Learning-based Firewall Technology

Internet technology is used widely now, there are some reasons to make the Internet security problem getting worse. These reasons include: the open and freedom of the Internet, and the fault of the TCP/IP.As the Internet technology developed these years, intrusions are getting complex. It's necessary to adopt positive security measures .To protect the electronic materials, firewall and intrusion detection system are getting more and more important.Machine learning is a field to study the computer self-studying processing. It hasn't as powerful studying capability as human beings, but some machine learning algorithms have been developed for some study mission. People also write many computer programs to realize the different studying missions, such as voice distinguish, human-machine playing chess. In the field of the Internet security, many machining learning algorithm ,for examples, neural network ,are used widely in the data mining and intrusion detection system.I studied the machine learning algorithm and write a program of ID3 algorithm. I studied the firewall technology and Dos attack, analyzing the fault when it's facing attacking, used TFN2K to attack the PC in our lab, collect the data from the attack. I chose the network data which has better attribute as a training example to realize a learning module ,then I added this module into the T.Rex firewall to generate the firewall rules. My work's goal is to make the firewall has learning ability.At last, I summarized the fault of studying and the future work, and analyzed the future of the machine learning used in firewall technology.