基于校园网的网络性能测量与评价技术的研究

发布时间：2018-09-01 08:01

【摘要】：随着网络技术的迅速发展和网络应用业务的逐渐丰富，网络影响到生活的各个领域，用户对网络的稳定性、可用性和响应时间的要求逐渐变高。为了提高网络服务质量（Quality of Service，QoS）以确保网络能够承载各种网络应用业务，网络测量和网络评价技术显得尤为重要。网络测量以网络设备为主要被管对象，采集相关网络性能数据；网络评价以这些性能数据为基础，建立QoS的评价模型，对网络上不同的网络业务的运行情况进行分析和评估。本文针对重庆大学校园网络管理系统，从具体的网络应用业务的用户体验质量（Quality of Experience，QoE）的角度对网络测量和评价技术进行了相关的研究。研究的内容主要有以下几点： ①研究网络性能的测量指标及测量方法，提出一种自适应的流量数据采集算法。该方法通过计算在一段时间t内，流量数据大小的浮动值D相对于测量平滑区间n内的平均测量值D n的平均变化幅度以及变化速率，来调整下一次流量数据采集的时间间隔，从而获取更多流量数据变化的细节信息，减少了数据采集次数。 ②提出一种改进的SNMP性能数据采集方法。采用SNMPv2中的GetBulk操作与SNMP Get操作相结合的方式，通过区分标量对象、一次“下一个对象操作”的列对象和多次“下一个对象操作”的列对象，最大程度地增加每个Response数据包中返回的信息量，从而降低了网络管理站与代理设备的通讯次数，提高了代理设备处理Request报文的响应速度，大大减少了数据采集的冗余数据量。 ③针对不同厂商网络设备的Trap数据格式和数据项的差异，设计了一种通用的SNMP Trap数据采集方法，可以很及时地对网络中的故障和告警信息进行展示和统计，以方便网络管理者进行网络维护。 ④从网络业务质量的角度出发，提出一种面向网络业务的基于模糊层次分析法的质量评估模型。该模型考虑到网络中各个QoS性能指标参数之间的相互影响作用以及具体的网络应用业务对不同性能指标参数的敏感程度，，结合统计学方法，对校园网中两条网络路径的业务质量进行了评估和分析。 ⑤在重庆大学校园网络管理系统上进行具体的实践。通过结合ICMP与SNMP中轮询及Trap的测量方式，对校园网内网络路径的性能状况进行测量和评价。
[Abstract]:With the rapid development of network technology and the gradual enrichment of network application services, network affects every field of life, and users' requirements for network stability, availability and response time become higher and higher. In order to improve the quality of service (Quality of Service,QoS) to ensure that the network can carry a variety of network applications, network measurement and network evaluation technology is particularly important. The network measurement takes the network equipment as the main managed object, collects the related network performance data, and based on these performance data, establishes the evaluation model of QoS, and analyzes and evaluates the operation of different network services on the network. Based on the campus network management system of Chongqing University, this paper studies the network measurement and evaluation technology from the point of view of user experience quality (Quality of Experience,QoE) of specific network application services. The main contents of the research are as follows: 1 the measurement index and measurement method of network performance are studied, and an adaptive traffic data acquisition algorithm is proposed. This method adjusts the time interval of the next flow data acquisition by calculating the average change amplitude and the change rate of the floating value D of the flow data relative to the average measured value D n in the measured smoothing interval n within a period of time t. In order to obtain more detailed information of traffic data change and reduce the number of data acquisition. 2 an improved SNMP performance data acquisition method is proposed. By combining GetBulk operation with SNMP Get operation in SNMPv2, this paper distinguishes between scalar object, column object of "next object operation" and column object of "next object operation". The amount of information returned in each Response packet is increased to the greatest extent, which reduces the communication between the network management station and the agent device, and improves the response speed of the agent device in processing the Request message. The redundant data amount of data acquisition is greatly reduced. (3) aiming at the difference of Trap data format and data items of different manufacturers, a general SNMP Trap data acquisition method is designed. The fault and alarm information in the network can be displayed and counted in a timely manner, in order to facilitate network maintenance for network managers. 4 from the point of view of network service quality, A quality assessment model based on fuzzy analytic hierarchy process (FAHP) for network services is proposed. The model takes into account the interaction between each QoS performance parameter in the network and the sensitivity of specific network application services to different performance index parameters. The service quality of two network paths in the campus network is evaluated and analyzed. 5 the concrete practice is carried out on the campus network management system of Chongqing University. The performance of network path in campus network is measured and evaluated by means of polling and Trap measurement in ICMP and SNMP.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.18

【参考文献】