Optimal Distributed Resource Allocation in 5G Virtualized Networks

The concepts of network function virtualization (NFV) and end-to-end (E2E) network slicing are two promising technologies empowering 5G networks for efficient, flexible and dynamic network deployment and service management. Optimal resource allocation is one of the challenging problems to address in such networks. In this paper, we propose a resource allocation model for 5G virtualized networks in a heterogeneous cloud infrastructure. In our model, each network slice has a resource demand vector for each of its building virtual network functions (VNFs). We then formulate the optimal resource allocation as a convex optimization problem maximizing the overall system utility as a function of the slice thicknesses with the constraints of the data centers' resource capacities. The slice thickness variables together with the demand vectors determine the amount of resources allocated to each slice. We further propose a distributed solution for the resource allocation problem based on auction/game theory by forming a resource auction between the slices and the data centers (DCs). It is shown that the resource allocation game has a unique Nash equilibrium and its solution is the same as the solution of the centralized system optimization problem, i.e., in equilibrium the slice thicknesses maximize the overall system utility. Numerical analysis are provided to show the validity of the results, evaluate the convergence of the distributed solution and also comparing the performance of the optimal scheme with heuristic ones.