弹性光数据中心网络中收益最大化的带宽分配算法

弹性光数据中心网络中的带宽分配需要兼顾云服务提供商收益和用户连接请求性能,对此提出收益最大化的带宽分配算法,其以子载波为基本单位将光纤带宽有效地分配给多个用户连接请求。首先,基于收益模型和虚拟子载波的概念建立带宽分配问题的整数线性规划模型;接着提出一种基于动态规划的带宽分配算法以及一种启发式的带宽分配算法,前者致力于获得最大的分配收益,后者则力求获得更好的时间效率;最后,通过仿真实验对所设计的带宽算法的性能进行验证。实验结果表明,相较于已有的弹性光数据中心网络中的带宽分配算法,设计的两种算法均能在带宽分配中获得更高的系统收益。

Algorithm for maximizing revenue in bandwidth allocation over elastic optical data center network

While conducting bandwidth allocation over elastic optical data center network EODCN,it is important to balance the revenue of cloud service providers and the connection performance of tenants.This paper proposed a bandwidth allocation algorithms to effectively assign optical bandwidth for multiple connection requests based on subcarriers so as to maximize revenue.Firstly,on the basis of a revenue model and a concept of virtual subcarrier,it constructed an integer linear programming formulation of the studied bandwidth allocation problem.Then,it proposed a dynamic programming-based algorithm and a heuristics-based algorithm correspondingly,which respectively focus on achieving maximum net revenue and gaining better time efficiency.Finally,it conducted simulations to validate the efficiency of the designed algorithms.The simulation results show that both of the proposed algorithms are superior to the existed methods in system revenue obtaining in bandwidth allocation over EODCN.