Cloud Network Convergence Disaster Recovery Approach for Reliable Virtual Network Function

The evolution of Network Function Virtualization (NFV) has transformed telecommunications by shifting from Physical Network Functions (PNFs) to Virtual Network Functions (VNFs), offering improved flexibility, scalability, and resource efficiency. However, this transition has also introduced new challenges, particularly vulnerabilities stemming from Cloud Operating System failures. Traditional NFV reliability measures, which focus on resource allocation and fault recovery, have struggled to bridge the reliability gap between cloud infrastructure and network services. Addressing these limitations, this paper proposes a novel disaster recovery framework that integrates cloud technologies, including multi-region and multi-Availability Zone (AZ) architectures, to significantly enhance VNF reliability. The framework is designed to leverage the fault tolerance and redundancy of cloud architectures for improved disaster recovery performance in VNFs. We evaluate its feasibility and effectiveness through a comparative analysis of disaster recovery performance between traditional NFV setups and the proposed framework. Experimental results show that the 3AZ architecture improves Recovery Time Objective (RTO) through fully automated recovery processes and a fault-isolated design, while the 2AZ with Arbiter setup also achieves notable RTO improvements.