Deployment Of Applications Research Articles

Optimizing traffic management for public services during high-demand periods using cloud load balancers

In times of high demand, effective traffic management is essential for public service platforms to maintain reliable, continuous access for users. This paper explores the role of cloud load balancers as a solution for optimizing traffic distribution in public services during peak periods. By dynamically distributing network traffic across multiple servers, cloud load balancers enhance system scalability, flexibility, and response time, which are critical for minimizing service disruptions. The paper further discusses best practices for configuring load balancers, including geographic balancing and redundancy, and highlights the importance of integrating dynamic scaling and monitoring tools to adapt to sudden traffic surges. Additionally, the potential of artificial intelligence (AI) and predictive analytics in advancing cloud load balancing is considered, emphasizing their value for predictive scaling and automated traffic management. Key challenges, including data privacy and dependency on third-party infrastructure, are also addressed, with recommendations for enhancing load balancer deployment in public service applications. This paper concludes with strategic recommendations for public service providers to leverage cloud technology effectively, ensuring improved reliability, security, and cost-efficiency in service delivery. Keywords: Cloud Load Balancers, Traffic Management, Public Service Platforms, Dynamic Scaling, Predictive Analytics, Service Reliability.

Demystifying Serverless Architecture for Scalable Web Applications

Serverless computing has revolutionized the development and deployment of scalable web applications by abstracting infrastructure management and enabling automatic scaling. This article aims to demystify serverless architecture by thoroughly analyzing its core concepts, benefits, and practical implementation strategies. We differentiate between Function-as-a-Service (FaaS) and Backend-as-a-Service (BaaS) models and explore how they contribute to scalability and cost efficiency. Deployment best practices are outlined using AWS Lambda, Azure Functions, and Google Cloud Functions. This article addresses common challenges associated with serverless computing, including cold starts and execution time limitations, and propose solutions to mitigate these issues. Integration with DevOps practices is also discussed, demonstrating how serverless applications can be incorporated into continuous integration and continuous deployment (CI/CD) pipelines. Through case studies, this article illustrates the effective utilization of serverless architecture in building scalable and resilient web applications.

The Role of Technophilia and User Goals in the Intention to Use a Mobility Management Travel App

The ubiquitous use of mobile devices along with the amount of traffic, transportation services, and travel pattern data available has led to the emergence and deployment of smartphone applications for providing information about personal travel management. Several of these travel apps are aimed at voluntary travel behavior change (VTBC) to support and increase sustainable mobility, and have led to the development of research to investigate their influence on travel behavior. Here, the aim is to study the role of technophilia and goal-framing theory in the intention to adopt and situationally use a prospective VTBC travel app. A Structural Equation Model is developed with the aim of empirically testing a sample of 971 respondents collected in two suburban corridors in the Lisbon Metropolitan Area. The results support that goal-framing theory is important for explaining the adoption of VTBC travel apps. Gain and normative motives are more relevant than hedonic motives, pointing to the importance of their tangible benefits. Frequent car users may benefit from VTBC travel apps in terms of encouraging behavioral changes, supporting sustainable mobility management solutions. The results also outline the importance of technophilia and the current use of travel apps in influencing the intention to use VTBC apps.

Path Algebra-Driven Classification Solution to Realize User-Centric Performance-Oriented Virtual Network Embeddings

The intense diversity of the Next-Generation Networking environments like 6G and the forthcoming deployment of immersive applications with varied user-specific requirements transform the efficient allocation of resources into a real challenge. Traditional solutions like the shortest path algorithm and mono-constraint methodologies are inadequate to handle customized user-defined performance parameters and effectively classify physical resources according to these intricate demands. This research offers a new evaluation mechanism to successfully replace the aforementioned traditional path ranking and path selection techniques. Specifically, the proposed framework is integrated with optimization-oriented metrics, each indicating a unique aspect of performance for evaluating candidate network paths. The deployed metrics are then algebraically synthesized to provide a distinctive multidimensional description of the examined substrate resources. These primary and composite metrics adhere to the fundamental monotonicity and isotonicity properties of a Path Algebra; hence, the validity and optimality of the proposed evaluation mechanism is guaranteed by design. To tackle the complexity created by the variety of human-centric customization, a novel methodology that analyzes and determines the weighted influence of the synthesized metrics depending on the characteristics of the served user-centric application is also introduced. The chosen suitable weights address performance-oriented mission-critical tailored objectives for adaptive optimizations. Its innovative algebraic design allows it to successfully describe and rank candidate paths in a versatile way, whether in legacy or modern architectures. The experimental data of the first scenario show that 62.5% and 50% of highlighted path evaluations proposed by the shortest path and unidimensional constraint strategies, respectively, suffer from moderate performance-oriented values compared to the proposed framework. Likewise, the results of the second examined scenario reveal that the proposed composite metric yields more suitable path rankings by 50% in contrast to its traditional counterparts, rendering the contested evaluation mechanisms obsolete.

