Mechanism of context-aware and adaptive data collection, aggregation and analysis

Abstract

With the arrival of 5G, networks will increase in complexity, heterogeneity and scale. Further, 5G is expected to cater to a diverse use cases with widely varying performance requirements with the help of network slicing - extending beyond enhanced Mobile Broadband (eMBB), massive Machine Type Communication (mMTC) and Ultra-reliable Low Latency Communication (URLLC) use cases. Given this context, it is essential to ensure that service KPIs (Key Performance Indicators) are adhered to through timely detection of fault or performance issues to enable resolution in accordance with the use case's requirements. The monitoring framework should thus enable timely detection of fault or performance issues based on service/slice KPIs under dynamic network conditions, taking into consideration user characteristics, location and other relevant criteria. This paper presents a context-aware and adaptive network data collection, aggregation and analysis framework for effective fault or performance anomaly-detection based on context (location, user and service/slice characteristics, environment factors), dynamic network conditions and service/slice target KPIs. An open-source based implementation of the proposed solution is also presented.