Anomaly Detection in Cloud Computing Workloads based on Re- source Usage

Abstract In recent years, the widespread adoption of cloud computing for data storage and processing has underscored the critical importance of security within cloud computing servers. Adherence to cloud service regulations mandates that all aspects, from resource procurement to data backup and security, are the responsibility of the service provider. Consequently, providers are actively seeking methods to detect and mitigate attacks promptly to safeguard their interests and users. However, the effectiveness of existing attack detection methods varies depending on the attack type, with known attacks being more easily identified and prevented through artificial intelligence-based techniques. Yet, even minor alterations in attack characteristics pose challenges to detection, highlighting an ongoing challenge in the field. Additionally , prior methods often rely on a fixed number of features during both learning and operational phases, limiting their adaptability.To address these challenges, this research proposes Pattern-AD, a novel anomaly detection method that models attacks as anomalies against the system’s normal states. Unlike traditional approaches, Pattern-AD extracts frequent patterns using the Apriori data mining algorithm, offering flexibility regardless of pattern length. Comparative analysis against established anomaly detection methods demonstrates Pattern-AD’s superior accuracy and reduced error rates, showcasing its potential as a reliable security solution for cloud computing environments.