Journal of Security in Computer Networks and Distributed Systems

Comprehensive Review of Data Poisoning Attacks in Green Distributed Systems

2026-02-06T09:51:33+00:00

Green distributed systems (GDS) represent a critical paradigm shift in computing, aiming to minimize energy consumption and carbon footprint through techniques like energy-aware scheduling, distributed learning, and edge-cloud offloading. However, the inherent trade-off between energy efficiency and system redundancy creates a novel and enlarged attack surface for data poisoning. This paper provides a comprehensive review of data poisoning attacks specifically targeting the sustainability objectives of GDS. The study defines data poisoning in this context not merely as an accuracy degradation threat, but as an energy integrity attack designed to induce energy-inefficient behavior, such as unnecessary recomputation or suboptimal carbon-aware scheduling. The unique attack vectors, including sponge poisoning that amplifies energy consumption in neural networks, and poisoning of energy-aware federated learning models are categorized. Furthermore, the critical defense challenges, where traditional security measures often incur prohibitive energy overheads are analyzed, negating the “green” objective. Finally, key future research directions, emphasizing the need for co-optimized robustness and energy-efficiency in GDS security policies are analyzed.

A Fault-Tolerant Framework for Improving Reliability in Distributed Cloud Systems

2026-04-04T09:20:36+00:00

Distributed cloud systems, which offer scalable and on-demand services to consumers globally, have emerged as the foundation of contemporary computer infrastructures. However, the availability and dependability of services can be severely impacted by these systems’ high susceptibility to hardware malfunctions, network outages, and node breakdowns. Although they increase system resilience, traditional fault-tolerance techniques like replication and checkpointing frequently result in higher computing overhead, storage costs, and system complexity. Hence, a lightweight fault-tolerant framework is presented in this paper to reduce resource overhead and enhance dependability in distributed cloud settings. To ensure service continuity in the event of node failures, the suggested model integrates effective recovery mechanisms, adaptive task reallocation, and failure detection. To assess system performance in the event of a failure, a simulation-based model was created. In comparison to non-fault-tolerant methods, experimental results show increased task completion rate, decreased recovery time, and improved system availability. The study emphasises how crucial it is to create fault-tolerance plans for next-generation distributed cloud systems that are both economical and effective.

Learning Without Sharing: A Privacy-aware Federated Learning Architecture for IoT Networks

2026-04-07T10:48:20+00:00

The boost of internet of things (IoT) devices in healthcare and everyday life has made it possible to monitor personal data such as vital signs, activity patterns, and information about people’s behavior, continuously. While that is in support of intelligent applications, centralized machine learning systems involve the transmission of sensitive data to cloud servers, which runs the risk of data breaches, misuse, and cyber-attacks. Federated learning (FL) is one way to get around this problem by allowing models to be trained in a decentralised manner without the sharing of raw data. However, the current federated frameworks assume that all devices involved are trustworthy, but current IoT environments are not composed of trustworthy devices operating in isolation, especially because some devices might be compromised or faulty, which creates resource limitations. The absence of dynamic trust assessment in existing federated systems is a gap that is identified in this research. To overcome this limitation, a trust-aware federated learning framework is proposed. The framework measures the device reliability with the help of update consistency, gradient deviation, and past performance before aggregating model updates. Adaptive weighting is used during the fusion of the models, which helps to reduce the impact of the suspicious or low-quality participants while still preserving decentralized privacy. Experimental analysis under the simulated adversarial scenarios shows the enhanced robustness against the poisoning attacks, stable convergence and the prediction accuracy is maintained compared to the standard federated averaging. The proposed approach achieves the objectives of providing improved security and reliability, which provides a scalable and practical solution for privacy-preserving intelligent IoT healthcare systems.

Cyber Kill Chain-based Multi-stage Attack Detection Using Graph Neural Networks

2026-04-08T06:20:39+00:00

Advanced persistent threats (APTs) and multi-stage cyberattacks have grown considerably in sophistication, systematically evading conventional rule-based intrusion detection systems. This study presents a novel detection framework that integrates the cyber kill chain (CKC) model with heterogeneous graph attention networks (HGAT) to enable accurate, stage-aware identification of complex attack campaigns in enterprise networks. By representing network telemetry, system call traces, and authentication logs as heterogeneous graphs, the proposed approach captures structural dependencies and temporal correlations across all seven CKC phases—from reconnaissance through actions on objectives. Experimental evaluation on four benchmark datasets—DARPA TC, CICIDS-2018, LANL 2015, and UNSW-NB15—demonstrates an overall detection accuracy of 97.40%, a macro F1-score of 96.93%, and a false positive rate of 1.80%, outperforming seven state-of-the-art baseline methods by significant margins. Stage-level analysis confirms consistently high detection rates across all CKC phases, while ablation studies validate the critical contribution of each architectural component. The results confirm that graph-based relational modelling combined with kill chain semantics provides a robust and interpretable solution for next-generation threat detection, with computational characteristics consistent with near-real-time enterprise deployment pending field validation.

An Intelligent Behavioral Authentication Framework for Securing Student Information Systems Using Hidden Markov Models

2026-04-08T06:44:39+00:00

The rapid digital transformation of educational institutions has significantly improved administrative efficiency in student registration and result processing; however, it has also exposed student information management systems (SIMS) to increasing cybersecurity threats. Traditional authentication mechanisms based on static credentials are inadequate for protecting sensitive academic records from unauthorized access, insider misuse, and sophisticated cyberattacks. This study presents an intelligent behavioral authentication framework for securing student information systems using hidden Markov models (HMMs). The proposed model analyzes sequential user behavior patterns such as login time, IP address, device information, and interaction frequency to distinguish legitimate users from suspicious actors. The system was trained and evaluated using authentication logs and the NSL-KDD dataset, with performance assessed using precision, recall, F1-score, and Matthews correlation coefficient (MCC). Experimental results demonstrate high detection capability across multiple attack categories, achieving strong precision and recall values for normal and DoS classes, and overall attack detection accuracy rates of 98.4% for SQL injection, 97.0% for cross-site scripting (XSS), 99.5% for brute-force login attempts, and 96.0% for malicious bot registration. The findings confirm that the HMM-based behavioral authentication model significantly improves anomaly detection performance and reduces false positives compared to traditional rule-based systems. The study establishes the effectiveness of probabilistic sequential modeling in strengthening adaptive access control mechanisms within student information systems.