A DAG-free Method to Event Correlation for Revealing Abnormal Connections of Cloud Instances. Issue 1 (1st July 2022)

• Record Type: Journal Article
• Authors: Li, Xiaojian; Hu, Hangping; Jiang, Yiyi; Zhang, Zhen; Yang, Haopeng; Liu, Liqin
• Is Part Of: Journal of physics. Volume 2303:Issue 1(2022)
• Publication Date: 2022-07-01
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/2303/1/012083
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: We proposes an event correlation method with rule evaluation(ECRE) to reveal abnormal connections from cloud Instances. We combine some characteristics of status-pair to categorize all abnormal status-connections as far as possible. Our method can classify abnormal connections of instances but doesn't require the DAG of tasks. Those labels that unmatched the supervision rules are removed to remain more and better samples for training. The results that our method achieved show high accuracy, and a significant correlation between the accuracy and the usage count of supervision-rules.