Analysis And Study Of The Correlation Of The Internet Primary Performance Parameters

With the Internet developing and the network size rapidly expansion have been greatly affected people's lifestyle. The Internet is becoming an indispensable tool for human beings, while the network performance has higher requirements. Because of the heterogeneous nature of the Internet, people have few understanding of the characteristics of the network behavior and the corelation of the network key performance parameters. So, carrying out the Internet measurement and analysis as well as the corelation analysis of key performance parameters is providing a basis for quantitative analysis. Qualitative and quantitative analysis of the Internet, is providing the basis for scientific managing, effective controlling and scientific planning the next-generation Internet. Network behavior analysis, and main performance parameters correlation analysis, have certain reference value and real-time forenoticing effect providing a basis for network future behavior analysis.Starting from the network applications of measurement, this paper study the network behavior, the Internet measurement parameters and the correlation analysis of key parameters. On this basis, this experiment testing the 15-25 hops of 33 nationwide sites, utilizing the tools of Traceroute and the Ping which wrote by myself under LINUX. Finally, we classify and mathematically analyse the collected experimental data, while doing a certain extent data mining, and then we get the expected results.The main work are as follows:First, the routing behavior, pack behavior, quantitative analysis of system behavior, and its measurement, measurement parameters, measurement methods were introduced. Network behavior can be modeled on human behavior analysis of vehicle behavior, and ultimately as a basis of network behavior analysis and forecasting. Simple network behavior can be analyzed and predicted.Second, mathematical model was built and the main performance parameters which can explain the corelation of latency and packet loss were analyzed.Besides, the correlation has been quantified to some extent.Through the experiment, the correlation of the main performance parameters of network latency and packet loss were analyzed and through mathematical and statistical methods, the corelation coefficient between them was given. Third, in this paper, network experiment was conducted, real data on the network was collected.According to selection,setting and optimization calculation of the major parameters such as packet size and contracting interval. End-to-end packet loss and throughput correlation, as well as packet loss and routing hops coralation were quantitativly analyzed on the basis of tests and data collection of 33 end-to-end network paths. The corelation coefficient between them was given too. Through the actual measurements of the network, this paper indicated that the Experimental analysis of key performance parameters could be a very good extension, and the corelation analysis had some role in guiding and forcasting for real-time applications.