VCG-based Bandwidth Allocation Scheme for Network Virtualization

Network virtualization has been proposed as a powerful and flexible way to support multiple heterogeneous architectures on a shared infrastructure. A major challenge in this respect lies in the resource allocation problem that allocates the resources of Infrastructure Providers (InPs) among Service Providers (SPs) fairly and efficiently. Because of the complexity in the interaction between InPs and SPs, this problem is more complicated and needs more investigation. This paper introduces an InP-SP-End user (ISE) system model in network virtualization environment, in which we present a bandwidth allocation scheme based on VCG (Vickrey-Clarke-Groves) to maximize the total revenue of SPs by inhibiting selfish action of SPs, and design a Q-learning strategy algorithm in order to obtain optimal bidding strategies for SPs. Simulations and experimental studies demonstrate the effectiveness and the fairness of the proposed scheme, and the convergence of the proposed algorithm.