Enhancing network security via error-correcting codes.

摘要

Network security has received increased attention because of the explosive growth of computers, networks, and information technologies on one side and the threats or attacks that have happened, are happening, or may happen on the other side. One of the major problems over the Internet today is the denial of service (DoS) attacks. Multicasting is an efficient solution for group communication on the Internet. Multicast security is motivated by the enforcement of group access control, confidentiality and authentication of data transmission, and protection of the network infrastructure. In this dissertation, we develop an efficient IP traceback scheme based on error-correcting codes to defend against DoS attacks. The new scheme aims at releasing the cooperation of all the intervening ISPs, reducing the overhead in the IP header, and increasing the feasibility compared to available IP traceback schemes. We also construct three practical multicast re-keying schemes by using the Reed-Solomon codes and Algebraic Geometric codes. The analysis shows that there are no computational assumptions required for our schemes and the storage complexity is reduced greatly. Furthermore, we use Algebraic Geometric codes to solve the traitor tracing problem over the multicast channel. The obtained public key tracing scheme is deterministic and can trace all the participated traitors. Compared with other schemes available, our scheme can accommodate more users and tolerate more colluders given a fixed length of private keys.