RASSH - Reinforced adaptive SSH honeypot

The wide spread of cyber-attacks made the need of gathering as much information as possible about them, a real demand in nowadays global context. The honeypot systems have become a powerful tool on the way to accomplish that. Researchers have already focused on the development of various honeypot systems but the fact that their administration is time consuming made clear the need of self-adaptive honeypot system capable of learning from their interaction with attackers. This paper presents a self-adaptive honeypot system we are developing that tries to overlap some of the disadvantaged that existing systems have. The proposed honeypot is a medium interaction system developed using Python and it emulates a SSH (Secure Shell) server. The system is capable of interacting with the attackers by means of reinforcement learning algorithms.