Resource consumption and security-aware multi-tenant service function chain deployment based on hypergraph matching

Abstract

Network function virtualization (NFV) technology enables cloud service providers (CSPs) to deploy their service function chains (SFCs) into the common shared physical infrastructures to operate the network or cloud service. However, designing a suitable placement scheme for multi-tenant SFCs to balance resource consumption with performance and security is a tricky issue. In this paper, we first proposed security constraints based on physical isolation and formulated the problem of reasonably placing multi-tenant SFCs subject to multiple constraints to minimize resource consumption. The proposed problem is more intractable than previous works because of the extra introduced security constraints. To solve it, a new non-uniform weighted hypergraph construction approach is proposed by defining hyperedge weights to represent the resource consumption on one SFC. A hypergraph matching algorithm is proposed to find the maximum weight subset of hypergraphs whose vertices do not intersect to give a high degree of physical isolation by dealing with the conflict graphs . Simulation results show that the proposed hypergraph matching algorithm outperforms the comparison schemes in terms of the resource consumption while satisfying the security requirements.