Self-managing policies provide a self-management behavior for autonomic systems developed with ASSL (Autonomic System Specification Language). With ASSL we have successfully developed special autonomic prototypes of both the NASA ANTS (Autonomous Nano-Technology Swarm) concept mission and NASAs Voyager Mission. In these prototypes, we applied ASSL self-managing policies to drive the missions in critical situations in response to environmental or system changes. Therefore, the logical correctness of the ASSL specification of such policies appears to be of major importance. Experience has shown, however, that ASSL specifications may contain logical faults causing improper behavior. To handle such behavior, self-managing policies are often tested with manually injected inputs triggering events and satisfying constraints to allow for the activation, execution, and deactivation of these policies. The logical correctness of an ASSL self managing policy currently depends solely upon the relation between inputs and conclusion. In this paper, we present our initial work on a novel tool, part of the ASSL framework, that generates test cases based on change-impact analysis. Our main goal is to reduce testing costs and effort and improve the quality of testing, thus eventually assuring the logical correctness of the self-managing policies developed with ASSL.
展开▼
