Deep learning technology: enabling safe communication via the internet of things

Frontiers in Communications and Networks, Journal Year: 2025, Volume and Issue: 6

Published: Feb. 4, 2025

Introduction The Internet of Things (IoT) is a new technology that connects billions devices. Despite offering many advantages, the diversified architecture and wide connectivity IoT make it vulnerable to various cyberattacks, potentially leading data breaches financial loss. Preventing such attacks on ecosystem essential ensuring its security. Methods This paper introduces software-defined network (SDN)-enabled solution for vulnerability discovery in systems, leveraging deep learning. Specifically, Cuda-deep neural (Cu-DNN), Cuda-bidirectional long short-term memory (Cu-BLSTM), Cuda-gated recurrent unit (Cu-DNNGRU) classifiers are utilized effective threat detection. approach includes 10-fold cross-validation process ensure impartiality findings. most recent publicly available CICIDS2021 dataset was used train hybrid model. Results proposed method achieves an impressive recall rate 99.96% accuracy 99.87%, demonstrating effectiveness. model also compared benchmark classifiers, including Cuda-Deep Neural Network, Cuda-Gated Recurrent Unit, (Cu-DNNLSTM Cu-GRULSTM). Discussion Our technique outperforms existing based evaluation criteria as F1-score, speed efficiency, accuracy, precision. shows strength detection highlights potential combining SDN with learning assessment.

Language: Английский

---

Centralised vs. decentralised federated load forecasting in smart buildings: Who holds the key to adversarial attack robustness?

Energy and Buildings, Journal Year: 2024, Volume and Issue: 324, P. 114871 - 114871

Published: Oct. 4, 2024

Language: Английский

---

Charting a Path Forward for the International Journal on Networked and Distributed Computing

˜The œInternational journal of networked and distributed computing, Journal Year: 2024, Volume and Issue: 12(2), P. 165 - 169

Published: Aug. 9, 2024

Abstract The International Journal of Networked and Distributed Computing has been pioneering research that advances our understanding networked distributed computing. As the newly appointed Editor-in-Chief, in this editorial, I articulate my vision for future journal, emphasizing its commitment to maintaining rigorous standards while embracing technological advancements. Key areas focus will be extended include Quantum Internet, Serverless Computing, Intelligence, convergence HPC Cloud Continuum, sustainable computing practices. Innovative initiatives, such as enhancing editorial board, forging strategic partnerships, and, possibly, expanding article types, are introduced elevate journal’s impact relevance. feasibility establishing an ad hoc periodic series works realized collaboration with key researchers different fields, focused on recent trends, findings, roadmaps investigated. process characterizes aimed at ensuring academic integrity transparency, not affected.

Language: Английский

---

New Continual Federated Learning System for Intrusion Detection in SDN‐Based Edge Computing

Concurrency and Computation Practice and Experience, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 27, 2024

ABSTRACT Software Defined Networking (SDN) is an open network approach that has been proposed to address some of the main problems with traditional networks. However, SDN faces cybersecurity issues. To provide a defense against attacks, Intrusion Detection System (IDS) needs be updated and included into architecture on regular basis. Machine learning methods have proved effective in detecting intrusions SDN. Moreover, these techniques pose problem significant computational overload absence updates when new cyber‐attacks appear. issues, we propose SDN‐based cloud intrusion detection system called Continual Federated Learning (CFL). In CFL, modify classical federated process by granting more important dynamic role each participating client. On one hand, it can trigger this whenever type detected. other once model identified, customer decide whether or not deploy his network. addition, verify accuracy CFL system, formally specified communication protocol. This specification organizes exchanges between different communicating entities involved CFL. specification, described using PROMELA language checked associated SPIN tool. experimental side, deployed computing environment. We defined scenarios, client decides locally newly obtained model. The decision based modified metric where integrate severity intrusions. Experimental results private local datasets show efficiently accurately detect types while preserving confidentiality. Thus, considered promising for edge computing.

Language: Английский

---

Learning Automata-Based Enhancements to RPL: Pioneering Load-Balancing and Traffic Management in IoT

Published: July 17, 2024

The Internet of Things (IoT) signifies a revolutionary technological advancement, enhancing various applications through device interconnectivity while introducing significant challenges due to these devices' limited hardware and communication capabilities. To navigate complexities, the Engineering Task Force (IETF) has tailored Routing Protocol for Low-Power Lossy Networks (RPL) meet unique demands IoT environments. However, RPL struggles with traffic congestion load distribution issues, negatively impacting network performance reliability. This paper presents novel enhancement by integrating learning automata designed optimize distribution. enhanced protocol, Learning Automata-based Load-Aware (LALARPL), dynamically adjusts routing decisions based on real-time conditions, achieving more effective balancing significantly reducing congestion. Extensive simulations reveal that this approach outperforms existing methodologies, leading notable improvements in packet delivery rates, end-to-end delay, energy efficiency. findings highlight potential our enhance operations extend lifespan components. effectiveness refining processes within offers valuable insights may drive future advancements networking, aiming robust, efficient, sustainable architectures.

Language: Английский

---

NFV and Secure Cognitive SDN for Educational Backbone Network Deployment

Advances in wireless technologies and telecommunication book series, Journal Year: 2024, Volume and Issue: unknown, P. 132 - 153

Published: June 14, 2024

Software defined networks (SDN) and wireless cognitive radio (CRN) are examined within the context of dynamic spectrum management. The features include control data plane separation, centralized control, adopting open-source standards, programmability, quality service (QoS) management, security. transformative impact network function virtualization (NFV) is explored with a perspective on its architecture applications in SDN, internet things (IoT), cloud computing, blockchain. security aspect SDN specific focus mitigating denial-of-service (DoS) attacks vulnerabilities associated open flow protocol also addressed. cognitive-inspired mechanisms adapt to evolving threats integrating machine learning (ML) artificial intelligence (AI) based algorithms for threat detection mitigation exemplified through case studies. Adoption software-defined perimeter, zero trust, blockchain, quantum-safe cryptography future discussed. Finally, IoT

Language: Английский

---

Ensuring IoT Security in 5G Era: Examining Protocols, Architectures, and Security Measures

Advances in Science, Technology & Innovation/Advances in science, technology & innovation, Journal Year: 2024, Volume and Issue: unknown, P. 135 - 145

Published: Jan. 1, 2024

Language: Английский

---

Increasing Security and Capacity in SDNFV Structures Via Edge Node Extension

Advances in Networks, Journal Year: 2024, Volume and Issue: 11(1), P. 8 - 16

Published: Dec. 16, 2024

The rapid proliferation of IoT devices and corresponding requirements for efficient data processing, Software-Defined Networking Function Virtualization (SDNFV) has come to be a key vehicle agile management network resources. advanced SDNFV model proposed in this study is intended resolve the two main challenges security scalability. sensitivity transmitted through networks as they grow size intricacy requires improved procedures hold ill-suited access their information ensure its integrity. Encryption & Authentication Protocols: Integration encryption authentication protocol together model, that secures streams against potential cyber threats threats, enhancing paradigm. Additionally, tackles scalability challenge by implementing multi-edge node support distributed processing better manage high volumes data. Such expansion especially notable since it solves latency issues bottlenecks so more resilient structure. current compares simulation results with existing models showcases amidst numerous architectures, suggested provides higher efficiency terms privacy capability. This latest development may fundamental future platforms are capable providing custom non-functioning backbone cope big today's ever-growing surrounding such devices.

Language: Английский

---