Data Distribution Research Articles

Imaging swiFTly: streaming widefield Fourier Transforms for large-scale interferometry

We describe a scalable distributed imaging algorithm framework for next-generation radio telescopes, managing the Fourier transform from apertures to sky (or vice versa) with a focus on minimising memory load, data transfers, and computation. Our algorithm uses smooth window functions to isolate the influence between specific regions of spatial-frequency and image space. This allows the distribution of image data between nodes and the construction of segments of frequency space exactly when and where needed. The developed prototype distributes terabytes of image data across many nodes, while generating visibilities at throughput and accuracy competitive with existing software. Scaling is demonstrated to be better than cubic in problem complexity (for baseline length and field of view), reducing the risk involved in growing radio astronomy processing to large telescopes like the Square Kilometre Array.

Prevalence, factors associated and management of needle phobia among the general population in Saudi Arabia and Egypt

ObjectiveThis study aims to assess the prevalence of needle phobia among Saudi and Egyptian adult populations. In addition, underlying causes and strategies that can be utilized to address needle fear were investigated.MethodsA cross-sectional online survey study was conducted in Saudi Arabia and Egypt between 1 May and 30 June 2023. Participants aged 18 years and above and living in Saudi Arabia and Egypt were eligible to complete the survey. Participants were invited to participate in this study through social media platforms (Facebook, X, Snapchat, and Instagram). A convenience sampling technique was used to recruit the study participants. A 21-item questionnaire consisting of four sections including a Likert scale score was used to answer the research objectives. Numeric data were presented as mean ± SD. For categorical variables, percentages were used. Comparison between groups were made by Student’s t-test or Mann Whitney test according to data distribution. Chi squared tests for categorical values were conducted. A binary logistic regression analysis was conducted to investigate factors associated with needle phobia.ResultsA total of 4065 participants were involved in this study (Saudi Arabia: 2628 and Egypt: 1437). Around one-third of the study participants (36.5%) confirmed that they have needle phobia. Most of the study participants (81.1%) reported that they have had needle phobia since they were under 18 years of age. Pain, general anxiety, and fear of making a mistake during the procedure were the most commonly reported contributors for fear of needles during or before a medical procedure. Around 15.8% of the study participants reported that they have tried to get rid of phobia from needles. Non-surgical alternatives (such as oral medications and patches) and using smaller/thinner needles were the most commonly reported interventions that reduced fear of needles. Binary logistic regression analysis identified that females, those who are aged (41–50 years), widowed, those with bachelor’s degrees and higher education, and those unemployed were more likely to have needle phobia compared to others.ConclusionOur study highlighted the high prevalence of needle fear within an adult population in Egypt and Saudi Arabia. Females, those who are aged (41–50 years), those widowed, those with higher education degrees, those unemployed, those working in the health sector and people with low income were more likely to have needle phobia compared to others.

Provable Unrestricted Adversarial Training without Compromise with Generalizability.

Adversarial training (AT) is widely considered as the most promising strategy to defend against adversarial attacks and has drawn increasing interest from researchers. However, the existing AT methods still suffer from two challenges. First, they are unable to handle unrestricted adversarial examples (UAEs), which are built from scratch, as opposed to restricted adversarial examples (RAEs), which are created by adding perturbations bound by an lp norm to observed examples. Second, the existing AT methods often achieve adversarial robustness at the expense of standard generalizability (i.e., the accuracy on natural examples) because they make a tradeoff between them. To overcome these challenges, we propose a unique viewpoint that understands UAEs as imperceptibly perturbed unobserved examples. Also, we find that the tradeoff results from the separation of the distributions of adversarial examples and natural examples. Based on these ideas, we propose a novel AT approach called Provable Unrestricted Adversarial Training (PUAT), which can provide a target classifier with comprehensive adversarial robustness against both UAE and RAE, and simultaneously improve its standard generalizability. Particularly, PUAT utilizes partially labeled data to achieve effective UAE generation by accurately capturing the natural data distribution through a novel augmented triple-GAN. At the same time, PUAT extends the traditional AT by introducing the supervised loss of the target classifier into the adversarial loss and achieves the alignment between the UAE distribution, the natural data distribution, and the distribution learned by the classifier, with the collaboration of the augmented triple-GAN. Finally, the solid theoretical analysis and extensive experiments conducted on widely-used benchmarks demonstrate the superiority of PUAT.

Measurement Guidance in Diffusion Models: Insight from Medical Image Synthesis.

