Design And Implementation Of The Traffic Monitoring And Early Warning System In Mobile Data Network

With the popularity of smart phones and the rapid growth of mobile Internet services, the amount of data carried by mobile data network is growing with each passing day. Therefore, the traditional reactive type of network traffic management style can not satisfy the telecom operators’requirements in network traffic data management any more. Thus, a new management mode with real-time monitoring of network traffic and dynamic risk early warning system is the direction for future development of network traffic management mode.This paper design and implement a monitoring and early warning system for data network of China Mobile’s city branch. The system is used to monitor the data network traffic and.predict its future tendency of the traffic, and on this basis, do the forewarning for the network traffic. Thus, it is helpful for the network optimization staffs to control network status and optimize network in time and prevent the occurrence of network failures effectively. The data network monitoring and early warning system is composed of the data processing module, the database module, the data presentation module and the traffic forecasting and early warning module. The data processing module is able to accomplish data acquisition, data processing, data storage and so on. Based on Netstream, the traffic data is collected from the Gb interface unit of mobile data network. The data sampling rate and sampling interval are adjustable, and the data is stored in the database by IP or equipment. The database module is used to store the original network traffic data itself and its forecasting and early warning data, the data of the systematic parameters is stored in database as well. The human-computer interaction interface of the system is the data presentation module which function is to presenting and querying the system configuration and traffic data according to the requirements of the optimization staffs. Based on the historical data of network traffic, some prediction algorithms of time series are employed by the traffic forecasting and early warning module to predict the development trend of the network traffic. And the LSSVM module is selected as the system prediction model according to forecast results, and the early warning result is given when the network traffic is abnormal based on the prediction. The system development platform is VS2008, and the development language is C++. The modeling platform of the traffic forecasting and early warning module is Matlab2012. SQL Server2008is used as the database of the system. The monitoring and early warning system for data network has been running for six months in one company of China mobile. The running result shows that each module running is stable and reliable. The prediction result of modeling the traffic data of125days to forecast the9days shows that the MRE of the prediction result is lower than10%and the modeling time is less than3seconds for one configuration each day. It can meet the requirements of the China mobile company, and is suitable to do online monitoring and forewarning for mobile data network traffic.