Analysis Of Mechanisms For Mobile Internet Services Identification

The applications of the mobile network have developed faster as the3-G has becomemore and more popular. The detection and flow statistics of the services have directrelationship with quality and efficiency of the network, it can also reduce the economicloss and protect the network from being attacked by efficient service detections.We firstly introduce the background and the tendency of the mobile network detection,and briefly summarize the relevant technology, i.e., deep packet investigation（DPI）、regular expreviewssion、SMI and currentrent popular service detection technology, andanalyses the advantages and the shortcomings of all the technology.In this paper, the previewsent mobile network services are being considered, includingWetChat、Neteasy Twitter、bit download and mobile QQ service, whose pattern andrelevant protocol are discussed in detail, and based on that, particular service detectionmethods are proposed. We propose two methods, i.e., based on flow analysis and based onregular expreviewssion matching. Based on flow analysis is based on the five elementsclassify, i.e., five same elements are classified as a kind. The based on regularexpreviewssion matching methods are using SVM learn training, set up four kinds ofservice regular expreviewssions. Also, we can combine the two methods, i.e., we use flowanalysis to classify the abundant TCP data, and then we use regular expreviewssion tosearch the IP packet application layer data, finally the successfully searched data areclassified to a same kind of service.According to the Identification scheme,this paper previewpared appropriate rocedures.the results indicate that the way based on analysis and regular expreviewssions is a simpleand efficient method of identification.