面向电力信息网络多维感知识别技术的研究

发布时间：2019-07-08 20:10

【摘要】：随着智能电网的建设以及“三集五大”工作的进一步深化推进,电力信息网络承载了越来越多的业务。电力信息网络的安全性、高带宽、负载均衡性、支撑视频等均对网络提出了较高的性能要求。同时,电力信息网络缺乏对内部业务精确描述的问题日益明显。如何精确掌握网络业务性能特性成为电力信息网络研究的核心问题。网络业务识别是网络资源动态调配、网络流量预测、精细化网络运维、网络业务差异化保障控制的基础。电力信息网络内部业务精细化识别和分析成为亟待解决的问题。因此必须加强业务类型、用户等级、网络(位置、终端类型、应用提供方等)三方面的感知识别技术,以提升管道的智能化程度,为后期的业务控制、网络运维管理、优化设计等提供依据。本文针对电力信息网络业务识别中面临的主要问题,研究目前网络流量识别技术,提出一种适合电力信息网络的基于综合特征的多维感知识别模型。某省电力信息网络公司对此进行了研究,主要创新如下:首先,研究分析VPN隧道、匿名加密和未知业务或业务变化的业务感知实时识别模型。其次,建立电力信息网络多维感知识别方法的多识别引擎。从业务、用户、网络三个方面提出一种基于综合特征的多维感知识别模型。能够对电力信息网络业务性能进行精细化识别与分析,有利于对电力信息网络故障预判、网络设计与优化、流量控制、资源管理及分配。提高网络中业务的可视化程度,为进一步实现业务的差异化保障和控制奠定基础。最后,根据电力信息网络业务特性,采用C4.5算法进行业务识别。
[Abstract]:With the construction of smart grid and the further deepening of the work of "three sets and five major", the power information network carries more and more business. The security, high bandwidth, load balancing and supporting video of power information network all put forward high performance requirements for the network. At the same time, the lack of accurate description of internal business in power information network is becoming more and more obvious. How to accurately master the performance characteristics of network service has become the core problem of power information network research. Network service identification is the basis of dynamic allocation of network resources, network traffic prediction, fine network operation and maintenance, and network service differentiation guarantee control. Fine identification and analysis of internal services in power information network has become an urgent problem to be solved. Therefore, it is necessary to strengthen the perceptual recognition technology of service type, user level, network (location, terminal type, application provider, etc.) in order to improve the intelligence of pipeline and provide the basis for later service control, network operation and maintenance management, optimization design and so on. In this paper, aiming at the main problems in power information network service identification, the current network traffic identification technology is studied, and a multi-dimensional perceptual recognition model based on comprehensive features is proposed, which is suitable for power information network. The main innovations of a provincial power information network company are as follows: firstly, the real-time service perception recognition model of VPN tunnel, anonymous encryption and unknown business or service change is studied and analyzed. Secondly, the multi-recognition engine of multi-dimensional perceptual recognition method in power information network is established. A multi-dimensional perceptual recognition model based on comprehensive features is proposed from three aspects: service, user and network. It can accurately identify and analyze the service performance of power information network, which is beneficial to fault prediction, network design and optimization, flow control, resource management and allocation of power information network. Improve the degree of visualization of services in the network, and lay the foundation for further realization of service differentiation guarantee and control. Finally, according to the service characteristics of power information network, C4.5 algorithm is used for service identification.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM73;TM76

【相似文献】