A design framework to information system functionality defense through diversity

Diversification is one of the most effective approaches to defend information systems against worms, virus, and malicious behavior. However, how to design an integrated information system to achieve effective diversity is still a challenge due to combinatorially exploded solution space and multiple conflicting design objectives. In the paper, we present a systematic framework employing a combination of configuration evaluation through controlled system simulations and a neural network based feedback learning mechanism to explore the solution space, thus achieving an effective design solution for the integrated system. A simulation model is employed to evaluate design solutions, and an artificial neural network is trained to approximate the behavior of the system using system feedback. Guided by the trained neural network, a multiple-objective evolutionary algorithm (MOEA) is proposed to search the solution space and identify potential good solutions. The MOEA incorporates the concept of Herbert Simon's satisficing. It integrates the decision maker's preference and uses his/her aspiration level for the performance as its search direction. Potentially good solutions are then evaluated through simulation. The newly obtained simulation results can refine the neural network. The exploration process stops until the result convergences or a satisfied solution is found. We demonstrate and validate our framework through a case study.