The Telco community has foreseen fully autonomous networks able to perform tasks, handle problems, and be capable of re-configuring themselves without requiring human intervention. To achieve this vision, Elisa Automate is developing an automation system for RAN fault management. In the context of the product they are developing, I propose (i) a system to automate the estimation of the time required by humans to handle alarms to accelerate the proof of concept with new customers, and (ii) an approach that leverages machine learning to automatically extract the counters from the Performance Management data to estimate the impact of alarms on network performance and consequently on the end-users. The value creation chain of fault management automation with its managerial implications concludes the research question.
The system (i) has shown promising results and little more work, following the suggestions provided, will allow having a more mature functionality ready to be tested in real scenarios. On the other hand, system (ii) has had a more challenging development due to several problems, mainly the lack of data for the experiments. Furthermore, the final approach has not yet been identified and will require further research to find a workable conclusion.
Together, systems (i) and (ii) will make it possible to monitor the direct and indirect effects of automating fault management, the manual labor hours saved, and the network performance increase, respectively.
