A GRASP-based approach for Dynamic and Reliable Virtual Network Embedding
Amine Rguez, Yassine Hadjadj‐Aoul, Gerardo Rubino · 2025
Ensuring reliable placements of network functions is critical in modern infrastructure, especially in Telco-cloud environments where high availability of services is paramount. In fact, the massive hardware infrastructure required for the emerging 6G and the multiplication of softwarized services to be deployed in the edge-cloud continuum have exacerbated failure concerns, underscoring the importance of addressing reliability challenges.In this context, we propose a new approach for providing reliable placement of network slices dynamically. This approach leverages the Greedy Adaptive Search Procedure (GRASP), a metaheuristic algorithm that combines greedy randomized construction with local search optimization. We compared the results of our approach against some traditional ones in terms of (1) the acceptance rate when reliability serves as a hard constraint and (2) the placement reliability values when reliability is an optimization objective. Our results demonstrate that our GRASP-based algorithm achieves up to 33% improvement in terms of reliability compared to baseline approaches.