Although fingerprinting techniques are helpful for security assessment, they have limited support to advanced security related applications. We have developed a new security framework focusing especially on the authentication reinforce- ment and the automatic generation of stateful firewall rules based on behavioral fingerprinting. Such fingerprinting is highly effective in capturing sequential patterns in the behavior of a device. A new machine learning technique is also adapted to monitor high speed networks by evaluating both computational complexity and experimented performances.