基于SDN的可计算网络体系架构的研究及Q类业务保障的实现
发布时间：2018-07-17 08:28
【摘要】：近年来,随着Internet网络规模的急剧增大和应用类型的日益丰富多样,网络的结构和功能日趋复杂性,管控能力的逐渐变弱,网络资源分配问题越来越明显。当前网络难以为各种业务提供QoS(Quality of Service,服务质量)保证。因此,为新一代Internet网络提供可靠的服务质量保障是当前计算机网络研究的重要课题。SDN(Software Defined Network,软件定义网络)的出现对实现网络的可控可管具有十分重要的意义。然而,当前SDN框架中并没有实现对QoS业务的保障,仍有很大发展的空间。故需要一个全新的网络体系架构来满足当前网络的发展,提供高质量的QoS保障。为改变传统网络及其设备的无限状态、不可精细控制、不可计算的现状,本文提出了可计算网络模型。且针对当前网络无法为各类业务提供高质量保证的QoS服务的现状,本文将SDN体系架构与可计算网络模型相结合,提出基于SDN的可计算网络体系架构。为实现对网络的状态进行描述实现网络的可计算,在可计算网络模型中提出网络状态量化表示、网络可计算原理、业务映射、Q业务映射和X业务映射等概念。本文提出的基于SDN的可计算网络体系架构主要为用户提供两类业务服务:Q类和X类业务。其中,Q类业务主要包括语音、视频等实时性业务;X业务主要指传统IP的尽力而为业务。本文主要采取资源预留的方式为Q类业务提供QoS保证。本文首先将网络资源分为Q类和X类资源,对应的Q类和X类业务分别交由Q类和X类控制器进行管理和控制。其次,本文设计了数据包格式用于资源预留的请求和释放,并在控制器中设计了处理Q类业务的功能模块,具体包括:Q类资源分配与管理模块、路径选择模块、流量测量与监控模块。Q类资源分配与管理模块主要实现对Q类资源的分配与释放;路径选择模块主要根据Q类业务的QoS要求完成路径的计算;流量测量与监控模块主要实现对Q业务的监控,判断和验证它是否按照订购的QoS要求执行,并对Q类流的违规行为采取一定的处理策略。最后,本文以带宽为例对Q业务进行实验和测试验证,结果证明本文设计的处理Q类业务的各个功能模块可以正常工作,能够达到为Q业务提供QoS保障的目的。
[Abstract]:In recent years, with the rapid increase of the scale of Internet network and the increasing variety of application types, the structure and function of the network become more and more complex, the ability of management and control becomes weaker, and the problem of network resource allocation becomes more and more obvious. At present, it is difficult to provide QoS (quality of Service) guarantee for all kinds of services. Therefore, it is very important to provide reliable QoS assurance for the new generation Internet network. SDN (Software defined Network) is one of the most important topics in the research of computer network. The emergence of SDN (Software defined Network) is of great significance to realize the controllable management of the network. However, the current SDN framework does not guarantee QoS services, there is still a lot of room for development. Therefore, a new network architecture is needed to meet the current network development and provide high quality QoS guarantee. In order to change the infinite state of the traditional network and its equipment, the model of computable network is presented in this paper. Aiming at the current situation that the network can not provide high-quality QoS services for all kinds of services, this paper combines the SDN architecture with the computable network model, and proposes a computable network architecture based on SDN. In order to describe the state of the network and realize the computability of the network, the concepts of network state quantization representation, network computable principle, service mapping Q service mapping and X service mapping are put forward in the computable network model. In this paper, the architecture of SDN-based computable network provides two types of services: Q and X services for users. Among them, Q service mainly includes real time service such as voice, video and so on. This paper mainly adopts the way of resource reservation to provide QoS guarantee for class Q services. In this paper, the network resources are divided into Q class and X class resources. The corresponding Q class and X class services are managed and controlled by Q class and X class controllers, respectively. Secondly, this paper designs the request and release of data packet format for resource reservation, and designs the function module of dealing with class Q service in the controller, including: class Q resource allocation and management module, path selection module. The resource allocation and management module of class Q mainly realizes the allocation and release of class Q resources, and the path selection module completes the path calculation according to the QoS requirements of class Q services. The flow measurement and monitoring module mainly realizes the monitoring of Q traffic, judges and verifies whether it is executed according to the QoS requirements of the order, and adopts a certain strategy to deal with the violation of class Q flow. Finally, this paper takes the bandwidth as an example to test and verify the Q service. The results show that the functional modules designed in this paper can work normally and can provide QoS guarantee for the Q service.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02
本文编号：2129800
[Abstract]:In recent years, with the rapid increase of the scale of Internet network and the increasing variety of application types, the structure and function of the network become more and more complex, the ability of management and control becomes weaker, and the problem of network resource allocation becomes more and more obvious. At present, it is difficult to provide QoS (quality of Service) guarantee for all kinds of services. Therefore, it is very important to provide reliable QoS assurance for the new generation Internet network. SDN (Software defined Network) is one of the most important topics in the research of computer network. The emergence of SDN (Software defined Network) is of great significance to realize the controllable management of the network. However, the current SDN framework does not guarantee QoS services, there is still a lot of room for development. Therefore, a new network architecture is needed to meet the current network development and provide high quality QoS guarantee. In order to change the infinite state of the traditional network and its equipment, the model of computable network is presented in this paper. Aiming at the current situation that the network can not provide high-quality QoS services for all kinds of services, this paper combines the SDN architecture with the computable network model, and proposes a computable network architecture based on SDN. In order to describe the state of the network and realize the computability of the network, the concepts of network state quantization representation, network computable principle, service mapping Q service mapping and X service mapping are put forward in the computable network model. In this paper, the architecture of SDN-based computable network provides two types of services: Q and X services for users. Among them, Q service mainly includes real time service such as voice, video and so on. This paper mainly adopts the way of resource reservation to provide QoS guarantee for class Q services. In this paper, the network resources are divided into Q class and X class resources. The corresponding Q class and X class services are managed and controlled by Q class and X class controllers, respectively. Secondly, this paper designs the request and release of data packet format for resource reservation, and designs the function module of dealing with class Q service in the controller, including: class Q resource allocation and management module, path selection module. The resource allocation and management module of class Q mainly realizes the allocation and release of class Q resources, and the path selection module completes the path calculation according to the QoS requirements of class Q services. The flow measurement and monitoring module mainly realizes the monitoring of Q traffic, judges and verifies whether it is executed according to the QoS requirements of the order, and adopts a certain strategy to deal with the violation of class Q flow. Finally, this paper takes the bandwidth as an example to test and verify the Q service. The results show that the functional modules designed in this paper can work normally and can provide QoS guarantee for the Q service.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02
【参考文献】
相关期刊论文 前1条
1 刘千里,汪泽焱,倪明放,戴浩;一种基于多条件约束的QoS路由选择优化算法[J];计算机研究与发展;2001年03期
，本文编号：2129800
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2129800.html
