An agent-based inter-application information flow control model

Access control in an application prevents information leakage in the application. The prevention can be achieved by controlling information flows. Many information flow control models have been developed. Since applications may cooperate, controlling information flows among cooperating applications is necessary. Our survey reveals that no existing model offers the control. We thus developed a model to control information flows among cooperating object-oriented applications. In designing the model, we require that cooperating applications communicate with one another through JAVA RMI (remote method invocation). Our model is based on the following considerations: when a RMI occurs, the security level of an argument should not be higher than that of the parameter receiving the argument's value, and the security level of a variable receiving a method return value should not be lower than that of the method return value. The model is agent-based. Moreover, different applications can embed different information flow control models. With this, the proposed model coordinates heterogeneous information flow control models. This paper presents our inter-application information flow control model.