基于压缩感知的网络层析成像技术研究

项目名称： 基于压缩感知的网络层析成像技术研究

项目编号： No.61501135

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 无线电电子学、电信技术

项目作者： 杨京礼

作者单位： 哈尔滨工业大学

项目金额： 20万元

中文摘要： 网络测量是理解网络特征和行为最有效的途径，是对互联网进行网络控制、实施QoS保证和提高网络性能的基础。网络层析成像技术以其端到端和非协作测量的特点，成为目前国内外网络测量领域的研究热点。目前，网络层析成像技术的研究虽然取得了一定的成果，但仍然无法满足互联网规模急剧扩张趋势下各种网络业务对互联网实时准确测量的需求，亟需引入新理论、新方法，开辟新的研究思路。本项目在深入研究稀疏表示和压缩感知理论的基础上，突破图上网络性能参数稀疏表示、测量路径优化选择和链路性能参数快速重构等关键技术，解决网络层析成像框架下的网络性能参数实时测量问题，在保证测量精度的前提下提高网络测量的效率，降低测量过程造成的网络负载，满足大规模互联网实时精确测量的需求。本项目的研究可以为及时了解互联网运行状况、检测网络拥塞、管理和优化资源配置的提供重要依据，也可为保障网络安全，防范大规模网络攻击提供预警信息。

中文关键词： 网络测量；网络层析成像；压缩感知；扩散小波；随机游走

英文摘要： Network measurement is most effective way to understand the network characteristics and behavior, and it is also the base of the network control, QoS implementation and network performance improvement. Network tomography is the research hotspot of network measurement field for the features of end-to-end and non-cooperation. Research of network tomography technology has achieved some results, but still unable to meet the various network services of Internet scale rapid expansion trend of Internet real-time accurate measurement needs, so there is an urgent need to introduce new theories and methods to solve the problem. In order to improve the efficiency of network measurement and reduce the network load caused by measurement process without losing the premise of accuracy, this project will research the theory of sparse representation and compressive sensing that can help to break through the key points of sparse representation for network performance parameters, measurement paths optimization and fast reconstruction of the network performance parameters. The research of this project can provide the important conditions for the Internet status monitor, network congestion detection and network resource optimization, and also can help to ensure the network safety and prevent massive network attack.

英文关键词： Network measurement;Network tomography;Compressive sensing;Diffusion wavelet;Random walk