Detecting malicious activities in a computer cluster for developing dynamic honeypot

The need of securing computer networks in this digital era has remarkably raised. This is because advance in the information technology affects the risk of the networks. Moreover, highly confidential data are likely stored in a database server within the network. In order to mitigate possible security problems, we propose a mechanism which is able to detect a possible network attack. Next, honeypot is deployed according to this detecting result. It is by considering both internal and external factors, including the number of receiving packets, source machine being used for sending those packets, and dead processes. The experimental results which are taken by using a real environment show that the proposed method has been able to achieve better performance than existing algorithms.