TTERCL: An onSite Real-time Alarm Root-Cause Location Algorithm

Ding, Jianbing; Wang, Xidong; Xiao-zhou, YE; Ouyang, Ye; Chai, Yuanyuan

doi:10.1109/bigdata52589.2021.9672017

Cited by 2 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Li, T. et al propose a novel association rules mining based alarm correlation analysis system to find interesting association rules between alarm events [5]. The online real-time alarm RCL algorithm proposed can fully explore the relationship between alarms in the dimension of time and space, and achieve online alarm data compression in [6]. Although root cause analysis have been well studied for some complicated system [7], most of them merely aim at finding the root cause alarms, and studying in the wireless network [8], access network [9], IP network [10] and cloud [11] areas.…”

Section: Introductionmentioning

confidence: 99%

Fault root cause location based on machine learning in 5G core network

Jiang,

Yu,

Yan

et al. 2023

Third International Conference on Green Communication, Network, and Internet of Things (CNIoT 2023)

View full text Add to dashboard Cite

With the development of the virtualization of 5G core (5GC) network, the network architecture is more complex. Once faults occur, the fault root cause location (RCL) is difficult. Traditional operation and maintenance (O&M) cannot satisfy the requirements of RCL based on manual experience. This paper proposes the k-nearest neighbor and Bagging (kNN-Bagging) scheme for locating the root cause of 5GC network faults. It first concludes the fault root causes from the historical fault orders, and extracts the fault features based on the related alarm databases. Then, it conducts the pre-processing of the alarm data in order to better understand and the input data for machine learning (ML) algorithms. At last, k-nearest neighbor (kNN), Naive Bayes (NB), Support Vector Machines (SVM), kNN-NB-SVM (KNS), SVM-Bagging and kNN-Bagging algorithms are adopted to establish the fault RCL models. Experimental results show that the average accuracy of fault RCL of kNN-Bagging algorithm is optimal value, which are 92.6% and 91.6% respectively in the offline and online environment of the real-world 5GC network. The proposed approach can effectively shorten the fault processing time and improve the O&M efficiency of the 5GC network.

show abstract