Scholarlink

<p>The continued emergence of large amounts of heterogeneous sensor and system logs is due to the rapid deployment of Internet of Things (IoT)-based Wireless Sensor Networks (WSNs). The security, reliability, and effectiveness of storage of such information are still a big challenge because of resource limitation, dynamic network environment, and rising security threats. Current solutions to log management and blockchain-based storage scale have been noted to have high computation cost, low scalability, and poor trust security, limiting their use in large-scale IoT. In order to overcome these problems a safe and effective data storage system is introduced that combines smart validation of logs based on the blockchain technology. The mesh normal filtering with shape awareness is utilized to improve the quality of data and a hybrid context axial reverse attention network incorporating a dual-channel convolutional neural network (hyb-CARAN-DCCNN) is applied to learn and validate features successfully with the help of a Builder Optimization Algorithm (BOA). Moreover, the low-latency, tamper-resistant data storage is ensured with the help of a dynamic and trusted blockchain consensus mechanism. The analysis of experimental evaluation results with the TON-IoT dataset shows that it has higher security and accuracy, in addition to a more efficient storage compared to traditional methods. The proposed framework achieves 99.9% accuracy with a block generation time of 1.6&#xa0;s, consensus latency of 210&#xa0;ms, and an immutability rate of 99.9% on the TON-IoT dataset.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Secure and Efficient Block Chain-Enabled Log Storage Framework For IoT-Based Wireless Sensor Networks Using a Hybrid Context Axial Reverse Attention Network with a Dual-Channel Convolutional Neural Network

  • 期刊论文

  • 发表日期：

    2026年5月26日

查看全文
  • J. M. Hamer Shield,
  • K. Suresh

摘要

The continued emergence of large amounts of heterogeneous sensor and system logs is due to the rapid deployment of Internet of Things (IoT)-based Wireless Sensor Networks (WSNs). The security, reliability, and effectiveness of storage of such information are still a big challenge because of resource limitation, dynamic network environment, and rising security threats. Current solutions to log management and blockchain-based storage scale have been noted to have high computation cost, low scalability, and poor trust security, limiting their use in large-scale IoT. In order to overcome these problems a safe and effective data storage system is introduced that combines smart validation of logs based on the blockchain technology. The mesh normal filtering with shape awareness is utilized to improve the quality of data and a hybrid context axial reverse attention network incorporating a dual-channel convolutional neural network (hyb-CARAN-DCCNN) is applied to learn and validate features successfully with the help of a Builder Optimization Algorithm (BOA). Moreover, the low-latency, tamper-resistant data storage is ensured with the help of a dynamic and trusted blockchain consensus mechanism. The analysis of experimental evaluation results with the TON-IoT dataset shows that it has higher security and accuracy, in addition to a more efficient storage compared to traditional methods. The proposed framework achieves 99.9% accuracy with a block generation time of 1.6 s, consensus latency of 210 ms, and an immutability rate of 99.9% on the TON-IoT dataset.