Network Qos Monitoring And Research In IPv4/IPv6Dual-stack Environment

With the rapid development of network technology, the Internet as therepresentation of information network, has gradually penetrated into all aspects of ourlife, and plays a significant role in politics, military, economics, civilization and otherfields, this makes both network operators and users are increasingly concerned about thenetwork QoS.Based on collecting relevant data, network QoS monitoring describes features ofthe QoS. And it’s the basic of timely knowing the network performance and fullyunderstanding the network. It is the premise of network controlling and performancemanagement, it has very important significance. In this paper, takes the ChongqingUniversity IPv4/IPv6dual-stack campus network as the research object, then from theuser’s perspective and network management point of view to study the network QoS.Mainly include the following:1. The quality of user-end service is researched from the available bandwidth andWEB application. According to the PRM model of available bandwidth measurementand the characteristics of the IPv6protocol, an available bandwidth algorithm in IPv6networks is proposed. This method uses the IPv6header of flow label and traffic classto minimum the influence of the measurement and to ensure the path only, based on thatand uses the rate difference of sender and receiver to calculate the one-way availablebandwidth. The aspects that will impact the user’s perception when accessing the WEBare analyzed, from the above aspects to monitor the WEB workflow and collectstatistics of each performance, to reflecting the QoS that the network provided whenusing the WEB service.2. The network performance is monitored from the perspective of networkmanagement. According to the current situation of Chongqing University campusnetwork, the design goals and the overall structure of the Chongqing University NMSare proposed. The shortcomings of polling methods that the NMS commonly used areanalyzed then put forward an intelligent polling method. This method uses the linearregression to predict the sample values, according to the deviation of the actual andpredictive value, combined with the state of managed equipment and the occupancy rateof manage bandwidth, it will dynamically adjust the polling time to get more details ofthe managed object when the data changes by a big margin. 3. According to the above researching, the terminal QoS monitoring module andChongqing University NMS are achieved, and the NMS has deployed. The monitoringresults show that both methods can correctly get data, and accurately monitor thenetwork QoS.