期刊
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
卷 18, 期 6, 页码 4059-4068出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TII.2021.3088938
关键词
Industrial Internet of Things; Training; Computer crime; Protocols; Botnet; Servers; Data models; Cybersecurity; federated learning (FL); fog; edge; gated recurrent unit (GRU); industrial IoT (IIoT) distributed denial of service (DDoS); iterative model averaging (IMA)
类别
资金
- National Natural Science Foundation of China [U1804263, 62072109, 62072356]
- Australia Research Council DECRA [DE210101458]
- Key Research and Development Program of Shaanxi [2019ZDLGY12-08]
This article proposes a protocol that combines federated learning and fog/edge computing to combat malicious codes by training a globally optimized model on distributed datasets. The FL-based detection protocol maximizes the value of distributed data samples, improving detection accuracy, reducing mitigation time, and increasing attackers' cost within a defense alliance. Comprehensive evaluations show a 2.7 times increase in cost, a 72% decrease in mitigation response time, and a 47% increase in accuracy on average, with a detection accuracy of approximately 98% in FL.
Due to resource constraints and working surroundings, many IIoT nodes are easily hacked and turn into zombies from which to launch attacks. It is challenging to detect such networked zombies rooted behind the Internet for any individual defender. In this article, we combine federated learning (FL) and fog/edge computing to combat malicious codes. Our protocol trains a global optimized model based on distributed datasets of collaborators while removing the data and communication constraints. The FL-based detection protocol maximizes the values of distributed data samples, resulting in an accurate model timely. On top of the protocol, we place mitigation intelligence in a distributed and collaborative manner. Our approach improves accuracy, eliminates mitigation time, and enlarges attackers' expense within a defense alliance. Comprehensive evaluations confirm that the cost incurred is 2.7 times larger, the mitigation response time is 72% lower, and the accuracy is 47% higher on average. Besides, the protocol evaluation shows the detection accuracy is approximately 98% in the FL, which is almost the same as centralized training.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据