Application And Implemetion Of Content Audit State Tag And Pre-Processing In Network Monitor

Theories and products about Network Monitor become more and more, but various challenges come out here and there in Network Monitor system. About all, one of the most important challenges is Network Monitor speed. With the rapid development of Network Monitor transmission, huge amounts of NetFlow makes speed the neck of Network content audit. This thesis focuses on the research of Network audit speed and proposed a method with which the speed of audit system can be improved dramatically.As a matter of fact, lots of network contents are the same (such as hotspot). As is known that CyberEconomy is also eye-catch economy. That is to say the same contents are always transmitted in the network repeatedly. In the same LAN(local area network), users tend to pay attention to the same content especially because of their similar identity. So it is obvious that Content Audit Network Monitor Systems do a lot of repeated work .But most of the present Network Monitor systems neglect this problem and thus waste lots of system resources and slow down the audit speed.Based on what has been discussed above, this thesis put forward 'content audit state tag' mechanism to reduce unnecessary audit. The idea includes the following two steps: the first one is to check-up whether the content to be audited is in the database, if so, the last audit result can be used directly; otherwise, normal network content audit is needed . The second step is to save the audit result to state tag database in order that it can be queried later.According to state tag mechanism, the scheme of design and implementation of model framework, system flow, databases are proposed. It formed standard plugin(module) finally. Update strategy about audit state tag and the identification strategy of the content to be refreshed, a strategy named multi-random vectors content identification techniques, are also discussed in this thesis. It also brings up an idea named pre-processing mechanism based on the fact that the server load drops dramatically in early morning, so the system startup pre-processing audit forwardly at that time.