SDN中面向控制器与交换机可靠通信的资源规划

发布时间：2018-10-21 12:29

【摘要】：软件定义网络(Software Defined Networking, SDN)是一种创新的网络形式,引起了业界的广泛研究和重视,被认为是未来网络的一项重要技术。SDN将传统网络设备中的控制计算模块和数据处理模块分离,并将控制计算模块集中在一个中央控制器上,OpenFlow协议对这种分离进行了良好的支持。在SDN网络中,交换机只能依靠控制器计算的流表项才能工作。因此,如何确保控制器和交换机互相通信的可靠性是SDN网络中的一个关键问题。控制器和交换机互相通信的可靠性主要取决于两个方面的因素,一是SDN网络的拓扑,二是控制器的部署情况。因此在网络设计阶段,当给定网络拓扑时,合理的控制器部署方案能够有效提高控制器和交换机之间通信的可靠性。为了进一步保障控制器与交换机互相通信的可靠性,通常还需要对控制网络中的控制器和控制路径进行适当备份,以避免控制器或者控制路径发生故障。物理网络中的故障一部分是由于网络组件自身故障造成的,还有一部分是由于网络组件所处的环境发生的灾害所引起的。当网络所处的自然环境出现灾害时,很可能导致网络中的区域故障。此时,除了网络中的工作资源可能会受到区域故障的影响,预先为这些资源准备的备份资源也可能同时受到影响,此时这些备份资源将变得不可使用,无法发挥恢复故障资源的功能。综上所述,为了提高SDN网络中控制网络的可靠性,需要在部署控制网络时进行合理的资源规划。本文讨论了控制网络资源规划中的控制器部署问题和控制网络资源备份问题,主要的研究内容包括:(1)提出了一种基于最小覆盖的控制器部署方法。该方法的目标是在SDN网络初始部署阶段,以合理的方式为网络部署若干个控制器,使得控制网络的可靠性得以保障,同时控制器和交换机之间的传输时延满足要求。方法首先定义了衡量控制网络可靠性的指标,然后根据该指标提出了 SDN网络中的邻域和最小覆盖两个概念,将控制器部署问题转化成为求解SDN网络最小覆盖的问题,然后根据二分查找的思路,提出了一种算法搜索最小覆盖。(2)提出了一种面向控制网络区域故障的资源备份方法。本文先建立模型来描述和刻画区域故障的影响范围,衡量区域故障对网络的影响。对于控制器的备份,本文将主控制器考虑为区域故障的源头,衡量网络中其它节点受到的区域故障影响,同时考虑每一个节点与其它节点之间的平均传输距离,即同时考虑备份资源的工作效率,综合这两方面因素决策出合理的控制器备份方案。对于控制路径的备份,将控制路径中的一条随机链路考虑为区域故障的源头,衡量其它备份路径的生存性,从中选择出生存性最高的路径作为备份路径。
[Abstract]:The software defined network (Software Defined Networking, SDN) is an innovative network form, which has attracted extensive research and attention of the industry, and is regarded as an important technology in the future network. SDN separates the control computing module from the data processing module in the traditional network equipment. The control computing module is centralized on a central controller, and the separation is well supported by OpenFlow protocol. In SDN networks, switches can only work on the stream table items computed by the controller. Therefore, how to ensure the reliability of communication between controllers and switches is a key problem in SDN networks. The reliability of the communication between the controller and the switch mainly depends on two factors, one is the topology of the SDN network, the other is the deployment of the controller. Therefore, in the phase of network design, a reasonable controller deployment scheme can effectively improve the reliability of communication between the controller and the switch when the network topology is given. In order to ensure the reliability of the communication between the controller and the switch, it is usually necessary to backup the controller and the control path in the control network to avoid the failure of the controller or the control path. The failure of physical network is partly caused by the failure of the network component itself, and partly by the disaster of the environment in which the network component is located. When a disaster occurs in the natural environment of the network, it may lead to regional failure in the network. In this case, in addition to the work resources in the network may be affected by regional failures, and the backup resources prepared for these resources may also be affected, these backup resources will become unusable, Unable to perform the function of recovering a fault resource. In conclusion, in order to improve the reliability of the control network in SDN network, reasonable resource planning is needed when deploying the control network. This paper discusses the problem of controller deployment in control network resource planning and the problem of controlling network resource backup. The main research contents are as follows: (1) A controller deployment method based on minimum coverage is proposed. The aim of this method is to deploy several controllers for the SDN network in a reasonable manner during the initial deployment phase of the network, so that the reliability of the control network can be guaranteed, and the transmission delay between the controller and the switch can meet the requirements. Methods firstly, the index to measure the reliability of the control network is defined. Then, the concepts of neighborhood and minimum cover in SDN network are proposed, and the problem of controller deployment is transformed into the problem of solving the minimum coverage of SDN network. Then, according to the idea of binary search, an algorithm is proposed to search the minimum coverage. (2) A resource backup method for fault control network is proposed. In this paper, a model is established to describe and characterize the impact of regional faults on the network. For the backup of the controller, the main controller is considered as the source of the regional fault, and the influence of the other nodes in the network is measured. At the same time, the average transmission distance between each node and the other nodes is considered. Considering the efficiency of backup resource and synthesizing these two factors, a reasonable backup scheme of controller is obtained. For the backup of the control path, a random link in the control path is considered as the source of the regional fault, the survivability of other backup paths is measured, and the path with the highest survivability is selected as the backup path.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.0

【相似文献】