面向SDN网络的组播路由问题研究
发布时间：2018-07-14 10:56
【摘要】：软件定义网络(Software Defined Networking,简称SDN)作为一种新型网络架构,弥补了传统网络的诸多缺陷,为网络创新提供了更多的机会。由于组播业务在传统网络里已比较成熟,SDN在革新传统网络过程中也有必要实现组播业务,以完善自身所能提供的网络服务。所以,如何设计和实现SDN架构下的组播服务成为了 SDN领域内亟待解决的主要问题之一。组播路由问题是提供组播服务所要解决的关键问题。在网络业务中,服务质量(Quality of Service,简称QoS)体现了网络服务所能满足用户需求的程度,通常网络业务的QoS要求会通过若干网络性质(如带宽、时延等)的阈值约束来体现,如何求解满足服务质量要求的QoS组播路由问题已成为当今学术界研究的前沿课题之一。目前人们多采用进化算法来求解QoS组播路由问题,包括蚁群优化(Ant Colony Optimization,简称ACO)算法。然而现有ACO算法存在着耗时过长、全局搜索能力不强等缺陷,为此本文提出了一种新型ACO算法来求解带宽、时延约束组播路由问题,该算法采用新型蚂蚁寻路规则和信息素更新策略,弥补了已有ACO算法耗时过长的缺点,使其局部搜索能力和全局搜索能力有了较大提升。本文自主生成了若干组播请求场景,用于模拟带宽、时延约束组播路由问题。组播请求场景的仿真实验表明,本文提出的新型ACO算法相较于已有ACO算法在算法运行时间性能上优势明显,求得问题解的质量更佳。但是,新型ACO算法耗时仍然较长,而果蝇优化(Fruit Fly Optimization,简称FFO)算法这种新兴进化算法具有参数少、实现简单、运行速度快等优点,正弥补了新型ACO算法的不足。本文首次将FFO算法应用于求解带宽、时延约束组播路由问题,采用了一种新型果蝇位置表述方式,以使算法能用于求解组播路由问题。算法采用了果蝇嗅觉随机搜索和视觉定位的新策略,以提升算法求解能力。通过仿真实验表明,本文采用的FFO算法相较于多种目前主流进化算法在算法运行时间、收敛速度和求解质量等性能指标上均有较大优势,证明了本文FFO算法性能的优越性。本文设计并实现了一套SDN网络QoS组播服务系统。该系统应用Mininet仿真工具、Ryu控制器和OpenFlow协议及技术标准,构建了组播组管理、拓扑管理、环路控制、QoS组播路由、请求调度等5大系统模块。系统实现了本文提出ACO算法和FFO算法,用于组播路由计算环节。仿真实验表明,本系统能在SDN网络中提供较完整的组播服务,对如何在SDN架构下提供组播服务提供了一定的参考价值。
[Abstract]:As a new network architecture, Software defined Network (SDN) makes up for many defects of traditional network and provides more opportunities for network innovation. Since multicast services are mature in traditional networks, it is necessary for SDN to realize multicast services in the process of reforming traditional networks in order to perfect the network services that they can provide. Therefore, how to design and implement multicast services based on SDN has become one of the most important problems in SDN. Multicast routing is a key problem in providing multicast services. In network services, quality of Service (QoS) reflects the degree to which network services can meet the needs of users. Usually, the QoS requirements of network services are reflected by the threshold constraints of some network properties (such as bandwidth, delay, etc.). How to solve QoS multicast routing problem which meets QoS requirements has become one of the leading topics in academic research. At present, evolutionary algorithms are widely used to solve QoS multicast routing problems, including Ant colony optimization (ACO) algorithm. However, the existing ACO algorithms have some shortcomings, such as long time consuming and weak global search ability. Therefore, a new ACO algorithm is proposed to solve the bandwidth, delay constrained multicast routing problem. The algorithm adopts new ant routing rules and pheromone updating strategy, which makes up for the long time consuming of the existing ACO algorithm and improves its local search ability and global search ability. In this paper, several multicast request scenarios are generated independently to simulate bandwidth and delay constrained multicast routing problems. The simulation results of multicast request scene show that the proposed new ACO algorithm is superior to the existing ACO algorithm in the performance of the algorithm's running time, and the quality of the solution to the problem is better than that of the existing ACO algorithm. However, the new ACO algorithm still takes a long time. Fruit fly Optimization (FFO), a new evolutionary algorithm, has the advantages of less parameters, simpler implementation and faster running speed, which makes up for the shortcomings of the new ACO algorithm. In this paper, the FFO algorithm is first applied to solve the bandwidth, delay constrained multicast routing problem. A new Drosophila location representation method is used to solve the multicast routing problem. The algorithm adopts a new strategy of random search of Drosophila olfactory and visual location to improve the ability of the algorithm. The simulation results show that the proposed FFO algorithm is superior to other popular evolutionary algorithms in terms of performance indexes such as running time, convergence speed and solution quality, which proves the superiority of the FFO algorithm in this paper. This paper designs and implements a set of QoS multicast service system in SDN network. Based on Mininet simulation tool, Ryu controller, OpenFlow protocol and technical standard, five system modules, such as multicast group management, topology management, loop control, QoS multicast routing and request scheduling, are constructed. ACO algorithm and FFO algorithm are proposed in this paper, which can be used in multicast routing computation. The simulation results show that the system can provide a complete multicast service in SDN network, which provides a certain reference value for how to provide multicast services in SDN architecture.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.02
