面向SDN的网络编码组播机制研究

In the field of communications networks, it is an important research direction to seek for high-performance point-to-multiple-points data transmission technologies. As a new architecture that meets the trend of future network evolution, software defined networking (SDN) is in urgent need of an efficient multicast mechanism that is able to support multicast applications with increasing bandwidth demand and stringent quality of service requirements. On the other hand, network coding is superior to the traditional store-and-forward based routing, in terms of the network throughput, the network load balancing, etc. Considering this, the proposal introduces network coding into the framework of SDN, with the aim of exploring an efficient multicast mechanism in a network background featured with centralized control, decoupled control and forwarding functionalities, and dynamic network environment. It will be focused on investigating the fundamental theory and key technologies to construct and optimize the network coding based multicast. The research contents include the multicast routing scheme in a dynamic network environment and the generating and delivery methods for forwarding data streams. The key issues to be addressed are the modeling of multi-objective multicast routing sub-graph optimization problems and evolutionary algorithm based approaches to solve them. As a result, a relatively complete network coding based multicast mechanism for SDN will be reported, including its supporting theory and technologies. In the meantime, simulations and experiments will be conducted. The expected research achievements will contribute to the future research and development in the field of SDN and network coding, in the sense of providing important supporting theory and references.

探索高效的点到多点数据传输技术是通信网领域的重要研究内容之一。作为一种符合未来网络发展趋势的新型体系结构，软件定义网络（简称SDN）亟需一种有效的组播机制以支持带宽需求日益增多、服务质量保证要求严格的组播业务。另一方面，相比于 “存储-转发”式路由，网络编码技术在增加吞吐量、均衡负载等多方面有显著优势。本课题将网络编码引入SDN中，旨在探索一种集中控制、控制与转发解耦、环境动态变化等网络特征下的高效组播机制，重点研究网络编码组播路由构建与优化核心机理与关键技术，研究内容包括动态网络环境下的组播路由策略和数据流转发规则的生成与下发机制，着重解决多目标组播路由子图优化问题的建模及其基于进化算法的求解方法等核心问题，构建一套较完整的适于SDN的网络编码组播技术理论与实现方案，完成相关的仿真与实验，研究成果将为未来SDN和网络编码领域的研究与发展提供重要的理论依据和参考价值。

软件定义网络(Software Defined Networking，简称SDN)具备支持云计算、移动互联、大数据等背景下各类应用和网络创新的能力，有着光明的应用前景。点到多点的组播数据通信功能，是SDN支撑多种网络业务数据传输与内容分发的有效手段。相比于传统“存储-转发”式组播，网络编码组播不仅能确保达到组播吞吐量上限，而且具有提高带宽利用率、平衡网络负载、降低能量消耗、提升数据传输安全等级等多种优势。本课题旨在将网络编码组播引入SDN中，结合二者特长，发掘新型网络架构上的高效组播数据传输潜力，探索一种适合SDN的网络编码组播技术。本课题从网络编码组播路由的构建着手，围绕网络资源消耗、网络负载均衡、服务质量(Quality-of-Service，简称QoS)保证等实际需求，构建最优组播路由方案。在此基础上，探讨了多目标网络编码组播路由特征，并对基于进化算法的路由策略和动态网络环境下高效组播进行了深入研究。.主要研究成果包括：深入研究网络编码资源优化问题，分别提出基于蚁群算法和基于粒子群优化算法的两种求解方法，并获得了高质量解。围绕网络编码组播负载均衡问题展开研究，构建了负载均衡优化问题模型，并应用种群增量学习算法进行求解，通过引入双概率向量协同进化策略和局部搜索启发算子策略提升算法全局搜索和局部发掘能力，得到了理想的优化效果。考虑影响组播性能的多种因素，对多目标网络编码组播路由优化问题进行建模，并提出了基于分解的多目标进化算法用于研究，仿真证明了所获非支配解集在距离、延展性和分散均匀性等方面能很好地逼近帕累托最优前沿。研究了动态网络环境下网络编码组播优化问题，对该问题进行建模学习，并应用量子衍生进化算法和种群增量学习算法对待优化问题进行求解。围绕SDN中的组播通信、数据安全、故障恢复、组网架构等问题展开研究，重点探讨转发规则生成与下发、入侵检测、链路失效检测和光电混合组网方式等内容。本课题的研究成果在SDN高性能路由领域有着重要的应用前景。