Research On P2P Traffic Identification Method Based On Deep Packet Inspection

P2P technology has developed rapidly in recent years,which brings a abundant of resources to web users ,however, as a result of open charastic of P2P network,the P2P network also brings a lot of bandwidth and security problems, it is necessary to understand and analsis on P2P traffic activity in order to inspect P2P traffic. In this paper,we analysis the advantages and disadvantages of current P2P flow indentification methods,on this basis, we proposes a system of P2P traffic identification based on DPI and DFI, As we all know, the advantage of DPI is the high precision of identification, meanwhile, DFI can identify the P2P traffic which is encrypted, so the system we proposed can combine the advantage with each other efficiently, at the same time,we adopt three-filtration in order to reduce the load of DPI and DFI.On the basis of this model, we achieve the design and implementation of P2P traffic identification based on DPI and DFI, we propose the whole framework of the system in chapter three,the system contains three modules, there are the collection and filtration of traffic,DPI and DFI, firstly,the traffic will be filtered ,secondly,most known P2P traffic has been identificatied by DPI,finally we consider the result of DPI as samples to train and study , then we can identificatied the encrypted and unknown traffic.We achieved this system in chapter four,and the critical technique, deal process are described in detail in the core modules.at the same time,we test the accuracy,missing rate,misjudgement rate and performances of the system, and the test results prove that the system of P2P traffic identification based on DPI and DFI has a lower misjudgement rate and higher identification accuracy. Furthermore,the system can draw the graph of the traffic distribution for any host computers.it can also notes the results in the database in order to effectively manage the P2P traffic in the furture.This project is a novel P2P traffic indentification and management software which can be used in telecommunications industries (Telecom, China Netcom, China Unicom, China Mobile, CRC), Internet cafes, universities and research institutions, corporations, government agencies and so on.