Policy based network management in high assurance environments

We examine the constraints and problems caused by attempting to use policy based network management (PBNM) in high assurance and distributed environments. There has been much previous work in both policy and security. However, policy management and security architectures have always been somewhat orthogonal, particularly for multiple security domains. PBNM could be used to bring together disparate systems and networks, driven from a minimum set of interfaces and high level goals, but the security constraints may still force separation of networks and precise specification of low level actions. We propose a solution that can meet the strong security constraints of high assurance environments while still retaining many of the benefits of PBNM. The solution has been designed for an application scenario consisting of an unplanned naval coalition network. However, the results may be applicable in other high assurance environments such as high availability or safety critical networks.