Multiobjective Hybrid Al-Biruni Earth Namib Beetle Optimization and Deep Learning based Task Scheduling in Cloud Computing

With the rapid development of computing networks, cloud computing (CC) enables the deployment of large-scale applications and meets the increased rate of computational demands. Moreover, task scheduling is an essential process in CC. The tasks must be effectually scheduled across the Virtual Machines (VMs) to increase resource usage and diminish the makespan. In this paper, the multi-objective optimization called Al-Biruni Earth Namib Beetle Optimization (BENBO) with the Bidirectional-Long Short-Term Memory (Bi-LSTM) named as BENBO+ Bi-LSTM is developed for Task scheduling. The user task is subjected to the multi-objective BENBO, in which parameters like makespan, computational cost, reliability, and predicted energy are used to schedule the task. Simultaneously, the user task is fed to Bi-LSTM-based task scheduling, in which the VM parameters like average computation cost, Earliest Starting Time (EST), task priority, and Earliest Finishing Time (EFT) as well as the task parameters like bandwidth and memory capacity are utilized to schedule the task. Moreover, the task scheduling outcomes from the multi-objective BENBO and Bi-LSTM are fused for obtaining the final scheduling with less makespan and resource usage. Moreover, the predicted energy, resource utilization and makespan are considered to validate the BENBO+ Bi-LSTM-based task scheduling, which offered the optimal values of 0.669J, 0.535 and 0.381.

Automated Diagnosis of Knee Osteoarthritis Using ResNet101 on a DEEP:PHI: Leveraging a No-Code AI Platform for Efficient and Accurate Medical Image Analysis.

Background: Knee osteoarthritis (OA) is a prevalent degenerative joint disease significantly impacting global health. Early and accurate diagnosis is crucial for effective management, but traditional methods often rely on subjective assessments. This study evaluates the efficacy of a deep learning model implemented through a no-code AI platform for diagnosing and grading knee OA from plain radiographs. Methods: We utilized the Osteoarthritis Initiative (OAI) dataset, comprising knee X-ray data from 1526 patients. The data were split into training (47.0%), validation (26.5%), and test (26.5%) sets. We employed a ResNet101 model on the DEEP:PHI no-code AI platform for image analysis. The model was trained to classify knee OA into five grades (0-4) based on the Kellgren-Lawrence scale. Results: Our AI model demonstrated high accuracy in distinguishing between different OA grades, with particular strength in early-stage detection. The model achieved optimal performance at 20 epochs, suggesting efficient learning dynamics. Grad-CAM visualizations were used to enhance the interpretability of the model's decision-making process. Conclusions: This study demonstrates the potential of AI, implemented through a no-code platform, to accurately diagnose and grade knee OA from radiographs. The use of a no-code AI platform such as DEEP:PHI represents a step towards democratizing AI in healthcare, enabling the rapid development and deployment of sophisticated medical AI applications without extensive coding expertise. This approach could significantly enhance the early detection and management of knee OA, potentially improving patient outcomes and streamlining clinical workflows.

A review of AI edge devices and lightweight CNN and LLM deployment

Artificial Intelligence of Things (AIoT) which integrates artificial intelligence (AI) and the Internet of Things (IoT), has attracted increasing attention recently. With the remarkable development of AI, convolutional neural networks (CNN) have achieved great success from research to deployment in many applications. However, deploying complex and state-of-the-art (SOTA) AI models on edge applications is increasingly a big challenge. This paper investigates literature that deploys lightweight CNNs on AI edge devices in practice. We provide a comprehensive analysis of them and many practical suggestions for researchers: how to obtain/design lightweight CNNs, select suitable AI edge devices, and compress and deploy them in practice. Finally, future trends and opportunities are presented, including the deployment of large language models, trustworthy AI and robust deployment.

