키워드 기반 라벨 인코딩을 수행한 IDS와 디셉션 기법을 통합한 포그 컴퓨팅 환경 보안 아키텍처 = Security Architecture for Fog Computing Environments Combining IDS with Keyword-Based Label Encoding and Deception Techniques

다국어 초록 (Multilingual Abstract)

nsion model of cloud computing is gaining attention and is being widely used in large-scale IoT-based distributed environments with many research efforts underway. However, these environments face increasing cybersecurity challenges due to vulnerabilities in networks and edge devices. To address these issues, this study presents a security framework integrating a supervised learning-based intrusion detection system trained on the TON_IoT dataset with deception techniques. The IDS performance was improved by performing keyword-based preprocessing during label encoding of specific columns and including simulation data in training. And to select an IDS model suitable for the fog computing environment, we analyzed the inference time and proposed a redirection method to reduce latency. However, overfitting of MitM data, limitations in detecting zero-day attacks, and further validation of redirection methods are challenges for future works.