A Hierarchical Tree-Based Syslog Clustering Scheme for Network Diagnosis
S. Siva Sankar Rao, Minghui Wang, Cuixia Tian, Xinan Yang, Xiangqiao Ao · 2021
With the continuous development of Information Technology, modern networks have been widely utilised. Since the complex network structure causes growing difficulties in maintenance, log analysis has been widely studied in recent years for network diagnosis. System log clustering is mainly focused for root cause analysis. In this paper, a hierarchical tree-based clustering scheme is proposed that could accurately group system logs according to both time and network constraints without any training and parameter settings. Furthermore, it largely accelerates the matching process by reducing matching times and significantly boosts the performance of hit rate (100%) and match efficiency (16%) comparing to other clustering strategies, which greatly helps with precise network diagnosis.