Cephalometric landmark annotation using transfer learning: Detectron2 and YOLOv8 baselines on a diverse cephalometric image dataset

BackgroundCephalometric landmark annotation is a key challenge in radiographic analysis, requiring automation due to its time-consuming process and inherent subjectivity. This study investigates the application of advanced transfer learning techniques to enhance the accuracy of anatomical landmarks in cephalometric images, which is a vital aspect of orthodontic diagnosis and treatment planning. MethodsWe assess the suitability of transfer learning methods by employing state-of-the-art pose estimation models. The first framework is Detectron2, with two baselines featuring different ResNet backbone architectures: rcnn_R_50_FPN_3x and rcnn_R_101_FPN_3x. The second framework is YOLOv8, with three variants reflecting different network sizes: YOLOv8s-pose, YOLOv8m-pose, and YOLOv8l-pose. These pose estimation models are adopted for the landmark annotation task. The models are trained and evaluated on the DiverseCEPH19 dataset, comprising 1692 radiographic images with 19 landmarks, and their performance is analyzed across various images categories within the dataset. Additionally, the study is extended to a benchmark dataset of 400 images to investigate how dataset size impacts the performance of these frameworks. ResultsDespite variations in objectives and evaluation metrics between pose estimation and landmark localization tasks, the results are promising. Detectron2's variant outperforms others with an accuracy of 85.89%, compared to 72.92% achieved by YOLOv8's variant on the DiverseCEPH19 dataset. This superior performance is also observed in the smaller benchmark dataset, where Detectron2 consistently maintains higher accuracy than YOLOv8. ConclusionThe noted enhancements in annotation precision suggest the suitability of Detectron2 for deployment in applications that require high precision while taking into account factors such as model size, inference time, and resource utilization, the evidence favors YOLOv8 baselines.

Design and Implementation of Virtualization Cloud Computing System Intelligent Terminal Application Layer

Cloud task scheduling has become a trend, and the shortcomings of traditional scheduling algorithms can be optimized through mathematical models of other cloud task scheduling scenarios. In order to improve the virtualization data processing effect of intelligent terminal application layer, this paper proposes an improved krill swarm optimization algorithm based on adaptive weight. The optimization of cluster load balancing and task average response time ratio are used to improve the convergence and accuracy of task scheduling algorithm. Moreover, this paper uses CloudSim simulation tool to conduct experiments to verify the effectiveness of the proposed model. In addition, this paper proposes an application-based virtualization method, which virtualizes the application programs inside the host machine into the virtualization software inside the virtual machine, so that the virtual machine can access it. Finally, this paper verifies the reliability of the proposed method with experiments, thus providing a theoretical reference for the subsequent design of intelligent terminal application layer virtualization cloud computing system. Compared with the traditional way of using physical hardware, using virtual machine hardware is more flexible, efficient and safe, which brings great convenience to the development and deployment of applications.

Bi-directional SMF-NDF-optical/5G NR wireless converged systems

This study presents the transmission of fifth-generation (5G) new radio (NR) signals over bi-directional single-mode fiber (SMF)-negative dispersion fiber (NDF)-optical/5G NR wireless converged systems, using two cascaded reflective semiconductor optical amplifiers (RSOAs). Downstream transmission of 5G NR 16-quadrature amplitude modulation (QAM) signals achieved an aggregate net bit rate of 228.037 Gb/s, comprising 59.813, 74.766, and 93.458 Gb/s at 150, 250, and 325 GHz, respectively. The system achieved a low bit error rate (BER) and clear constellation patterns. However, systems without NDF exceed the forward error correction 7% threshold, indicating the importance of NDF in compensating for fiber dispersion. Moreover, for upstream transmission, two cascaded RSOAs are used to transmit 5G 16-QAM-orthogonal frequency division multiplexing signals, achieving a total data rate of 40 Gb/s. The use of two cascaded RSOAs in the system significantly improved upstream performance, resulting in low BERs and clear constellation patterns. This system not only achieved high data rates but also extended transmission coverage to support the deployment of advanced 5G NR applications.

