Scholarlink

With the increase in data in today’s world comes the problem of data vulnerability. Denial of Service (DoS) has become a very common form of attack, as also Distributed Denial of Service (DDoS), where the target is overwhelmed by a flood. They waste a lot of precious resources, so federated learning is used as a reliable approach to improving the reliability of DDoS attack detection systems. So, in order to increase data security and remove the single point failure in the federated learning process, the conventional federated learning process, which is implemented using a feedforward neural network, is integrated with blockchain technology in our work. The performance of the model is evaluated on various learning rates and number of clients and variations are analyzed to reach to the optimal accuracy. To accomplish this, we try to study and improve the safety and efficiency of DDoS attack detection in a CIC-DDoS 2019 dataset using federated learning and blockchain.

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

DDoS Attack Detection Using Federated Learning Integrated with Blockchain

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Rajashree Konwar,
  • Ayush Utkarsh,
  • Shreyoshi Chakraborty,
  • Pavan Kumar Kannojiya,
  • Indrajit Banerjee

摘要

With the increase in data in today’s world comes the problem of data vulnerability. Denial of Service (DoS) has become a very common form of attack, as also Distributed Denial of Service (DDoS), where the target is overwhelmed by a flood. They waste a lot of precious resources, so federated learning is used as a reliable approach to improving the reliability of DDoS attack detection systems. So, in order to increase data security and remove the single point failure in the federated learning process, the conventional federated learning process, which is implemented using a feedforward neural network, is integrated with blockchain technology in our work. The performance of the model is evaluated on various learning rates and number of clients and variations are analyzed to reach to the optimal accuracy. To accomplish this, we try to study and improve the safety and efficiency of DDoS attack detection in a CIC-DDoS 2019 dataset using federated learning and blockchain.