Survivable Virtual Network Mapping Based on Two-Stage Potential Games for Cloud Infrastructure

We study the survivable virtual network mapping (SVNM) problem for allocating virtual machines (VMs) from multiple data centers (DCs) with the objective of minimizing the total cost of networking and computing resources to tolerate any single link failure under two constraints: the computing capacity constraint at each DC and the networking capacity constraint at each link. We first describe graph models of SVNM, formulate the SVNM problem, and prove that SVNM is NP-complete. We then formulate the problem as an integer linear programming (ILP) and give results for small-scale cases. A heuristic approach, named Virtual Star Mapping First (VSMF), is proposed with two consecutive steps: 1) partitioning a virtual network to a set of virtual stars; and 2) allocating stars to the cloud through a two-stage capacity-constrained potential game which is proved to be convergent to a pure Nash Equilibrium. Numerical results show that VSMF achieves low cost, which is close to the optimal solution, in both small-scale and large-scale cases.