Deployment of mobile application using a novel CNN model for the detection of COVID-19 thoracic disease

In January 2021 and January 2022, COVID-19 caused roughly 13,000 and 6,000 deaths respectively per day. In August 2022, 26,000 deaths per day were estimated to be caused by COVID-19, followed by 13,000 deaths per day in February 2024. The timely identification and treatment of malignant diseases can potentially lower the mortality rate. Nonetheless, the use of manual methods for diagnosing these conditions requires a meticulous and comprehensive examination, making it susceptible to errors, burdensome for healthcare professionals, and time-intensive. Hence, the objective of this study is to design and deploy a novel deep-learning model for the detection of COVID-19 thoracic diseases. A Convolutional Neural Network (CNN) with less trainable parameters was implemented. This proposed model was deployed on a mobile device using Android Studio and Flutter for the detection of COVID-19 thoracic diseases. Specificity, accuracy, precision, sensitivity, f1-score, ROC, and PR curves were used to evaluate the model's performance. Moreover, the carbon footprint as well as how responsible the proposed model is according to Responsible AI rules was also assessed. The model's evaluation results show an overall accuracy of 93.27 %, specificity of 97.33 %, precision of 93.75 %, sensitivity of 94.42 %, F1-Score of 94.06 %, ROC rate of 98.0 %, PR rate of 96.8 %. The evaluation of the mobile application shows higher generalizability on the COVID-19 dataset. Also, the overall FACETS Score representing responsible AI is 83 % and the carbon footprint (representing the amount of carbon emission emitted into the environment during model training and testing) of 416.73 g with equivalent tree months of 0.45 was obtained. This application with better performance and a low carbon footprint was deployed using Android Studio and Flutter and can assist physicians in the diagnosis of COVID-19 and related diseases.

Nucleo-SAFARI: Automated Identification of Fragment Ions in Top-Down MS/MS Spectra of Nucleic Acids.

Recent progress in top-down mass spectrometry analysis of progressively larger nucleic acids has enabled in-depth characterization of intact, modified RNA molecules. Development of methods for desalting and MS/MS fragmentation allows rapid acquisition of high-quality top-down MS/MS spectra of nucleic acids up to 100 nt, which has spurred the need for development of software approaches to identify and validate nucleic acid fragment ions. We have implemented an R-based approach to aid in analysis of MS/MS spectra of nucleic acids based on fragment ions observed directly in the m/z domain. This program, entitled Shiny Application for Fragment Assignment by Relative Isotopes (Nucleo-SAFARI), utilizes the Shiny HTML framework for deployment of a user-friendly application for automated annotation of top-down MS/MS spectra of nucleic acids recorded on Orbitrap mass spectrometer platforms. This approach proceeds through in silico generation of fragment ions and their isotopic distributions, followed by algorithmic assessment of the experimental isotopic distributions. Nucleo-SAFARI is available for download at https://github.com/mblanzillotti/Nucleo-SAFARI.

Submerged and completely open solid–liquid triboelectric nanogenerator for water wave energy harvesting

AbstractTriboelectric nanogenerator (TENG) is an emerging wave energy harvesting technology with excellent potential. However, due to issues with sealing, anchoring, and difficult deployment over large areas, TENG still cannot achieve large‐scale wave energy capture. Here, a submerged and completely open solid–liquid TENG (SOSL‐TENG) is developed for ocean wave energy harvesting. The SOSL‐TENG is adapted to various water environments. Due to its simple structure, it is easy to deploy into various marine engineering facilities in service. Importantly, this not only solves the problem of difficult construction of TENG networks at present, but also effectively utilizes high‐quality wave energy resources. The working mechanism and output performance of the SOSL‐TENG are systematically investigated. With optimal triggering conditions, the transferred charge (Qtr) and short‐circuit current (Isc) of SOSL‐TENG are 2.58 μC and 85.9 μA, respectively. The wave tank experiment is taken for fully demonstrating the superiority of the SOSL‐TENG network in large‐scale collection and conversion of wave energy. Due to the excellent output performance, TENG can harvest wave energy to provide power for various commercial electronic devices such as LED beads, hygrothermograph, and warning lights. Importantly, the SOSL‐TENG networks realizes self‐powered for electrochemical systems, which provides a direction for energy cleanliness in industrial systems. This work provides a prospective strategy for large‐scale deployment of TENG applications, especially for harvesting wave energy in spray splash zones or at the surface of the water.image

