart cities. The multi-decoder segmentation network leverages data from heterogeneous domains and utilizes strong learning representations to accurately segment infections. Performance evaluation of the multi-decoder segmentation network was conducted on three publicly accessible datasets, demonstrating robust results with an average dice score of 89.9% and an average surface dice of 86.87%. To address scalability and latency issues associated with centralized cloud systems, fog computing (FC) emerges as a viable solution. FC brings resources closer to the operator, offering low latency and energy-efficient data management and processing. In this context, we propose a unique FC technique called PANDFOG to deploy the multi-decoder segmentation network on edge nodes for practical and clinical applications of automated COVID-19 pneumonia analysis. The results of this study highlight the efficacy of the multi-decoder segmentation network in accurately segmenting infections from cross-domain CT scans. Moreover, the proposed PANDFOG system demonstrates the practical deployment of the multi-decoder segmentation network on edge nodes, providing real-time access to COVID-19 segmentation findings for improved patient monitoring and clinical decision-making.

A CNN-Based Framework for Video Analysis and Accident Detection

This research investigates the development and deployment of a Convolutional Neural Network (CNN) model for automatic accident detection in CCTV footage. The ever-increasing reliance on video surveillance necessitates efficient and accurate methods for accident identification. CNNs, with their inherent ability to learn complex spatial relationships within images, are particularly well-suited for this task. This study proposes a CNN architecture that utilizes a pre-trained MobileNetV2 base for feature extraction, followed by a custom classification head tailored to the specific task of accident vs. no accident classification. The model is trained on a dataset of grayscale video frames, achieving an impressive accuracy of 92% on the testing set. This high level of accuracy suggests that CNNs hold significant promise for real-world accident detection applications. Furthermore, to bridge the gap between research and practical implementation, the model is converted to a TensorFlow Lite (TFLite) format for deployment on resource-constrained devices. Additionally, a user-friendly frontend application is developed, empowering users to interact with the model and analyze both images and videos. This user-centric approach broadens the model's accessibility and paves the way for potential improvements in road safety through real-time accident detection.

ML Assisted Foot Step Power Generation using Piezoelectric Sensors

In our rapidly evolving world, the escalating demand for energy coupled with the finite nature of traditional resources underscores the imperative for sustainable, pollution-free, and inexhaustible energy solutions. This paper introduces a pioneering approach to harnessing the kinetic energy expended during human locomotion through the innovative use of piezoelectric sensors. Leveraging the piezoelectric effect, these sensors efficiently convert mechanical energy generated by footstep pressure into electrical energy, thereby mitigating wastage and addressing the increasing energy needs. Our model advocates for the deployment of an extensive sensor network along footpaths, complemented by an RFID-based mobile charging system for enhanced convenience and functionality. Moreover, we introduce an innovative method integrating Machine Learning (ML) techniques to enhance power generation efficiency through intelligent modulation of piezoelectric element resistance. Additionally, we leverage ML algorithms to enhance the requisite daily footstep count necessary to fulfill the energy demands of specific areas. This proactive approach ensures optimal deployment of footstep power generation devices based on actual foot traffic patterns. Our research underscores the significance of this technology in the context of urban energy sustainability, particularly in densely populated regions like China and India, where foot traffic is abundant. By harnessing mechanical energy and leveraging advanced ML algorithms, our approach promises to revolutionize energy harvesting paradigms, paving the way for greener and more efficient power generation systems

Динамическое использование спектра в LTE и NR для развертывания 5G

Both Static allocation of the frequency spectrum as deployed in 4G-LTE as well as dynamically configured 5G NR physical layer channels does not satisfy the demand of current exponential growth of wireless users. As 5G NR will perform better in Line of Sight (LOS) scenarios while 4G-LTE provides remarkable performance in non-Line of sight (NLOS) scenarios. Dynamic Spectrum Sharing (DSS) allows operators to run LTE and 5G NR on the same band simultaneously. This paper illustrates the relationship and working theory of Dynamic Spectrum Sharing to comprehend how the latest 5G standard will result in a substantial improvement in average user throughput and improved network efficiencies. This is an effort to propose a generic framework to initial Deployment of 5G network.

