| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 91 === With the growth of computer networking and e-commerce climate, people have heavy relied on the internet. For this reason, more and more important information on the internet, security of networking systems has become an important research issue. TCP port 80 is the default port for HTTP Protocol. Web servers purposely listen for requests on this port. This provides hackers a good chance to attack on the Web server, and that poses a serious security threat. A Web server provides logs about every interaction between a site visitor and Web server. If an intrusion attempt has occurred, Web server logs provides a Web site manager with critical information for analyzing the security of the Web server and detecting signs of intrusion. However, the contents of logs are usually too huge for people to judge the meaning of it. To improve the traditional manual analysis method, it is necessary for the intrusion detection system to do real-time detection. Applying the Neural Network approach to intrusion detection, provide the potential to classify network activity based on adaptable learning method. In this paper, we propose a Hybrid-IDS framework that provides the ability to detect Microsoft IIS Web server attacks. This framework apply backpropagation neural network to build a real-time Web IDS. Using a set of benchmark data from ISS Internet Scanner[1]，we demonstrate that efficient and accurate classifier can be built to detect intrusions.
|
|---|