In the field of healthcare, the acquisition of sample is usually restricted by multiple considerations, including cost, labor- intensive annotation, privacy concerns, and radiation hazards, therefore, synthesizing images-of-interest is an important tool to data augmentation. Diffusion models have recently attained state-of-the-art results in various synthesis tasks, and embedding energy functions has been proved that can effectively guide the pre-trained model to synthesize target samples. However, we notice that current method development and validation are still limited to improving indicators, such as Fréchet Inception Distance score (FID) and Inception Score (IS), and have not provided deeper investigations on downstream tasks, like disease grading and diagnosis. Moreover, existing classifier guidance which can be regarded as a special case of energy function can only has a singular effect on altering the distribution of the synthetic dataset. This may contribute to in-distribution synthetic sample that has limited help to downstream model optimization. All these limitations remind that we still have a long way to go to achieve controllable generation. In this work, we first conducted an analysis on previous guidance as well as its contributions on further applications from the perspective of data distribution. To synthesize samples which can help downstream applications, we then introduce uncertainty guidance in each sampling step and design an uncertainty-guided diffusion models. Extensive experiments on four medical datasets, with ten classic networks trained on the augmented sample sets provided a comprehensive evaluation on the practical contributions of our methodology. Furthermore, we provide a theoretical guarantee for general gradient guidance in diffusion models, which would benefit future research on investigating other forms of measurement guidance for specific generative tasks. Codes and models are available at: https://github.com/yangqy1110/MGDM.

Study of 113In(α, α) elastic scattering to determine α-optical potential relevant for astrophysical γ-process

Abstract The α-optical potential is one of the key input parameters used to measure the reaction rate of the (γ, α)-process using the Hauser-Feshbach statistical model and the principle of detailed balance. α-elastic scattering experiment on 113In p-nucleus was carried out in the energy range E lab = 26−32 MeV. The vacuum evaporation technique was used to prepare the 113In target (∼86 μg cm−2). An energy-dependent local optical potential parameters set was obtained by analysing the experimental elastic scattering angular distribution data. The local potential parameters are extrapolated for lower energies and are used to calculate the 113In(α, γ) reaction cross-section.

Meta-Learning for Multi-Family Android Malware Classification

With the emergence of smartphones, Android has become a widely used mobile operating system. However, it is vulnerable when encountering various types of attacks. Every day, new malware threatens the security of users’ devices and private data. Many methods have been proposed to classify malicious applications, utilizing static or dynamic analysis for classification. However, previous methods still suffer from unsatisfactory performance due to two challenges. First, they are unable to address the imbalanced data distribution problem, leading to poor performance for malware families with few members. Second, they are unable to address the zero-day malware (zero-day malware refers to malicious applications that exploit unknown vulnerabilities) classification problem. In this paper, we introduce an innovative meta -learning approach for m ulti-family A ndroid m alware c lassification named Meta-MAMC , which uses meta-learning technology to learn meta-knowledge (i.e. the similarities and differences among different malware families) of few-family samples and combines new sampling algorithms to solve the above challenges. Meta-MAMC integrates (i) the meta-knowledge contained within the dataset to guide models in learning to identify unknown malware, and (ii) more accurate and diverse tasks based on novel sampling strategies, as well as directly adapting meta-learning to a new few-sample and zero-sample task to classify families. We have evaluated Meta-MAMC on two popular datasets and a corpus of real-world Android applications. The results demonstrate its efficacy in accurately classifying malicious applications belonging to certain malware families, even achieving 100% classification in some families.

Artificial neural network, machine learning modelling of compressive strength of recycled coarse aggregate based self-compacting concrete.

This research study aims to understand the application of Artificial Neural Networks (ANNs) to forecast the Self-Compacting Recycled Coarse Aggregate Concrete (SCRCAC) compressive strength. From different literature, 602 available data sets from SCRCAC mix designs are collected, and the data are rearranged, reconstructed, trained and tested for the ANN model development. The models were established using seven input variables: the mass of cementitious content, water, natural coarse aggregate content, natural fine aggregate content, recycled coarse aggregate content, chemical admixture and mineral admixture used in the SCRCAC mix designs. Two normalization techniques are used for data normalization to visualize the data distribution. For each normalization technique, three transfer functions are used for modelling. In total, six different types of models were run in MATLAB and used to estimate the 28th day SCRCAC compressive strength. Normalization technique 2 performs better than 1 and TANSING is the best transfer function. The best k-fold cross-validation fold is k = 7. The coefficient of determination for predicted and actual compressive strength is 0.78 for training and 0.86 for testing. The impact of the number of neurons and layers on the model was performed. Inputs from standards are used to forecast the 28th day compressive strength. Apart from ANN, Machine Learning (ML) techniques like random forest, extra trees, extreme boosting and light gradient boosting techniques are adopted to predict the 28th day compressive strength of SCRCAC. Compared to ML, ANN prediction shows better results in terms of sensitive analysis. The study also extended to determine 28th day compressive strength from experimental work and compared it with 28th day compressive strength from ANN best model. Standard and ANN mix designs have similar fresh and hardened properties. The average compressive strength from ANN model and experimental results are 39.067 and 38.36 MPa, respectively with correlation coefficient is 1. It appears that ANN can validly predict the compressive strength of concrete.

