Research And Implementation Of BGP Instable Event Source-tracing System Based On Multi-perspective

With the rapid development of the Internet,its stable operation is strongly related to the country,individuals and industries.The Internet is composed of tens of thousands of autonomous systems.Their interconnection and inter-domain routing are achieved basicly by the Boundary Gateway Protocol.The stability of inter-domain routing plays an important role in the stable operation of the Internet.Instable inter-domain routing could disrupt the application service's connection,narrowen the available bandwidth,or block the traffic at the link entrance,thus seriously affecting the performance of the network and the upper service.There are many causes of instable routing events,including misconfiguration,equipment failure,network attack,etc.The rapid source-tracing of instable routing events is of great significance for maintaining the stable operation of the Internet.Inter-domain instable routing events are characterized by various abnormal mutations in the Routing Update Messages of Boundary Gateway Protocol.This study makes analysis of inter-domain instable routing events on both perspective of the monitoring point autonomous system and the source autonomous system.In the analysis,this study detects instable routing events by making the stability analysis based on the decrete wavelet transformation,and then sourcetraces the event's root cause by the method based on network topology building and connectivity variation calculating.The work of this study is constructed by three parts.First,in the data analysis and storage part,the real-time and historical data collection of the Routing Message is achieved.The database of the Routing Update Messages' raw data on the perspective of the monitoring point autonomous system is built.Also,a time series database of Routing Update Messages' time series data on the both perspectives is built.Second,in the event detection and sourcetracing part,the time series data is used to make event detection of the stability analysis based on the decrete wavelet transformation.Moreover,according to the abnormal time result of event detection,the event sourcetracing is achieved on the perspective of the source autonomous system,via detecting the affected source autonomous systems and calculating their connectivity variation in the related network topology.Finally,for the flexibility of the analysis,adjustment method for a various of system configurations is implemented.And the results of the analysis are well presented.Overall,the experimental analysis results show that the implemented system can sourcetrace the root cause of inter-domain instable routing events quickly and effectively.This theis provides data support for future network failure recovery caused by inter-domain instable routing events.