Bi-Level Programming Model and Algorithm for VNF Deployment With Data Centers Placement

Hejun Xuan

doi:10.1109/ACCESS.2019.2960395

Return

Bi-Level Programming Model and Algorithm for VNF Deployment With Data Centers Placement

2019-01-01

DOI：10.1109/ACCESS.2019.2960395

Original

Save

View PDF

Abstract

En 中文

Virtual network function (VNF) can provide various network services and is widely deployed in inter-data centers elastic optical networks (Inter-DC EONs). Routing and VNF deployment for VNF service chain (VNF-SC) in Inter-DC EONs is a very important and well-known NP-hard problem. For this problem, if determining the number and locations of data centers is additionally considered, it will be more complicated. In this paper, we investigate a network planning problem in Inter-DC EONs by determining all these factors, i.e, by determining not only the optimal routing and the optimal VNF deployment for VNF-SCs, but also the optimal number and locations of data centers. To achieve this purpose, we first establish a bi-level programming model in which the leaders objective is to minimize the number of data centers and find the best locations of data centers so that we can get a balanced VNF deployment on data centers. To determine the optimal routing and VNF deployment for VNF-SCs, the followers objective is to minimize the maximum index of used frequency slots and the number of used frequency slots. Then, to solve the proposed model effectively, tailor-made crossover, mutation and local search operators are designed, and based on these operators, an efficient bi-level hybrid memetic algorithm (BiHMA) is proposed. Finally, to test the effectiveness of the proposed model and the efficiency of the proposed algorithm, the simulation experiments are conducted on two widely used networks, and experimental results indicate that the proposed algorithm has a higher efficiency than compared algorithms.

Keywords:

EONs

data centers placement

bi-level optimization

Memetic algorithm

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.