Location-based clinical and angiographic profile of brain arteriovenous malformations - a single-center observational study.

The location of brain arteriovenous malformations (bAVM) is one of the most relevant prognostic factors included in surgical, endovascular and radiosurgical scores. However, their characteristics according to location are seldom described. The goal of this study was to describe the clinical and angiographic characteristics of bAVM classified according to their location. This retrospective observational study included patients diagnosed with bAVM and attending a national referral hospital during the period 2010-2020. Data regarding clinical and angiographic variables were extracted, including characteristics on nidus, arterial afferents, venous drainage and associated aneurysms. BAVM were classified in 8 groups according to their location: frontal, temporal, parieto-occipital, periventricular, deep, cerebellar, brainstem and mixed. Data distribution for each group was determined and between-group differences were assessed. A total of 269 bAVM (in 258 patients) were included. The most frequent location was parieto-occipital; and the least frequent, brainstem. Statistically significant differences were observed between groups for most studied variables, including: clinical presentation, functional status at admission; nidus size and density, classification according to the Spetzler-Martin, Buffalo and modified Pollock-Flickinger scales; number, diameter, origin and type of afferents; number, diameter, type and direction of venous drainage, retrograde venous flow; and presence and size of flow-related aneurysms. The clinical and angiographic differences observed between brain AVM groups allow the formulation of profiles according to their location.

Analysis of vertical differentiation of vegetation in Taishan World Heritage site based on cloud model

While the forests on Mount Taishan are predominantly man-made, there is a notable vertical variation in vegetation. This study employs the method of cloud model, quantifying uncertainty (fuzziness and randomness) of things. Utilizing digital elevation model (DEM) and vegetation distribution data, we constructed elevation cloud models for Mount Taishan’s deciduous broad-leaved, temperate coniferous, and mixed coniferous-broadleaved forests. Using three numerical features of the cloud model—Expectation (EX), Entropy (EN), and Hyper-entropy (HE)—we quantitatively analyzed the macro regularity and local heterogeneity of Mount Taishan’s forests vertical distribution from the perspective of uncertainty theory. The results indicate: (1) The EX of the core zone elevation of deciduous broad-leaved forest is 716.65 m, temperate coniferous forest is 1053.51 m, and mixed coniferous-broadleaved forest is 1384.09 m. The variation range of the core zone distribution height is smaller in the mixed coniferous-broadleaved forest (EN: 53.74 m) compared to deciduous broad-leaved forest (EN: 99.63 m) and temperate coniferous forest (EN: 121.70 m). (2) The fuzziness and randomness of the distribution height of the lower extension zones of deciduous broad-leaved forest and temperate coniferous forest (EN: 75.15 m, 184.56 m; HE: 24.09 m, 63.54 m) are greater than those of the upper extension zones (EN: 44.75 m, 42.49 m; HE: 14.48 m, 13.23 m). (3) The distribution fuzziness and randomness within temperate coniferous forests exceed those of deciduous broad-leaved forests. Within the core zones, the uncertainty regarding the vertical distribution of vegetation across different aspects remains consistent, which retains the characteristic of man-made forests. However, in transition areas, there is significant disparity, reflecting the adaptive relationship between vegetation and its environment to some extent. In the upper and lower extension zones of deciduous broad-leaved forests, the EX values for the vertical distribution height of mixed coniferous and broad-leaved forests differ significantly from those of deciduous broad-leaved forests (the difference is 22.82–39.15 m), yet closely resemble those of temperate coniferous forests (the difference is 4.79–7.94 m). This suggests a trend wherein deciduous broad-leaved tree species exhibit a proclivity to encroach upon coniferous forest habitats. The elevation cloud model of vertical vegetation zones provides a novel perspective and method for the detailed analysis of Mount Taishan’s vegetation vertical differentiation.

