Fault diagnosis in optical access network using Bayesian network

Due to ever rising need for higher bandwidth dictated by the increasing amounts and quality of services (Asymmetric Digital Subscriber Line (ADSL), Internet Protocol television (IPTV) and Voice over Internet Protocol (VOIP)), and the global development of technology, Fiber To The Home (FTTH) is becoming an affordable answer for the service provider. Before fully launching the service for the mass market the most important prerequisite is being able to provide customers with a high quality experience. To be able to achieve that, one of the main factors that service providers have to ensure is accurate and precise diagnosis of customer reported faults. Low quality of diagnostics prolongs the period of service degradation and increases the likelihood of repeated faults which results in dissatisfaction amongst customers. That is why our objective is to improve the accuracy of diagnostics performed by network technicians to at least 80%. In this paper we are introducing a solution based on Bayesian network and presenting the results of the applied method.