Runtime Enforcement of Data-centric Properties for Concurrent Service-Based Applications

Abstract

For service-based applications which are composed of multiple independent third-parties, continuous monitoring is required to assure that runtime behavior of the systems complies with specified properties. However, most existing work only detects the violation while not consider how to enforce the properties so that the constraint can not be violated at runtime. To address this limitation, this paper presents EnforceBCL, a framework for enforcing data-centric properties for concurrent service-based applications. Users of EnforceBCL can specify the properties to be enforced using the expressive behavior constraint enforcement language. Data-centric property is enforced at runtime by blocking the process whose next action would violate it. The impacted processes can be unblocked and allowed to execute when the specified property eventually reaches a safe state. EnforceBCL also provides the mechanism to detect possible deadlock during the enforcement of the property, and executes corresponding handler to solve the deadlock. To evaluate the effectiveness and efficiency of the proposed approach, we conducted several experiments. Results show that EnforceBCL is able to effectively enforce data-centric properties for concurrent service-based applications and also incurs less performance overhead.