A scheme for adaptive self-diagnosis of QoS degradation in future networks

The capability of a network to identify problematic situations, named self-diagnosis, enables it to react promptly and autonomously once an event or error has been identified. The use of service information in this process enables it to identify more composite problems and to act more targeted in order to solve complex errors. This paper proposes a novel fuzzy logic-based self-diagnosis mechanism for identifying Quality of Service (QoS) degradation events. Furthermore, we introduce a framework for the adaptation of the self-diagnosis scheme, which enables the network elements to evolve the way they interpret the context information. The adaptation scheme is based on the statistical analysis of the measurements and reacts accordingly without requiring any external human intervention. The adaptive self-diagnosis scheme has been evaluated through simulations in order to showcase the benefits from its application in IP networks for the VoIP service. The simulation results show that the adaptive self-diagnosis scheme performs very well compared to existing solutions, increases significantly the event detection rate and, as a result, the capability of controlling the QoS on top of the involved network elements.