An implementation and study of entropy and principal component analysis on Network Traffic

碩士 === 中原大學 === 通訊工程碩士學位學程 === 102 === It is a challenge task to conduct Entropy computation over the attributes of network packet header for high-speed computer network. In this thesis, we present a scheme to estimate entropy of high-speed network streams. It is based on the algorithm presented by...

Full description

Bibliographic Details
Main Authors: Hui-Ping Yu, 游輝評
Other Authors: Yu-Kuen Lai
Format: Others
Language:zh-TW
Published: 2014
Online Access:http://ndltd.ncl.edu.tw/handle/36244177384974993391
Description
Summary:碩士 === 中原大學 === 通訊工程碩士學位學程 === 102 === It is a challenge task to conduct Entropy computation over the attributes of network packet header for high-speed computer network. In this thesis, we present a scheme to estimate entropy of high-speed network streams. It is based on the algorithm presented by Ashwin Lall et al., we replace the exact count by Count Sketch and use Reservoir Sampling to resolve the issue of unknown length of data stream. We implement the measurement system on a NetFPGA-10G platform with principal component analysis (PCA) for real-time network traffic analysis.