Scholarlink

<p>This study introduces BDML-IDHIS, a hybrid intrusion detection system combining artificial neural networks (ANN) and Honeypot intelligence, implemented on Apache Flink for real-time big data processing. The system employs a Message Queuing Telemetry Transport (MQTT) Honeypot integrated with Decision and Redirection Engines to enhance system security Experimental evaluations demonstrate that the proposed model achieves a classification accuracy of 98.09%, significantly outperforming traditional methods such as Support Vector Machine (92.76%) and Random Forest (89.40%). Furthermore, the system’s scalability and real-time processing capabilities are validated under varying data sizes, showcasing superior throughput and latency performance compared to Apache Spark-based systems. However, limitations include the computational overhead associated with ANN training and reliance on pre-collected datasets. The study highlights the strengths of the BDML-IDHIS system, including precise attack filtering, real-time processing, and scalability for big data environments. Future work will focus on incorporating feature selection techniques to enhance model efficiency and reduce computational complexity.</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

Big data with machine learning enabled intrusion detection with honeypot intelligence system on apache Flink (BDML-IDHIS)

  • 期刊论文

  • 发表日期：

    2025年1月7日

查看全文
  • Akshay Mudgal,
  • Shaveta Bhatia

摘要

This study introduces BDML-IDHIS, a hybrid intrusion detection system combining artificial neural networks (ANN) and Honeypot intelligence, implemented on Apache Flink for real-time big data processing. The system employs a Message Queuing Telemetry Transport (MQTT) Honeypot integrated with Decision and Redirection Engines to enhance system security Experimental evaluations demonstrate that the proposed model achieves a classification accuracy of 98.09%, significantly outperforming traditional methods such as Support Vector Machine (92.76%) and Random Forest (89.40%). Furthermore, the system’s scalability and real-time processing capabilities are validated under varying data sizes, showcasing superior throughput and latency performance compared to Apache Spark-based systems. However, limitations include the computational overhead associated with ANN training and reliance on pre-collected datasets. The study highlights the strengths of the BDML-IDHIS system, including precise attack filtering, real-time processing, and scalability for big data environments. Future work will focus on incorporating feature selection techniques to enhance model efficiency and reduce computational complexity.