Network failure detection based on correlation data analysis

For a telecommunication operator, the effective detection of access and aggregation network failures is the key to providing continuous service. Although there are many monitoring systems on the market, analysis has shown that there is no possibility of automatically detecting all failures using standard monitoring systems. In this article, an innovative option for failure detection is proposed, based on correlation analysis of data retrieved in real time from the network. A key source of data is the Remote Authentication Dial-in User Service (RADIUS) which records events giving information about a user's Point-to-Point Protocol (PPP) session state sent to the monitoring system. It is shown here that the proposed solution enables an operator to detect all network events affecting customers. The detection of a greater number of events enables the operator to react quickly to them and to restore services to users as soon as possible, which ultimately improves the quality and continuity of provided services.