Background:
User behavior models have been widely used to simulate attack behaviors
in the security domain. We revised all patents related to response to attack behavior models. How to
decide the protected target against multiple models of attack behaviors is studied.
Methods:
We utilize one perfect rational and three bounded rational behavior models to simulate attack
behaviors in cloud computing, and then investigate cloud provider’s response based on
Stackelberg game. The cloud provider plays the role of defender and it is assumed to be intelligent
enough to predict the attack behavior model. Based on the prediction accuracy, two schemes are
built in two situations.
Results:
If the defender can predict the attack behavior model accurately, a single-objective game
model is built to find the optimal protection strategy; otherwise, a multi-objective game model is
built to find the optimal protection strategy.
Conclusion:
The numerical results prove that the game theoretic model performs better in the corresponding
situation.
