Research And Implementation Of Probability Packet Marking Based IP Traceback

With the development and popularization of computer network, our working and living environment has been greatly improved, the Internet provides a more convenient way to work and makes life easier. However, everthing has its other sides, hackers often make use of the vulnerable networks to attack innocent victims frequently. Facing these problems, we were used to passive defense, such as taking the patch, installing a firewall, reinforcing the system, which can only solve the problem temporarily.IP traceback technology is a kind of network security technology which can trace IP source, and its purpose is to find the attacker from the network. In this way, it can solve most of network security issues, such method has been taken more attentions in these years.This paper introduces the IP traceback technology firstly, analyses various technical branches and compares those functions or effects deeply.There are some disadvantages in traditional IP traceback methods, such as the loss of marking information, the difficulties to reconstruct attack path, low accuracy and so on. The thesis presents a new approach, called dynamic probabilistic packet marking (DPPM), to further improve the effectiveness of PPM. Instead of using a fixed marking probability, it proposes to judge whether the packet has been marked or not then choose a proper marking probability. DPPM may solve most of the problems in PPM methods. Formal analysis indicates that DPPM outperforms PPM in most aspects.In the end, in order to evaluate the performance of the method, experiments are made in different environments, which include real network environments and and network simulation software NS2.