Date Added: Dec 2011
This paper proposes a new design scheme for hybrid dynamic worm detection system. In the proposed scheme, the features of matching detection method based on protocol analysis, Bayesian detection method based on probability, and honeypot-based detection method are integrated for a relative complemented function to develop the advantages and avoid the disadvantages of each detection method. The designed hybrid dynamic worm detection system can be used for effective detection of abnormal data behavior of the hosts attacked by worms, and it also serves for dynamic real-time update of the rule base. Then a preliminary functional verification is performed on this system.