Design And Implementation Of Campus Network Service Quality Active Perception System

With the continuous development of college education informatization in recent years,the campus network has provided many conveniences for teachers and students in schools to access resources inside and outside the school,broadened the students' horizons,and more and more businesses have been transferred to online management.However,due to the increasing number of campus network users and the limited number of management personnel,when there is a problem with the network service quality in some areas of the campus,it is inevitable that the processing will not be timely.In this context,this paper uses data mining technology and big data processing technology to regularly detect abnormal networks,and provides technical support for network administrators to timely grasp the quality of network services on campus networks.The main work done in this article is as follows:(1)Test data packets are regularly sent to online users of the campus network to measure the delay,jitter,upload speed,download speed,packet loss rate and other index data.The Flume framework is used to collect data files in real time and upload them to the HDFS in the built cluster for storage.(2)The K-means clustering algorithm for detecting abnormal network data is studied.In view of the shortcomings of the traditional K-means algorithm,this paper improves the original algorithm,uses the gap statistics method to improve the problem that the K value is difficult to determine,and uses the dichotomy method to reduce the sensitivity to the initial cluster center selection.Then through the KDD99 data set,it is verified that the improved algorithm can effectively reduce the number of iterations and can produce a better clustering effect.(3)The method of evaluating the comprehensive quality of the network is studied.Aiming at the problem of strong subjectivity of the analytic hierarchy process of the typical subjective weighting method,this article uses part of the collected data as a data set,uses the entropy method to determine the objective weight distribution value,and uses objective correction to subjectively optimize the combination of the two.A better index weight distribution is obtained,which determines the formula for calculating network service quality in this paper.(4)Designed and implemented a campus network service quality active perception system.The back end of the system implements modules such as data collection,processing,analysis,and statistics;the front end of the system displays the current comprehensive quality of network services,data records of abnormal online users,alarm information,and some statistical information.It provides a basis for the administrator to keep abreast of the current campus network.The administrator can notify the relevant management personnel to deal with the network that generates the alarm information.