Genetic Programming for Resource Allocation in Network Function Virtualization
Nguyen Thi Tam, Do Duc Anh, Tran Huy Hung, Pham Van Hanh, Huỳnh Thị Thanh Bình, Lê Trọng Vĩnh · 2023
Network function virtualization is a promising ar-chitecture for replacing dedicated hardware middle boxes with adaptable software, commonly called virtual network functions. A service function chain, which is made up of a set of ordered virtual network functions, can serve as the network function virtualization equivalent of a service. Allocating resources at servers is one of the most challenging problems in network function virtualization because of resource restrictions and the increasing number of services/requests. This paper considers the path planning for service function chain requests in network function virtualization to maximize the number of accepted requests. We formulate the problem as a Mixed Integer Linear Programming problem to find the optimal solution. However, the formulated Mixed Integer Linear Programming becomes complex to solve with increasing decision variables and constraints with increased network size. Since the problem is NP-hard, we propose the genetic programming-based algorithm to obtain near-optimal solutions to the problem efficiently. The links' bandwidth and servers' resources are considered during the service function chain routing. Simulation results show that the proposed solution is very close to the optimum and outperforms existing works concerning the service acceptance ratio.