GPH-International Journal of Computer Science and Engineering

A Scalable Distributed and Fault-Tolerant Architecture for Cloud-Based Machine Learning and Data Analysis

2026-07-23T11:21:34+00:00

The rapid growth of data-intensive applications has necessitated the development of scalable and efficient architectures for cloud-based machine learning and data analysis. This study proposes a scalable, distributed, and fault-tolerant architecture designed to address the challenges of processing large-scale and dynamic datasets in cloud environments. The architecture integrates key components, including data ingestion, distributed storage, parallel processing frameworks, machine learning pipelines, and application deployment layers, enabling seamless data flow and modular system design. It supports both batch and real-time data processing, making it adaptable to diverse analytical workloads. A design science and experimental research methodology was adopted to develop and evaluate the proposed system. Mathematical modeling and performance analysis were employed to assess system scalability, throughput, and latency under varying load conditions. Experimental results demonstrated that the architecture achieves significant improvements in processing efficiency and resource utilization through horizontal scaling. However, the findings also revealed sub-linear scalability behavior due to factors such as communication overhead, synchronization delays, and resource contention, which are inherent in distributed systems. The architecture exhibited strong fault tolerance and resilience, ensuring continuous system operation through redundancy and dynamic resource management. Performance evaluations highlighted an optimal operating region where throughput is maximized and latency remains within acceptable limits, beyond which system performance begins to degrade. The proposed architecture provides a robust and flexible framework for large-scale machine learning and data analysis in cloud environments. It offers a balance between scalability, performance, and reliability, making it suitable for modern data-driven applications. Future research may focus on enhancing auto-scaling strategies, optimizing workload distribution, and incorporating intelligent resource management techniques to further improve system efficiency.

An Explainable Hybrid Behavioral Analytics Framework for Insider Threat Detection in Hypervisor-Based Cloud Environments

2026-08-08T08:58:26+00:00

Insider threats remain one of the most challenging cybersecurity problems in hypervisor-based cloud environments because privileged users can exploit legitimate access to compromise virtual machines, manipulate hypervisor configurations, and exfiltrate sensitive data without triggering conventional security mechanisms. Existing approaches largely focus on enterprise user activities and rarely integrate hypervisor telemetry, behavioral analytics, Explainable Artificial Intelligence (XAI), and automated response into a unified framework. This study designed and developed an Explainable Hybrid Behavioral Analytics Framework for insider threat detection in hypervisor-based cloud environments. The study adopted the Design Science Research Methodology (DSRM) to design, implement, and evaluate the proposed framework. Hypervisor telemetry, virtual machine lifecycle events, privileged user activities, and behavioral logs were processed using User and Entity Behavior Analytics (UEBA), while a hybrid artificial intelligence engine integrating Random Forest, XGBoost, Long Short-Term Memory (LSTM), Autoencoder, and Isolation Forest were employed for threat detection. Explainability was achieved using SHAP and LIME, with dynamic risk scoring and automated response supporting real-time mitigation. Experimental evaluation using the CERT Insider Threat, LANL, and TWOS datasets achieved accuracies of 98.70%, 97.90%, and 98.30%, respectively, with F1-scores ranging from 97.35% to 98.25%, false positive rates of 1.20–1.60%, detection latency of 42–55 ms, and explainability scores of 0.91–0.93. The study concludes that integrating hypervisor telemetry, hybrid AI, and XAI provides an accurate, transparent, scalable, and proactive solution for insider threat detection in modern hypervisor-based cloud infrastructures.

Scalable and Enhanced Security Mechanism for Cloud Database System

2026-08-12T11:19:42+00:00

Developing secure and trusted distributed systems remains a major challenge in cloud computing environments. Ensuring data confidentiality, integrity, authentication, and secure access is particularly critical in distributed architectures such as wireless, peer-to-peer, and cloud database systems, where structural limitations and trust-related issues increase security vulnerabilities. Data breaches in such environments may result in operational disruptions, financial losses, legal liabilities, compliance violations, and reputational damage. Therefore, the development of scalable and enhanced security mechanism has become increasingly essential for protecting distributed cloud data systems. This thesis presents a HybridPDFSecurityTrainer model designed to enhance data security and confidentiality in distributed cloud database environments. The developed model integrates Convolutional Neural Network (CNN), Rivest–Shamir–Adleman (RSA), and Term Frequency–Inverse Document Frequency (TF-IDF) algorithms to provide intelligent document classification and secure data protection. The TF-IDF algorithm is used to extract textual features from Portable Document Format (PDF) files by converting documents into machine-readable numerical feature vectors. These feature vectors are subsequently processed using a CNN model to classify documents as benign or malicious based on detected patterns, sensitive information, and malicious content indicators. To strengthen data confidentiality and integrity, the RSA cryptographic algorithm is incorporated to perform secure encryption and decryption of classified documents before cloud storage and transmission. RSA also facilitates secure key exchange and prevents unauthorized access to sensitive information within distributed cloud environments. The integration of machine learning and cryptographic techniques provides a scalable and intelligent security framework capable of addressing modern cloud security challenges. This developed model was implemented using Python programming language and the Object-Oriented Analysis and Design Methodology (OOADM). Experimental evaluation showed that the system achieved 98% security accuracy with a low error rate, outperforming existing approaches in detecting malicious PDF documents while maintaining secure data transmission and storage. This study demonstrates that the HybridPDFSecurityTrainer model significantly improves confidentiality, integrity, and security management in distributed cloud database systems, making it suitable for deployment in real-world cloud computing environments.

