绿色微数据中心与NGPON融合网络部署规划研究

发布时间：2018-04-23 05:05

本文选题：云计算 + 微数据中心　；参考：《大连海事大学》2015年硕士论文

【摘要】：为了实现云计算服务的易扩展,移动性强、部署简单,服务本地化和成本低等特性,一种新型的分布式微数据中心(Micro Data Centers, MicroDCs)与下一代无源光网络(Next-Generation Passive Optical Network, NGPON)的融合网架构。与传统的大型数据中心(Mag-DC)相比,这种融合网架构具有更高的经济效益,更低的传输延迟,更少的能源消耗和更高的扩展性。在本文中,我们主要研究的是MicroDC-NGPON融合网的部署规划和优化(Infrastructure Deployment and Optimization, IDO)问题。以最小化整个MicroDC-NGPON融合网部署花费为目标,不仅能够计算出MicroDC-NGPON网元设备的数目,还可以确定MicroDC-NGPON网元设备的最佳位置,旨在为完成了“最后一公里云服务”MicroDC-NGPON网络的部署规划提供方案。我们对IDO问题进行数学建模,并用Gurobi求解这个模型,求解结果表明我们提出的IDO问题是可解的。虽然通过Gurobi能够得到最优的部署结果,但是随着网络规模不断扩大,Gurobi求解问题的时间会成指数级增长。针对复杂度过高的问题我们提出新型启发式算法对IDO问题进行求解,旨在针对大规模网络部署规划优化时,能够快速获得近似最优解,提高模型优化求解效率。在仿真试验中,我们针对不同规模的网络进行部署从而验证网络模型的有效性,通过比较两种求解方法,可得出新型启发式算法能较快得到近似最优求解结果,对于大规模网络部署场景尤为适合。
[Abstract]:In order to realize the characteristics of cloud computing services such as easy expansibility, strong mobility, simple deployment, localization and low cost, a new distributed micro Data centers (microDCs) and Next-Generation Passive Optical Network (NGPON) are proposed. Compared with the traditional large data center (Mag-DC), this converged network architecture has higher economic efficiency, lower transmission delay, less energy consumption and higher scalability. In this paper, we mainly study the deployment planning of MicroDC-NGPON fusion network and the optimization of infrastructure Deployment and optimization (ido). Aiming at minimizing the deployment cost of the entire MicroDC-NGPON fusion network, not only can the number of MicroDC-NGPON network element devices be calculated, but also the best location of the MicroDC-NGPON network element devices can be determined. It is designed to provide a solution for the deployment planning of the Last kilometer Cloud Service (LKS) MicroDC-NGPON network. We model the IDO problem and use Gurobi to solve the model. The results show that the proposed IDO problem is solvable. Although the optimal deployment result can be obtained through Gurobi, the time of solving the problem will increase exponentially with the increasing of network size. A new heuristic algorithm is proposed to solve the IDO problem for the problem of high complexity, which aims at obtaining the approximate optimal solution quickly and improving the efficiency of the model optimization when the large-scale network deployment planning is optimized. In the simulation experiment, we deploy the network with different sizes to verify the effectiveness of the network model. By comparing the two methods, we can get the approximate optimal solution result quickly by the new heuristic algorithm. For large-scale network deployment scenarios are particularly suitable.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN929.1;TP308

【相似文献】