An interpretable deep learning model predicts RNA–small molecule binding sites

The intricate interplay between RNA molecules and small ligands plays a pivotal role in regulating biological processes, underscoring the necessity for accurate prediction models of RNA–small molecule binding sites. In this study, we develop CapBind, a novel interpretive deep learning framework utilizing a Capsule Network architecture, to precisely identify RNA–small molecule binding sites. This framework stands out for its multifaceted approach, which leverages both the sequence and structural information of RNA as input features, allowing for a thorough understanding of the complex interactions between RNA and small molecules. Leveraging multi-source bioinformatics data, including the distribution of RNA secondary structures, our model characterizes sequence and structural features with unparalleled depth. Through a unified alignment of these multi-source feature representations, we achieve dimensional homogeneity in the potential feature maps for both types of biological data. The crux of our methodology lies in the deployment of a multi-level Capsule Network, engineered to identify the nuanced interactions of RNA sequences with small molecules. The essence of our approach resides in the implementation of a multi-layered Capsule Network. Utilizing a sophisticated routing mechanism, this network demonstrates unparalleled efficacy in learning and identifying the intricate interactions between RNA sequences and small molecules. This mechanism enables dynamic adjustment of information flow across multiple levels, thereby capturing subtle features and patterns that might elude conventional neural network architectures. The experimental results demonstrate the superior performance of our model, particularly highlighting its progress in evaluation metrics compared to current state-of-the-art models. Case studies further validate the model’s precision in identifying binding sites, offering potential insights into gene expression pathways and novel therapeutic avenues for neurodegenerative diseases. Moreover, an extensive interpretability analysis of our silicon model reveals the internal interaction mechanisms between specific RNA regions and small molecules, contributing valuable perspectives for drug design and molecular biology research. These findings underscore the model’s predictive accuracy and biological applicability, establishing our approach as a significant leap forward in the computational prediction of RNA–small molecule binding sites.

A Lightweight Secure Scheme for Underwater Wireless Acoustic Network

Due to the open underwater channels and untransparent network deployment environments, underwater acoustic networks (UANs) are more vulnerable to hostile environments. Security research is also being conducted in cryptography, including authentication based on asymmetric algorithms and key distribution based on symmetric algorithms. In recent years, the advancement of quantum computing has made anti-quantum attacks an important issue in the field of security. Algorithms such as lattice and SPHINCS+ have become a research topic of interest in the field of security. However, within the past five years, few papers have discussed security algorithms for UANs to resist quantum attacks, especially through classical algorithms. Some existing classical asymmetric and symmetric algorithms are considered to have no prospects. From the perspective of easy deployment in engineering and anti-quantum attacks, our research focuses on a comprehensive lightweight security framework for data protection, authentication, and malicious node detection through the Elliptic Curve and Hash algorithms. Our mechanism is suitable for ad hoc scenarios with limited underwater resources. Meanwhile, we have designed a multi-party bit commitment to build a security framework for the system. A management scheme is designed by combining self-certifying with the threshold sharing algorithm. All schemes are designed based on certificate-less and ad hoc features. The proposed scheme ensures that the confidentiality, integrity, and authentication of the system are well considered. Moreover, the scheme is proven to be of unconditional security and immune to channel eavesdropping. The resource and delay issues are also taken into consideration. The simulations considered multiple variables like number of nodes, attackers, and message length to calculate proper values that can increase the efficiency of this scheme. The results in terms of delay, delivery ratio, and consumption demonstrate the suitability of the proposal in terms of security, especially for malicious node detection. Meanwhile, the computational cost has also been controlled at the millisecond level.

Ultra-Low Loss Fiber Deployment in Elastic Optical Networks With Fixed and Variable Topologies

Ultra-Low Loss Fiber Deployment in Elastic Optical Networks With Fixed and Variable Topologies

Channel modeling for NLoS visible light networks with integrated sensing and communication.

Inspired by the advanced integrated sensing and communication (ISAC), in this Letter, we explore the non-line-of-sight (NLoS) optical channels formed by reflections from the ground or objects to establish an integrated channel model for simultaneous communication and sensing. The integrated channel model can, on the one hand, perceive the changes in the surrounding environment and, on the other hand, determine whether these changes positively or negatively affect the quality of communication simultaneously. To validate the effectiveness of the proposed model, from sensing, we analyze the impact of various floor materials and visible light communication (VLC) users on the integrated channel; from communication, we characterize the influence of perceived environmental changes on communication performance by calculating throughput. Experimental results confirm the capability of the derived model, which can support the design and deployment of VL-based ISAC networks.

An invisible, robust copyright protection method for DNN-generated content

Wide deployment of deep neural networks (DNNs) based applications (e.g., style transfer, cartoonish), stimulating the need for copyright protection of such application’s production. Though some traditional visible copyright techniques exist, they often introduce undesired artifacts and compromise the aesthetic quality of the images. In this paper, we propose a novel invisible, robust copyright protection method, which is composed of two networks: the copyright encoder and the copyright decoder. The former projects the copyright information to the invisible perturbation with the drive of both the input of images and copyright information, thereby adding it to the image and yielding encoded images. The copyright decoder extracts copyright information from encoded images. Moreover, a robustness module is integrated to enhance the decoder’s ability to decipher images against various distortions encountered on social media platforms. Furthermore, the loss function is elaborately designed, taking into account both feature space and color space, to guarantee the quality of encoded and decoded copyright images. Extensively objective and subjective experiments validate the effectiveness of the proposed method. Additionally, the physical test is conducted by posting the encoded images to social media (e.g., Weibo and Twitter) and downloading them to verify the feasibility of the proposed method in practice.

