面向自感知网络的测试关键技术研究

发布时间：2018-10-14 09:03

【摘要】：随着网络智能化程度的不断增加,人们对感知技术的研究越来越多。网络感知技术是实现高度智能化自组织网络的基础,通过对网络自身特性及所承载业务属性的感知,网络的运营、维护和管理将越来越不需要人工的干预,从而可以实现高度自动化。感知网络模型的建立、具体感知参数的标准化、以及感知参数的相关测试技术等都是当前关于感知网络的研究热点,也是实现自感知网络的关键。本论文建立了基于内容感知的网络模型,并从物理层、网络层、应用层等分层次角度研究了相关的感知网络技术；从测试技术角度出发,将不同层次的感知网络参数和对应的检测及测试技术结合在一起展开研究,利用软件定义网络的优势,对相关的测试技术方法作了创新性改进。论文的主要研究工作包括以下几个部分的内容： (1)基于内容感知的网络模型研究。根据报文内容,结合排队论,分析并建立了基于内容感知的网络数学模型。运用深度报文检测(DPI)与启发式行为检测相结合的检测技术,有效解决了传统网络对阻塞业务的优先级甄别问题,确保了高优先级业务优先分配资源,同时也保证了资源的公平使用。 (2)物理层主要测试指标及测试方法研究。设计并开发出了针对数字误码的实用测试工具,该仪器可以模拟产生不同的误码率,既可以单独作为测试仪器使用,也可以作为内容感知的光网络探针来使用,还可以作为光性能检测模块(OPM)技术的验证平台；提出了基于自相关函数的光纤色度色散测试方案,并进行了相关的实验验证。 (3)网络层性能指标及测试技术研究。为了评价网络节点的实际处理能力,基于实际应用环境中有大量的混合流量,我们建立了网络层新的性能指标测试模型,给出了网络节点主要参数(如：吞吐量、混合包的转发能力、新建连接速率等)的测试评估分析方法,并进行了相关的实验数据验证。 (4)应用层测试技术研究。一方面对异常流量检测技术进行了相关研究,提出了主要基于主成份分析的异常流量检测算法；另一方面对实际测试工具作了研发,基于软件定义网络(SDN),设计了网络应用层测试工具,该测试工具可针对各种通信网络、OPM、应用流量模拟等场景进行使用,可应用于面向下一代智能光网络的监控及测试系统。 (5)基于软件工程的测试技术创新。从软件工程角度分析了感知网络面临的威胁及对应的消减措施,从测试技术软件工程方法入手,针对感知网络面临的安全问题,对安全测试各个阶段的测试工作进行了研究,主要针对基于威胁建模的测试用例设计以及如何完成安全测试工作展开研究。总之,本论文的工作主要集中在感知网络模型的建立及各个分层的关键参数测试技术研究上,从具体指标的测试到软件工程层面都分别进行了较为系统的研究。结合实际工程应用,从目前亟需要解决的实际问题出发,建立了相应的数学模型,进行了具体的测试工具开发以及具体参数指标的测试研究,提出了一些创新的研究思路和解决方案,同时进行了相关的实验验证,为下一代智能化光网络的建立提供了技术支持。
[Abstract]:With the increasing network intelligence, people have more and more research on perception technology. The network sensing technology is the basis of realizing highly intelligent self-organizing network. By sensing the characteristics of the network and the perceived service attributes, the operation, maintenance and management of the network will become more and more unnecessary. Therefore, it is possible to realize the high degree of automation. The establishment of the perceptual network model, the standardization of the specific sensing parameters, and the related test technology of the sensing parameters are the research hot spots about the perception network, and also the key to realize the self-sensing network. This paper establishes a network model based on content perception, and studies the related sensing network technology from physical layer, network layer, application layer, etc. Starting from the analysis of different levels of perceived network parameters and corresponding detection and test techniques, the advantages of the network are defined by using software, and the related test techniques are innovative. The main research work of the thesis includes the following parts: Content: (1) Content-aware web Based on the message content and the queuing theory, the content-aware network is analyzed and established. Using the detection technology of depth message detection (DPI) and heuristic behavior detection, this paper effectively solves the problem of priority discrimination of traditional network to blocking service, ensures the high priority service to allocate resources preferentially, and also ensures the resources. Fair use of physical layer. (2) Physical layer primary test finger The design and development of a practical test tool for digital error code, which can simulate the different bit error rate, can be used as the test instrument alone or as the content perception. The optical network probe can also be used as the verification platform of the optical performance detection module (OPM) technology, Relevant experimental verification is provided. (3) Network layer In order to evaluate the actual processing capacity of network nodes, we set up a new performance index test model of network layer in order to evaluate the actual processing capacity of network nodes, and give the main parameters of network nodes (such as: Test and evaluation analysis method of emetic quantity, forwarding capacity of mixed packet, newly built connection rate, etc., and go hand in hand Relevant experimental data validation was performed. (4) Application layer test technology research. On the one hand, the abnormal flow detection technology is researched, and the abnormal flow detection algorithm based on the main component analysis is put forward. On the other hand, the actual test tool is researched and developed. Based on the software definition network (SDN).), a network application layer testing tool is designed, which can be used for various communication networks, OPM, application flow simulation and other scenarios, and can be applied to the next Intelligent optical network monitoring and testing system. (5) Test technology innovation based on software engineering. From the software engineering point of view, the threat and corresponding mitigation measures are analyzed. From the test technology software engineering method, the security problems faced by the perception network are analyzed. The test work of all phases of the whole test is studied, and the test cases based on threat modeling are mainly designed. In conclusion, the work of this paper mainly focuses on the establishment of perception network model and the key parameter testing technology of each layer. On the basis of practical engineering application, the corresponding mathematical model is set up based on the practical problems which are urgently needed to be solved, and the specific test tool development and the test of specific parameters are carried out. In this paper, some innovative research ideas and solutions have been put forward, and relevant experimental verification has been carried out.
【学位授予单位】：北京邮电大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN915.06

【参考文献】