An error messages clustering-based fault management framework for adaptive home network middleware

This paper proposes an autonomous fault management framework consists of efficient fault detection, analysis and fault recovery based on clustering error messages. Most home network middlewares provide simple rule-based fault processing mechanisms optimized only to handle error messages individually. The proposed architecture focuses on detecting and analyzing error messages generated by identical faults as collectively as possible. For this, three relation graphs are used to decide the fault range and to find error messages. Message filters are also used to omit analyzing the previously found faults. Predefined fault recovery rules are used to handling the analyzed results and reconfigure home networks automatically. The proposed framework is implemented on an adaptive home network middleware based on the Universal Middleware Bridge (UMB) that provides adaptive autonomic configuration for the heterogeneous home networks.