Cognitive Network Fault Management Approach for Improving Resilience in 5G Networks

Resilience is one of the fundamental requirements of critical communication services such as ultra-reliable low latency (URLLC) services offered by 5G networks. In order to support the communication service, the 5G networks can take different approaches for deployment of network functions, i.e. the network functions can run on virtualized infrastructure (telco cloud) as well as on the specialized physical hardware instances (e.g. RAN functions). Irrespective of the deployment approach taken the adequate level of resilience needs to be supported on all parts of the network in order to achieve required level of service resilience. In this work, we aim at improving the resilience level of communication services by applying network fault management techniques specialized for 5G slicing-enabled networks taking jointly into account the aspects of virtualized and physical infrastructure. We describe the novel approach of designing flexible and cognitive fault management functions that can dynamically adapt their behavior based on the actual network slice requirements and current network context. We highlight the benefits of such an approach in achieving the
required level of resilience especially addressing the telco cloud domain.