Hierarchical Bandwidth Management and Intrusion Detection System for Campus Network

碩士 === 國立高雄應用科技大學 === 電子與資訊工程研究所碩士班 === 96 === The main direction of our research focuses on bandwidth management and detection of unusual attack and intrusion in campus network. There are two main research achievement would be proposed in this paper:Improved Hierarchy Bandwidth Management System an...

Full description

Bibliographic Details
Main Authors: Hong-Chi Shih, 施宏旗
Other Authors: Bin-Yih Liao
Format: Others
Language:zh-TW
Published: 2008
Online Access:http://ndltd.ncl.edu.tw/handle/37293298219293047165
Description
Summary:碩士 === 國立高雄應用科技大學 === 電子與資訊工程研究所碩士班 === 96 === The main direction of our research focuses on bandwidth management and detection of unusual attack and intrusion in campus network. There are two main research achievement would be proposed in this paper:Improved Hierarchy Bandwidth Management System and Detection of Network Attack and Intrusion Using PCA-ICA. For bandwidth management in campus network, our research utilize improved hierarchical management structure. The users’ capacity of network are the main consideration of classification, and the network structure is classified into subnet level and user level. The subnet level performs flow control over diversity of service according to each group and the number of people in group. For user level, user punishment mechanism is delivered on the people in unusual status according to the algorithm and connection status machine, and it will cohere with the balance of bandwidth distribution. For unusual attack and detection of intrusion in campus network, our research combine Principal Component Analysis (PCA) with Independent Component Analysis (ICA) to detect the network attack and intrusion. The main purpose to use PCA is for analyzing the feature of unusual intrusion and attack in network, and then ICA is used to fetch the independent basis for elevating the recognition ratio in whole. With the result of experimental simulation, obvious and better result to detect network attack and intrusion.