大用户量异构软件定义网络关键技术研究

针对大用户量异构无线局域网络性能显著下降的突出问题，本项目构建基于软件定义网络架构的无线局域网，从网络应用测量、性能模型分析和性能优化方法等方面，探索有效提升无线局域网性能的理论和方法。本项目计划设计和实现支持软件定义网络的具有单接入点-多网络接口无线接入点系统；开展无线局域网下多种网络应用的用户体验测量研究；面向一般性异构随机接入网络，基于嵌入式离散时间马尔可夫过程模型，提出一种通用的理论分析框架，用来研究网络的吞吐量、延迟和稳定性等重要网络性能指标，并通过实际网络测量对该理论模型进行性能评测；设计和实现支持软件定义网络的分布式无线接入点测量、管理和优化平台，在校园网络部署测试实验床，进行实际网络性能测试。本课题的研究成果，将对于推动实现可测、可管和可控的大规模无线局域网，具有重要的现实意义。

In massive heterogeneous WiFi networks, users may experience severe network performance degradation. In this project, we propose to apply the emerging software-defined network architecture (SDN) into WiFi networks. Using network measurement, theoretical modeling and performance optimization methods, we aim to explore different theories and methods to improve the performance of WiFi networks effectively. We plan to design and implement an SDN-based One-AP-Multiple-Interface architecture for 802.11 wireless access points. We also plan to conduct a measurement study of the quality-of-experience for multiple network applications in WiFi networks. For the general random access networks, based on the embedded discrete Markov Chain, we propose a unified model to analyze stability, throughput and delay. This model will be validated and evaluated using real WiFi traffic measurements on a campus network. Last, we plan to design and implement an SDN-based virtual network for the measurement, management and optimization of WiFi networks. This project is expected to promote a measurable, manageable and controllable large-scale wireless local area network.