Secured Cyber-Internet Security in Intrusion Detection with Machine Learning Techniques

The rapid proliferation of Internet-connected devices has elevated the significance of cybersecurity, making intrusion detection a critical aspect of maintaining network integrity. Traditional security measures often fail to provide adequate protection against sophisticated attacks, necessitating advanced and robust solutions. This paper introduces a comprehensive cyber-internet security framework that leverages machine learning techniques for real-time intrusion detection and prevention. The proposed methodology employs a hybrid approach, integrating supervised and unsupervised learning models to detect anomalies and classify intrusions effectively. Specifically, a combination of Support Vector Machine (SVM), Decision Trees (DT), and K-means clustering is used to enhance detection accuracy and reduce false-positive rates. The experimental results demonstrate that the proposed model achieved a detection accuracy of 97.8%, a precision of 96.5%, and a recall of 95.2% on the NSL-KDD dataset. The implementation also reduced the false-positive rate to 1.2% and the computational overhead by 15% compared to traditional detection systems. Additionally, the proposed system was tested on real-time traffic data, where it successfully identified and mitigated various cyber threats, including Distributed Denial of Service (DDoS) attacks and network infiltrations, with minimal latency and high reliability. In conclusion, the study presents an efficient and secured cyber-internet security framework that significantly enhances intrusion detection capabilities using machine learning techniques. The proposed system provides a scalable and adaptive solution for securing critical infrastructure and networks against evolving cyber threats, making it an ideal candidate for deployment in real-world cybersecurity applications.

YOLO-LFPD: A Lightweight Method for Strip Surface Defect Detection.

Strip steel surface defect recognition research has important research significance in industrial production. Aiming at the problems of defect feature extraction, slow detection speed, and insufficient datasets, YOLOv5 is improved on the basis of YOLOv5, and the YOLO-LFPD (lightweight fine particle detection) model is proposed. By introducing the RepVGG (Re-param VGG) module, the robustness of the model is enhanced, and the expressive ability of the model is improved. FasterNet is used to replace the backbone network, which ensures accuracy and accelerates the inference speed, making the model more suitable for real-time monitoring. The use of pruning, a GA genetic algorithm with OTA loss function, further reduces the model size while better learning the strip steel defect feature information, thus improving the generalisation ability and accuracy of the model. The experimental results show that the introduction of the RepVGG module and the use of FasterNet can well improve the model performance, with a reduction of 48% in the number of parameters, a reduction of 13% in the number of GFLOPs, an inference time of 77% of the original, and an optimal accuracy compared with the network models in recent years. The experimental results on the NEU-DET dataset show that the accuracy of YOLO-LFPD is improved by 3% to 81.2%, which is better than other models, and provides new ideas and references for the lightweight strip steel surface defect detection scenarios and application deployment.

Development of a new soil–structure contact stress sensor for underground construction applications

This paper describes the design, development, calibration and validation of a novel soil–structure contact stress sensor. The new sensor design combines a novel operating principle, fibre Bragg grating (FBG) strain sensing and data-driven mapping techniques to create a multi-axis contact stress sensor that is both economical and suitably robust for deployment in underground construction applications. The instrumentation process is informed using a ‘virtual twin’ of the sensor in which synthetic data is generated by extracting and interpolating virtual FBG strains obtained from a large number of 3D finite-element calculations. A physical prototype is subsequently developed to demonstrate proof of concept. Results from laboratory validation tests give confidence in the sensor's ability to provide accurate contact stress measurements in typical soil–structure interface shear applications. In particular, the novel sensor structure and operating principle was shown to achieve excellent measurement of effective normal stress. The new sensor design harnesses many of the inherent benefits of FBGs including immunity to electromagnetic noise and water ingress, and the use a single lightweight cable and connector, which significantly simplifies installation on site compared to electrical multi-axis sensors.

