Fault-Tolerance in Future RAN: Enabling Redundancy for Efficient Recovery of RAN Network Functions

In recent years, the reliability and resilience of communication networks and cellular networks in particular has become of increasing importance. For cellular networks, disaggregation and virtualization of RAN network functions offer the opportunity to improve the overall resilience of future networks. That is, improving the resilience of individual network functions is much less complex compared to improving the resilience of the overall communication network. Overall, the resilience of individual network functions increases resilience of the communication network. In this paper, we propose a mechanism to enable redundant network functions in the Radio Access Network (RAN) to reduce its total downtime in case of disruptions or failures. If a network function is disrupted, the redundant network function in warm standby is activated to recover the functionality of the disrupted network function. Different Radio Intelligent Controller (RIC) services are used to monitor RAN components and control the efficient recovery of disrupted network functions. We show that the proposed mechanism significantly reduces the overall down-time of the RAN compared to restarting a disrupted network function. Our prototypical implementation on the example of the Distributed Unit (DU) for replacing a disrupted network function in warm standby reduces the downtime by 76.86% (median 41.13%) compared to restarting the disrupted network function.