This paper describes a Node Profiler. A Node profiler is a Machine Learning based approach to profile Nodes based on underlying KPIs (Key Performance Indicator). It also explains how such profiling procedures can be employed to automate and accelerate Network Optimization tasks. Relying on Data Analytics and Machine Learning techniques, this process can help getting insights and assess underlying (common) Root Causes thus becoming an important part of a Network Engineer’s toolkit.
Due to an exponential increase in mobile telephony, the wireless ecosystem is in a constant state of flux.
Software patches, hardware swap-outs, new chipsets etc. continuously get applied to the system to cope with this increase. As a result, the KPIs/ health of the wireless system is under constant surveillance and any degradation in network performance requires quick analysis and recovery. Since hundreds of thousands of subscribers could be affected with such an operation (be it big or small), the analysis and recovery is generally done by OEMs, closely monitored by the TSPs. It is in this context that the current paper attempts to add value to both TSPs and OEMs.