A Regeneration Model for Mitigation Against Attacks on HTTP Servers for Mobile Wireless Networks

The widespread expansion of the internet has fueled a global surge in the utilization of various online transactions, with a significant portion of such services operating on mobile web platforms. Simultaneously, the deployment of innovative Mobile Ad Hoc Network (MANET) technologies and mobile applications has grown as solutions for diverse tasks. Unfortunately, this progress has attracted the attention of hackers, who continually devise new strategies to exploit the vulnerabilities inherent in mobile networks. This study aims to address the escalating challenges posed by cyber threats in the era of widespread internet expansion, particularly focusing on securing mobile web platforms against sophisticated attacks such as Structured Query Language Injection (SQLi) for web-based database solutions and Denial of Service (DoS/DDoS) for various applications. In response to the identified vulnerabilities, this paper proposes an HTTP regeneration (HReg) model that not only detects various cyber-attacks but also ensures the uninterrupted provision of critical services during such incidents. The model introduces an innovative regeneration algorithm capable of scanning both the connection channel and web application to detect attacks, creating survivable connections within the underlying TCP engine to replace compromised ones during an ongoing attack. In simulated environments using OMNeT++, where the server is subjected to attacks, the experimental results demonstrate the efficacy of the model. The response and performance metrics, including throughput (73%), delivery ratio (68.8%), delay (3s), and network load, showcase the model's ability to detect and neutralize attacks. A comparison with state-of-the-art approaches highlights the superior performance of the regeneration model, attributed to its additional survivability layer. While the regeneration model proves robust in simulated environments, its real-world application may encounter limitations. Future research should explore these limitations to enhance the practical applicability of the proposed model. The proposed HReg model's resilient performance under attack conditions ensures the survivability of web-based applications. This innovative approach offers practical implications for securing mobile web platforms, providing continuous delivery of critical services even in the face of persistent and evolving cyber threats. This research addresses a significant gap in existing efforts by not only focusing on attack detection but also emphasizing the development of a survivable TCP connection for HTTP servers during attacks. The introduced regeneration model stands out for its unique ability to maintain service continuity, showcasing originality in the approach to cybersecurity in the context of web-based applications.

National Emergency Tele-Critical Care in a Pandemic: Barriers and Solutions.

The COVID-19 pandemic caused tremendous disruption to the U.S. healthcare system and nearly crippled some hospitals during large patient surges. Limited ICU beds across the country further exacerbated these challenges. Telemedicine, specifically tele-critical care (TCC), can expand a hospital's clinical capabilities through remote expertise and increase capacity by offloading some monitoring to remote teams. Unfortunately, the rapid deployment of telemedicine, especially TCC, is constrained by multiple barriers. In the summer of 2020, to support the National Emergency Tele-Critical Care Network (NETCCN) deployment, more than 50 national leaders in applying telemedicine technologies to critical care assembled to provide their opinions about barriers to NETCCN implementation and strategies to overcome them. Through consensus, these experts developed white papers that formed the basis of this article. Herein, the authors share their experience and propose multiple solutions to barriers presented by laws, local policies and cultures, and individual perspectives according to a minimum, better, best paradigm for TCC delivery in the setting of a national disaster. Cross-state licensure and local privileging of virtual experts were identified as the most significant barriers to rapid deployment of services, whereas refining the model of TCC to achieve the best outcomes and defining the best financial model is the most significant for long-term success. Ultimately, we conclude that a rapidly deployable national telemedicine response system is achievable.

Reg-Tune: A Regression-Focused Fine-Tuning Approach for Profiling Low Energy Consumption and Latency

Fine-tuning deep neural networks is pivotal for creating inference modules that can be suitably imported to edge or field-programmable gate array (FPGA) platforms. Traditionally, exploration of different parameters throughout the layers of deep neural networks has been done using grid search and other brute force techniques. Although these methods lead to the optimal choice of network parameters, the search process can be very time consuming and may not consider deployment constraints across different target platforms. This work addresses this problem by proposing Reg-Tune, a regression-based profiling approach to quickly determine the trend of different metrics in relation to hardware deployment of neural networks on tinyML platforms like FPGAs and edge devices. We start by training a handful of configurations belonging to different combinations of \(\mathcal {NN}\scriptstyle \langle q (quantization),\,s (scaling)\rangle \displaystyle\) or \(\mathcal {NN}\scriptstyle \langle r (resolution),\,s\rangle \displaystyle\) workloads to generate the accuracy values respectively for their corresponding application. Next, we deploy these configurations on the target device to generate energy/latency values. According to our hypothesis, the most energy-efficient configuration suitable for deployment on the target device is a function of the variables q , r , and s . Finally, these trained and deployed configurations and their related results are used as data points for polynomial regression with the variables q , r , and s to realize the trend for accuracy/energy/latency on the target device. Our setup allows us to choose the near-optimal energy-consuming or latency-driven configuration for the desired accuracy from the contour profiles of energy/latency across different tinyML device platforms. To this extent, we demonstrate the profiling process for three different case studies and across two platforms for energy and latency fine-tuning. Our approach results in at least 5.7 \(\times\) better energy efficiency when compared to recent implementations for human activity recognition on FPGA and 74.6% reduction in latency for semantic segmentation of aerial imagery on edge devices compared to baseline deployments.