A Model for Privacy-Preserving Smart Contracts in Cloud Computing

2026-08-13T08:19:55+00:00

The increasing adoption of cloud computing and blockchain-based smart contracts has transformed digital service delivery through decentralized automation, transparency, and trusted transaction execution. However, existing smart contract frameworks continue to face challenges related to privacy preservation, secure computation, intelligent access control, execution integrity, and auditability. Most existing solutions rely on isolated privacy-preserving mechanisms, exposing sensitive information during computation and limiting scalability and overall system performance. This study developed a Model for Privacy-Preserving Smart Contract in Cloud Computing by integrating Zero-Knowledge Proofs (ZKP), Secure Multi-Party Computation (SMPC), Trusted Execution Environments (TEE), Federated Learning (FL), Differential Privacy (DP), Autoencoder-based anomaly detection, GraphSAGE Graph Neural Networks (GNN), Proximal Policy Optimization (PPO), and Blockchain Smart Contracts within a unified architecture. The study adopted the Design Science Research Methodology (DSRM), while Object-Oriented Analysis and Design (OOAD) guided system implementation. The proposed model was evaluated using the CICIDS2017 cybersecurity benchmark dataset across privacy, security, execution integrity, auditability, scalability, computational performance, and cost efficiency. Experimental results achieved 96% privacy preservation, 94% security strength, 99% execution integrity, 98% auditability, and 90% scalability, while the Artificial Intelligence Privacy Engine attained 98.91% validation accuracy, 0.9962 ROC-AUC, 0.9490 Macro F1-score, and 0.9718 Matthews Correlation Coefficient (MCC). Comparative analysis against RBAC, ABAC, and blockchain-based frameworks demonstrated superior performance in privacy preservation, secure computation, intelligent authorization, and auditability. The proposed model provides a practical, scalable, and intelligent solution for secure smart contract execution in privacy-sensitive cloud computing environments.

An Intelligent Privacy-Preserving Access Control Framework for Cloud-Based Smart Contracts

2026-08-13T08:36:17+00:00

Cloud computing has become the preferred platform for deploying blockchain-enabled smart contracts because of its scalability and flexibility. However, existing access control mechanisms such as Role-Based Access Control (RBAC), Attribute-Based Access Control (ABAC), and conventional blockchain authentication expose sensitive user information during authentication, rely on static authorization policies, and lack intelligent mechanisms for detecting evolving cyber threats. This study proposes an Intelligent Privacy-Preserving Access Control Framework for Cloud-Based Smart Contracts that integrates Modified Groth16 Zero-Knowledge Proofs (ZKP), Secure Multi-Party Computation (SMPC), Trusted Execution Environments (TEE), Federated Learning, Differential Privacy, GraphSAGE Graph Neural Networks, Autoencoder-based anomaly detection, Proximal Policy Optimization (PPO), and Blockchain Smart Contracts. The framework enables credential-free authentication, confidential collaborative computation, adaptive authorization, intelligent threat detection, and immutable blockchain auditing without compromising user privacy. The proposed framework was implemented and evaluated using the CICIDS2017 cybersecurity dataset. Experimental results achieved 96.4% privacy preservation, 94.1% security strength, 99.0% execution integrity, 98.7% auditability, 90.3% scalability, 88.6% computational performance, 86.9% cost efficiency, 98.91% validation accuracy, 99.62% ROC-AUC, 94.90% Macro F1-Score, and an overall system fitness of 94.23%. Comparative evaluation against Hawk, Zether, Ekiden, and a Federated Learning-only IDS demonstrated superior performance across all evaluation metrics. The proposed framework therefore provides an intelligent, scalable, and privacy-preserving access control solution suitable for next-generation cloud-based smart contract systems.

DESIGN AND IMPLEMENTATION OF A MACHINE LEARNING MODEL FOR NETWORK INTRUSION DETECTION

2026-08-13T10:59:56+00:00

Network intrusion detection has become difficult because enterprise traffic is high-volume, heterogeneous and continuously altered by encryption, cloud adoption, remote access and adversarial behaviour. This study designed and implemented a supervised machine-learning model that classifies network flows as benign or intrusive while preserving operational interpretability and low false-alarm rates. The proposed architecture integrates traffic capture, flow aggregation, data cleaning, encoding, scaling, feature screening, imbalance-aware model training, thresholded inference and alert generation. Five classifiers logistic regression, decision tree, random forest, gradient boosting and linear support vector machine were compared on a reproducible 15,000-record flow-like benchmark containing 32 statistical attributes and an 82:18 benign-to-intrusion ratio. Evaluation used a stratified 75:25 split and accuracy, precision, recall, F1-score, receiver operating characteristic area under the curve and confusion-matrix analysis. Random forest produced the strongest overall balance, attaining 94.72% accuracy, 97.10% precision, 73.32% recall and an F1-score of 83.55% in the implemented benchmark, while linear models offered lower computational cost but weaker nonlinear discrimination. The findings show that preprocessing, class-aware learning and per-class metrics are as important as classifier choice. The paper contributes an implementation-ready framework, pseudocode, data schema, model-comparison evidence and deployment controls for drift, privacy and retraining. Because benchmark accuracy may overstate field performance, the proposed system requires temporal validation and monitoring before production use. The architecture is suitable for campus, small-enterprise and cloud-edge networks where explainable flow-based detection is preferred to payload inspection.