Operational Fault Detection in cellular wireless base-stations

The goal of this work is to improve availability of operational base-stations in a wireless mobile network through non-intrusive fault detection methods. Since revenue is generated only when actual customer calls are processed, we develop a scheme to minimize revenue loss by monitoring real-time mobile user call processing activity. The mobile user call load profile experienced by a base-station displays a highly non-stationary temporal behavior with time-of-day, day-of-the-week and time-of-year variations. In addition, the geographic location also impacts the traffic profile, making each base-station have its own unique traffic patterns. A hierarchical base-station fault monitoring and detection scheme has been implemented in an IS-95 CDMA Cellular network that can detect faults at - base station level, sector level, carrier level, and channel level. A statistical hypothesis test framework, based on a combination of parametric, semi-parametric and non-parametric test statistics are defined for determining faults. The fault or alarm thresholds are determined by learning expected deviations during a training phase. Additionally, fault thresholds have to adapt to spatial and temporal mobile traffic patterns that slowly changes with seasonal traffic drifts over time and increasing penetration of mobile user density. Feedback mechanisms are provided for threshold adaptation and self-management, which includes automatic recovery actions and software reconfiguration. We call this method, Operational Fault Detection (OFD). We describe the operation of a few select features from a large family of OFD features in Base Stations; summarize the algorithms, their performance and comment on future work.

[1]  Chuanyi Ji,et al.  Proactive network fault detection , 1997, Proceedings of INFOCOM '97.

[2]  Malgorzata Steinder,et al.  Probabilistic fault localization in communication systems using belief networks , 2004, IEEE/ACM Transactions on Networking.

[3]  Mahamat Guiagoussou,et al.  Implementation of a diagnostic and troubleshooting multi-agent system for cellular networks , 1999, Int. J. Netw. Manag..

[4]  Salah Aidarous,et al.  Telecommunications Network Management into the 21st Century: Techniques, Standards, Technologies, and Applications , 1994 .

[5]  Klaus Jobmann,et al.  Automatic detection of abnormal deviations from network performance data values , 2002, NOMS 2002. IEEE/IFIP Network Operations and Management Symposium. ' Management Solutions for the New Communications World'(Cat. No.02CH37327).

[6]  Benjamin Cheung,et al.  Statistical algorithms in fault detection and prediction: Toward a healthier network , 2005, Bell Labs Technical Journal.

[7]  Michael T. Goodrich,et al.  Algorithm Design: Foundations, Analysis, and Internet Examples , 2001 .

[8]  C. S. Hood,et al.  Proactive network-fault detection [telecommunications] , 1997 .

[9]  Marina Thottan,et al.  Proactive anomaly detection using distributed intelligent agents , 1998, IEEE Netw..

[10]  Donald S. Jackson,et al.  Calculation of system availability using traffic statistics , 2003, Bell Labs Technical Journal.

[11]  L. Lawrence Ho,et al.  Real-time performance monitoring and anomaly detection in the Internet: An adaptive, objective-driven, mix-and-match approach , 1999, Bell Labs Technical Journal.

[12]  Benito R. Fernandez,et al.  A neural network based adaptive fault detection scheme , 1995, Proceedings of 1995 American Control Conference - ACC'95.

[13]  Samuel C. Yang,et al.  CDMA RF System Engineering , 1998 .