Uneven primary healthcare supply of rural doctors and medical equipment in remote China: community impact and the moderating effect of policy intervention

BackgroundUnequal access to primary healthcare (PHC) has become a critical issue in global health inequalities, requiring governments to implement policies tailored to communities’ needs and abilities. However, the place-based facility dimension of PHCs is oversimplified in current healthcare literature, and formulating the equity-oriented PHC spatial planning remains challenging without understanding the multiple impacts of community socio-spatial dynamics, particularly in remote areas. This study aims to push the boundary of PHC studies one step further by presenting a nuanced and dynamic understanding of the impact of community environments on the uneven primary healthcare supply.MethodsFocusing on Shuicheng, a remote rural area in southwestern China, multiple data are included in this village-based study, i.e., the facility-level healthcare statistics data (2016–2019), the statistical yearbooks, WorldPop, and Chinese GDP’s spatial distribution data. We evaluate villages’ PHC service capacity using the number of doctors and essential equipment per capita, which are the major components of China’s PHC delivery. The indicators describing community environments are selected based on extant literature and China’s planning paradigms, including town- and village-level factors. Gini coefficients and local spatial autocorrelation analysis are used to present the divergences of PHC capacity, and multilevel regression model and (heterogeneous) difference in difference model are used to examine the driving role of community environments and the dynamics under the policy intervention.ResultsDespite the general improvement, PHC inequalities remain significant in remote rural areas. The village’s location, aging, topography, ethnic autonomy, and economic conditions significantly influence village-level PHC capacity, while demographic characteristics and healthcare delivery at the town level are also important. Although it may improve the hardware setting in village clinics (coef. = 0.350), the recent equity-oriented policy attempts may accelerate the loss of rural doctors (coef. = − 0.517). Notably, the associations between PHC and community environments are affected inconsistently by this round of policy intervention. The town healthcare centers with higher inpatient service capacity (coef. = − 0.514) and more licensed doctors (coef. = − 0.587) and nurses (coef. = − 0.344) may indicate more detrimental policy effects that reduced the number of rural doctors, while the centers with more professional equipment (coef. = 0.504) and nurses (coef. = 0.184) are beneficial for the improvement of hardware setting in clinics.ConclusionsThe findings suggest that the PHC inequalities are increasingly a result of joint social, economic, and institutional forces in recent years, underlining the increased complexity of the PHC resource allocation mechanism. Therefore, we claim the necessity to incorporate a broader understanding of community orientation in PHC delivery, particularly the interdisciplinary knowledge of the spatial lens of community, to support its sustainable development. Our findings also provide timely policy insights for ongoing primary healthcare reform in China.

Broken Rotor Bar Detection Based on Steady-State Stray Flux Signals Using Triaxial Sensor with Random Positioning

This paper investigates the detection of broken rotor bar in squirrel cage induction motors using a novel approach of randomly positioning a triaxial sensor over the motor surface. This study is conducted on two motors under laboratory conditions, where one motor is kept in a healthy state, and the other is subjected to a broken rotor bar (BRB) fault. The induced electromotive force of the triaxial coils, recorded over ten days with 100 measurements per day, is statistically analyzed. Normality tests and graphical interpretation methods are used to evaluate the data distribution. Parametric and non-parametric approaches are used to analyze the data. Both approaches show that the measurement method is valid and consistent over time and statistically distinguishes healthy motors from those with BRB defects when a reference or threshold value is specified. While the comparison between healthy motors shows a discrepancy, the quantitative analysis shows a smaller estimated difference in mean values between healthy motors than comparing healthy and BRB motors.

Deep convolutional neural network based synthetic minority over sampling technique: a forfending model for fraudulent credit card transactions in financial institution

