A secure execution framework for Java
Manfred Hauswirth, Clemens Kerer, Roman Kurmanowytsch · 2000
The Java platform facilitates to dynamically load and execute code from remote sources which can threaten the security and integrity of a system and the privacy of its users. To address these problems, Java includes a securityarchitecture which is based on a closed policy model. Although this model is sufficient to specify arbitrary policies, it easily may become cumbersome to use and is not well-suited for administering a consistent security policy for a complete network. The Java Secure Execution Framework (JSEF) overcomes these drawbacks: it introduces higher-level abstractions which enhance the expressiveness of policy rules# it simplifies the maintenance of security configurations# and it provides additional functionality and tools to make administration less error-prone. In JSEF we propose a hybrid policy model which supports additive and subtractivepermissions with a denial-take-precedence rule to resolve conflon --- Security profiles can be expressed in terms of hierarchical groups where a subgroup inherits the policy defined by its parent. All members of a group share the same set of permissions and users can be members of an arbitrary number of groups. JSEF's administrative model supports the definition of a network-wide policy which users can tailor to their needs but not break. At runtime JSEF enforces the defined security policy and supports security negotiation in case of insufficient permissions. A set of graphical tools supports the user in defining security policies and configuring JSEF.