本文编号：2121420
[Abstract]:As a new network architecture, Software defined Network (SDN) makes up for many defects of traditional network and provides more opportunities for network innovation. Since multicast services are mature in traditional networks, it is necessary for SDN to realize multicast services in the process of reforming traditional networks in order to perfect the network services that they can provide. Therefore, how to design and implement multicast services based on SDN has become one of the most important problems in SDN. Multicast routing is a key problem in providing multicast services. In network services, quality of Service (QoS) reflects the degree to which network services can meet the needs of users. Usually, the QoS requirements of network services are reflected by the threshold constraints of some network properties (such as bandwidth, delay, etc.). How to solve QoS multicast routing problem which meets QoS requirements has become one of the leading topics in academic research. At present, evolutionary algorithms are widely used to solve QoS multicast routing problems, including Ant colony optimization (ACO) algorithm. However, the existing ACO algorithms have some shortcomings, such as long time consuming and weak global search ability. Therefore, a new ACO algorithm is proposed to solve the bandwidth, delay constrained multicast routing problem. The algorithm adopts new ant routing rules and pheromone updating strategy, which makes up for the long time consuming of the existing ACO algorithm and improves its local search ability and global search ability. In this paper, several multicast request scenarios are generated independently to simulate bandwidth and delay constrained multicast routing problems. The simulation results of multicast request scene show that the proposed new ACO algorithm is superior to the existing ACO algorithm in the performance of the algorithm's running time, and the quality of the solution to the problem is better than that of the existing ACO algorithm. However, the new ACO algorithm still takes a long time. Fruit fly Optimization (FFO), a new evolutionary algorithm, has the advantages of less parameters, simpler implementation and faster running speed, which makes up for the shortcomings of the new ACO algorithm. In this paper, the FFO algorithm is first applied to solve the bandwidth, delay constrained multicast routing problem. A new Drosophila location representation method is used to solve the multicast routing problem. The algorithm adopts a new strategy of random search of Drosophila olfactory and visual location to improve the ability of the algorithm. The simulation results show that the proposed FFO algorithm is superior to other popular evolutionary algorithms in terms of performance indexes such as running time, convergence speed and solution quality, which proves the superiority of the FFO algorithm in this paper. This paper designs and implements a set of QoS multicast service system in SDN network. Based on Mininet simulation tool, Ryu controller, OpenFlow protocol and technical standard, five system modules, such as multicast group management, topology management, loop control, QoS multicast routing and request scheduling, are constructed. ACO algorithm and FFO algorithm are proposed in this paper, which can be used in multicast routing computation. The simulation results show that the system can provide a complete multicast service in SDN network, which provides a certain reference value for how to provide multicast services in SDN architecture.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.02
【参考文献】
相关期刊论文 前2条
1 霍慧慧;李国勇;;基于自适应果蝇算法的神经网络结构训练[J];微电子学与计算机;2016年01期
2 陈杰;张洪伟;;基于自适应蚁群算法的QoS组播路由算法[J];计算机工程;2008年13期
，本文编号：2121420
本文链接：https://www.wllwen.com/shoufeilunwen/xixikjs/2121420.html