Today, fraudulent credit card transaction are committed by unauthorized individuals and organization exploring methods such as phishing and social engineering fraud. Several researchers proposed Machine Learning (ML) techniques to serve as solutions to credit card fraud detection. However, ML suffers imbalance class distribution, high dimensionality and sparsity problem. These concerns make it extremely difficult in credit card fraud detection. Therefore, this study proposes Deep Convolutional Neural Network (DCNN) with Synthetic Minority Oversampling Techniques (SMOTE) to serve as an ideal solution. Data augmentation method of SMOTE oversampling techniques is incorporated into the DCNN that solve the class imbalance trial and in improving the performance. Datasets with 284807 records with 31 features gotten via Kaggle repository is used. Implementation is carried out on Google Colab containing cloud-based platform embedding Jupyter notebook environment with Graphical processing unit (GPUs). Two experiment were carried out where the first experiment is used to determine suitable models among baseline techniques: Logistics Regression (LR), Random Forest (RF), Isolation forest, and single DL model of Multiple layer perceptron (MLP). The baseline models yielded an overfitting accuracy score, with recall, specificity, precision, and F1-score all presenting 1.00% respectively. This outcome is not sufficient in establishing findings on imbalance data distribution as its biased. This led to the construction of a new ML model absorbing Light Gradient Boosting Machine (LGBM), with Artificial Neural Network (ANN) and proposed DCNN+SMOTE for the second experiment phase alongside baseline models. Experimental results via simulation show the proposed DCNN+SMOTE yielded overwhelming superclass performance across board displaying 1.00% results respectively better than baseline models. It Error Rate (ER) and Null Error Rate (NER) is 0.00% distinctly. While, the False Positive Rate (FPR) yields 0.001% result lesser than the baseline models.

Implementasi Google Cloud Pub/Sub menggunakan Metode Subscription Pull dalam Pengiriman Data Promosi Toko di PT XYZ

As a company adopts a distributed business model, scalability issues often arise in the distribution of data from server to client. PT XYZ, as one of the companies facing this challenge, frequently encounters delays and request blocking when using synchronous communication models such as RESTful API. Additionally, there is resource and time wastage on the server side as it has to handle store requests one by one, resulting in losses for the store due to delayed access to required data for business processes. This research proposes a solution to address these issues by implementing a publish/subscribe communication model. Through the implementation of this architecture, clients and servers are separated by a broker, namely Google Cloud. Pub/sub technology is implemented using the Python programming language with the Waterfall development method. The results of this research include the design of a client-side application capable of receiving data from a previously deployed broker. Test results demonstrate that the client can receive data at any time without disrupting the data fetching process for other clients.

Multi-scale Target Detection Algorithm of Optical Remote Sensing Image Based on Improved YOLOv8

Abstract. With the progress of remote sensing sensors, the quality of optical remote sensing image is significantly improved, and target detection on it can extract rich feature information. However, due to the characteristics of remote sensing image with various target sizes and a large proportion of the number of small targets, increasing the difficulty in target detection for it. In response to this challenge, this paper proposes an improved YOLOv8 algorithm for multi-scale target detection of optical remote sensing images. First, we propose a PSPPF module, which improves the model's ability to adapt to different data distributions; Second, DSConv is introduced into the Backbone structure of YOLOv8 to reduce the complexity of the network while maintaining the performance of model detection; Finally, we replace the original loss function CIoU with MPDIoU to improve the localization accuracy of the prediction box. Applying the improved algorithm to the public dataset NWPU VHR-10, the mAP value of the our algorithm is 95.1%, which is 3.0% higher than that of the original YOLOv8, indicating that the proposed algorithm is able to effectively detect multi-scale targets in optical remote sensing images.

Sistem Informasi Pembelajaran Taman Pendidikan Al-Qur’an di Yayasan Al-Muttaqien Jadid (TPQ)

This research aims to determine the TPA academic system currently running at Al-Muttaqien Jadid. This research is descriptive research using qualitative methods. Data collection was carried out by interviews, observation and documentation, as well as using literature studies from various sources related to the questions asked in the research. It was found that there was a problem with the TPA Academic system, including regarding student registration, class division and scheduling, which still uses the bookkeeping method, it is difficult to get information regarding student data, class data, teacher data and subject data. Apart from that, the process of class division and scheduling cannot be completed quickly due to difficulties in finding data, the number of students being placed in excess of the quota often occurs, so mistakes or mistakes often occur in recording students' grades. So that registration, payment and distribution data can be processed quickly and accurately, a web-based application is needed. By implementing this application, it can help academic processing in preparing payment reports and student registration so that the management of TKA-TPA in the data collection process makes coordination and coaching easier. From the research results, it can be concluded that registration, payment, class division and scheduling using the bookkeeping method requires quite a long time in processing data and storing data.

