The premises of this paper are 1) security is application dependent because application semantics directly influence proper protection; but 2) applications are generally too complex to be trusted to implement security as specified by the given security policy. These problems are aggravated if the application operates over time and space. This paper proposes the use of a simple program (a "wrapper") that has enough knowledge about a specific application's potential states and the actions that are permissible in each state. Using this knowledge, it is able to filter requests that should not reach an application at a given point. Key words: Web Security, Application security, Access control, Wrappers, State-based authorisation.
Martin S. Olivier, Ehud Gudes