Towards Cognitive Autonomous Networks

9Cognitive Autonomy for Network Self‐Healing

Janne Ali‐Tolppa¹, Marton Kajo², Borislava Gajic¹, Ilaria Malanchini³, Benedek Schultz⁴, and Qi Liao³

¹Nokia Bell Labs, Munich, Germany

²Technical University of Munich, Munich, Germany

³Nokia Bell Labs, Stuttgart, Germany

⁴Bosch, Budapest, Hungary

While self‐optimization functions optimize a set of configuration parameters to improve the network performance from a given network state, the self‐healing functions focus on ensuring that the network can fulfil its purpose and serve its customers even in case of failures, unexpected changes, or events that risk a degradation in the network performance [1]. In other words, their objective is to make the network more resilient [2–5].

This chapter discusses the methods for improving resiliency of future cognitive autonomous networks. In particular, concepts for cognitive self‐healing functions that are designed to detect or even predict network performance degradations and apply corrective actions are presented. Unlike in optimization, the self‐healing functions can only monitor the symptoms and the root causes are not known a priori. Due to this and the less‐constrained problem space, the self‐healing process is often more complex than the optimization control loops. This complexity is accentuated by the distributed and heterogeneous nature of mobile networks. Finally, diverse fault states often occur only in very rare cases, which makes it difficult to collect statistically meaningful ...