A new poisson-exponential-gamma distribution for modelling count data with applications

A new poisson-exponential-gamma distribution for modelling count data with applications

Comparative analysis of key factors influencing urban green space in Mashhad, Iran (1988–2018)

This paper analyzed the role of national economic factors, in addition to some key city-level variables, in the variation of the urban green space (UGS) in Mashhad City (Iran) during three decades (1998–2018). The correlation result revealed the effects of the increasing trend of land price, population rate, and construction of built-up areas in the decreasing trend of UGS in the study area (R from − 0.95 to -0.99 at p-value > 95%). Also, the country-level economic factors, i.e., GDP per capita, oil price, export total value, and FDI, represented the overall increasing trend from 1988 to 2018, correlating with the decrease of green space areas (R from − 0.76 to -0.92 at p-value > 75%). Some statistical analyses, such as the run-test, the skewness and kurtosis tests, the Kolmogorov-Smirnov test, and the ANOVA test were done to confirm the normality of the data distribution and reliability of the results. Ultimately, clustering the research variables based on the significance and confidence levels of the estimated correlation results revealed that the change in the oil price and national export values in the petroleum-dependent economy of Iran can be assumed as the lead key economic factors to fluctuate all city-level variables, particularly the UGS variations.

An Adaptive Enhancement Method of Malicious Traffic Samples Based on DCGAN-ResNet System

A malicious traffic sample adaptive enhancement device based on Deep Convolutional Generative Adversarial Network (DCGAN) is designed to address the issue of imbalanced network traffic data distribution, aiming to enhance the accuracy and efficiency of anomaly detection. By leveraging generative adversarial network technology, this device can generate samples similar to real malicious traffic to balance the training dataset. It utilizes the generator and discriminator of the Deep Convolutional Generative Adversarial Network (DCGAN), combined with the residual network (ResNet) in the CNN model, to enhance the quality of generated samples. The device can switch states to adapt to various network environments and has been experimentally validated for its effectiveness and feasibility.Moreover, employing an adaptive device, the samples of malicious traffic are adjusted. Experimental analysis demonstrates that the device significantly enhances the accuracy of anomaly traffic detection, improves robustness, and provides robust support for network security protection.

Rapid determination of seismic influence field based on mobile communication big data-A case study of the Luding Ms 6.8 earthquake in Sichuan, China.

Smartphone location data provide the most direct field disaster distribution data with low cost and high coverage. The large-scale continuous sampling of mobile device location data provides a new way to estimate the distribution of disasters with high temporal-spatial resolution. On September 5, 2022, a magnitude 6.8 earthquake struck Luding County, Sichuan Province, China. We quantitatively analyzed the Ms 6.8 earthquake from both temporal and geographic dimensions by combining 1,806,100 smartphone location records and 4,856 spatial grid locations collected through communication big data with the smartphone data under 24-hour continuous positioning. In this study, the deviation of multidimensional mobile terminal location data is estimated, and a methodology to estimate the distribution of out-of-service communication base stations in the disaster area by excluding micro error data users is explored. Finally, the mathematical relationship between the seismic intensity and the corresponding out-of-service rate of communication base stations is established, which provides a new technical concept and means for the rapid assessment of post-earthquake disaster distribution.

Robust-DSN: A Hybrid Distributed Replication and Encoding Network Grouped with a Distributed Swarm Workflow Scheduler

In many distributed applications such as the Internet of Things (IoT), large amounts of data are being generated that require robust storage solutions. Traditional cloud solutions, although efficient, often lack trust and transparency because of centralized management. To address these issues, we present Robust-DSN, a distributed storage network leveraging the hybrid distributed replication and encoding network (HYDREN) and the distributed swarm workflow scheduler (DSWS) as its main components. Our system uses an interplanetary file system (IPFS) as an underlay storage network and segments it into multiple regions to distribute the failure domain and improve the data’s proximity to users. HYDREN incorporates Reed–Solomon encoding and distributed replication to improve file availability, while DSWS optimizes resource allocation across the network. The uploaded file is encoded into chunks and distributed across distinct optimal nodes leveraging lightweight multithreading. Additionally, Robust-DSN verifies the integrity of all chunks by preserving the hashes when uploading and validating each chunk while downloading. The proposed system provides a comprehensive solution for resilient distributed data storage, focusing on the key challenges of data availability, integrity, and performance. The results reveal that compared with a state-of-the-art system, the proposed system improves file recovery by 15%, even with a 50% peer failure rate. Furthermore, with replication factor 4 and the same failure resilience as IPFS, it saves 50% storage and enhances file recovery by 8%. Robust-DSN acts as a distributed storage platform for emerging technologies, expanding storage system capabilities in a wide range of distributed applications.