In this paper, we propose an "end-to-end" approach that supports dynamic reconfiguration of software architectures taking advantage of graphical modeling, formal methods and aspect-oriented programming. There are three ingredients of the proposal. The specification end of the solution is covered by a new UML profile enabling to specify the desired architectural style (model), its invariants and the intended reconfiguration operations. In order to verify the consistency of the model and the preservation of the invariants after every reconfiguration, we automatically generate formal specifications in Z notation from the defined model. At the runtime enforcing end of the solution, we propose to encode the enforcement logic as aspect in the AspectJ language. The third important ingredient that makes our approach end-toend is the automatic translation of formal specifications into aspect-based enforcement code.
Formal methods such as Z and Petri nets can be used to specify invariants that should hold during the execution of component-based applications such as those regarding changes in the architecture of the application and valid sequences of architecture reconfigurations. Integrating logic for checking and enforcing these invariants into the application's implementation is generally done by adding appropriate code to the functional application code. In this paper, we discuss several limitations of this approach that may ensue in a disconnection between the application implementation and its formal specification. We propose an approach for specifying and enforcing architectural constraints, which combines formal methods and Aspect-Oriented Programming. We use the Z notation for describing the architectural invariants of the application and Petri nets for modeling coordination protocols. At the implementation level, aspects intercept architecture reconfiguration events and check according to the formal specification and the coordination protocol whether a reconfiguration action can be performed.
scite is a Brooklyn-based startup that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.