Virtual Network Function Scheduling: A Matching Game Approach

Network function virtualization is a promising technique for telecom providers to efficiently manage network services at low cost. However, existing works mainly focus on resource allocation and thus leave behind an important issue: the virtual network function (VNF) scheduling. Current approaches, e.g., round-robin scheduling or heuristic algorithms, still expose some unsolved issues, such as high computational cost and inability to perform online scheduling. In this letter, we propose a matching-based algorithm to solve the NP-hard VNF scheduling problem. This approach can guarantee a stable scheduling, in which all network services are satisfied with the assignment. Finally, the effectiveness of our method is verified through numerical evaluation, showing that our approach can increase the number of completed VNFs by 36.8% compared with the current round-robin method.