Fault analysis and prediction in telecommunication access network

Abstract

Technological innovations and growing consumer demand have led to a variety of design and exploitation problems in telecommunication networks. In particular, local access networks have received a lot of attention, since they account for approximately 60% of total investments in communication facilities. Hence even marginal improvements in the access network design and maintenance can lead to substantial savings. Faults in a telecommunication network have direct impact on its availability and maintenance costs, so their quick elimination, prevention and removal of causes that generated them, is of special interest. The possibility to analyse the spatial distribution of faults and to predict places where future errors may arise can significantly help telecom operators who are in charge to detect and repair such problems. This data is also crucial for future planning and design of telecommunication networks as it can help to avoid problematic areas or use additional measures for protection of telecommunication cables and equipment. Existing statistical data (fault data and foreign factors data like precipitation, high-voltage installations etc.) is analysed and then a model of faults in telecommunication network is created using Bayesian methodology. The whole procedure of acquisition and interpretation of data is described, and an application called FATAN (fault analysis in telecom access network) that aggregates all procedures is created. The achieved result gives us a possibility to successfully determine the locations where it is likely that an error will arise.