基于DPI和BP神经网络的P2P流量识别研究  

P2P Traffic Identification Based on DPI and BP Neural Network

作　　者：万建伟 胡勇[1] WAN Jian-wei;HU Yong(College of Electronic Information, Sichuan University, Chengdu 610021)

机构地区：[1]四川大学电子信息学院,成都610021

出　　处：《现代计算机》2019年第10期31-35,共5页Modern Computer

摘　　要：为了准确地识别P2P流量,提出一种结合DPI和BP神经网络的新的检测模型,该模型对单一的识别方法进行优化以提高识别率。实验表明,该模型弥补DPI技术不能识别加密流量的不足,能够根据网络流量的特征有效地识别网络应用,提高检测的准确性。In order to accurately identify P2P traffic, proposes a new detection model combining DPI and BP neural network, which optimizes the single recognition method to improve the recognition rate. Experiments show that this model makes up for the inadequacy of DPI technology in identifying encrypted traffic, and can effectively identify network applications according to the characteristics of network traffic, thus im? proving the accuracy of detection.

关 键 词：P2P 识别 BP神经网络 DPI 

分 类 号：TP393.06[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]