FPGA Implementation of Compact Architecture for Lightweight Hash Algorithm for Resource Constrained Devices

The security implications of the emerging world of smart devices brought about by the Internet of things' recent rise are enormous. IoT devices typically have restricted resources, such as minimal memory, low processing power, and a short battery life, in addition to their stated security level. In the context of IoT, lightweight cryptographic primitives are presented taking into account the trade-off between speed and security assurance. In this research, we proposed lightweight cryptographic hash methods and optimized hardware for devices with limitations. The size, speed, efficiency, and power consumption of our suggested solutions are examined on FPGA systems. Bit permutation, linear transformation, and S-Box functionality are used in the suggested design. To analyze large data sets and Internet of Things applications, conventional hash algorithms need memory and time. Thus, the demand arises for a lightweight cryptographic protocol that is both rapid and safe. This technique offers the necessary security criteria of conventional hash functions and the parameters for lightweight cryptographic protocols. The proposed architecture is tested and validated based on three key criteria: memory, speed, and power consumption, which are important factors in determining its lightweight nature. Based on the findings, the proposed architecture outperforms other lightweight protocols regarding memory usage, performance, and power consumption. For an appropriate FPGA-based application deployment, a thorough comparison of our suggested designs is detailed on various FPGA families.

Sodium-Ion Battery at Low Temperature: Challenges and Strategies.

Sodium-ion batteries (SIBs) have garnered significant interest due to their potential as viable alternatives to conventional lithium-ion batteries (LIBs), particularly in environments where low-temperature (LT) performance is crucial. This paper provides a comprehensive review of current research on LT SIBs, focusing on electrode materials, electrolytes, and operational challenges specific to sub-zero conditions. Recent advancements in electrode materials, such as carbon-based materials and titanium-based materials, are discussed for their ability to enhance ion diffusion kinetics and overall battery performance at colder temperatures. The critical role of electrolyte formulation in maintaining battery efficiency and stability under extreme cold is highlighted, alongside strategies to mitigate capacity loss and cycle degradation. Future research directions underscore the need for further improvements in energy density and durability and scalable manufacturing processes to facilitate commercial adoption. Overall, LT SIBs represent a promising frontier in energy storage technology, with ongoing efforts aimed at overcoming technical barriers to enable widespread deployment in cold-climate applications and beyond.

Optimizing Cloud Computing Performance by Integrating the Novel PSBR Service Broker Policy and Load Balancing Algorithms

As cloud computing advances, organizations' IT infrastructure and application deployment processes are moving to the cloud because cloud computing provides everything as a service over the Internet. The performance of a cloud-based application is based on proper datacenter selection and workload distribution within the selected datacenter. Service broker policies are used for suitable datacenter selection, and load balancing algorithms(LBA) are applied to distribute workloads. This paper is to evaluate the effect of a proposed service broker policy (PSBR) on the performance of cloud-based applications with LBA. To achieve the objective, the behavior of the TikTok application was modeled using the worldwide users’ statistics on the cloud simulation framework, namely CloudAnalyst. As a result, the average response time and data center processing time are measured. Next, the PSBR provides better results than the existing service proximity-based policy. This paper supports cloud service providers' benefits, from coordination between data center configuration, data center selection, and workload distribution to cloud users' identification of the appropriate procedures for their organization or application. PSBR with Active Monitoring had the best average response time of 75.1 ms, while SPR consistently exhibited higher average times across all algorithms, with the highest being 84.5 ms for Round Robin. Under the PSBR policy, Throttled had the lowest average processing time (4.67), while Round Robin had the highest (5.72). Similarly, under the SPR policy, Throttled maintained its efficiency with the lowest average (4.8), while Round Robin showed the highest (5.79).

Investigations on collision entropy and its variant for machine condition monitoring

Abstract The efficacy of sparsity measures and complexity measures in machine condition monitoring is well acknowledged. However, a comprehensive exploration of their theoretical underpinnings and interrelations remains incomplete. This study endeavors to address this gap by employing collision entropy as an illustrative case and establishing its relationship with sparsity measures. Through mathematical derivations, it is demonstrated that collision entropy exhibits a phenomenon termed as “bilateral reduction,” a characteristic indicative of complexity measures. Motivated by foundational concepts such as Shannon entropy and relative entropy, this paper introduces the concept of variation in collision entropy, which serves as a sparsity measure. Notably, both Shannon entropy and collision entropy, as specific cases of Rényi entropy, manifest the “bilateral reduction” effect and can be transformed into sparsity measures through certain methodologies. These insights contribute significantly to the exploration of the theoretical properties of Rényi entropy and its conversion into a cluster of sparse measures. This research augments the understanding of sparsity measures and complexity measures, offering theoretical insights crucial for their practical deployment in real-world applications.