Analysis of Telecom Service Operation Behavior with Time Series

Abstract

Operation of complex telecom services is a field that mixes technology, processes and teams. Despite the existence of detailed protocols and automation, the real behavior is hard to measure and predict. The human factor is a source of uncertainty, and this fact is of special relevance when facing stressful situations. Informal team working culture, time shifts or external pressure are main sources of change. In this research we use time series analysis as a statistical proxy to detect this kind of drift in teams that solve network failures if three live services: IPTV, Cloud Infrastructure and IoT. This task is known as incident management. This would provide not only a numerical evidence of the uncertainty in troubleshooting of digital services but also an assessment about the economic and operational impact of service releases. Changes in best fitting models may reflect different informal work cultures among the operation teams.