A self-managing software system should be able to monitor and analyze its runtime behavior and make adaptation decisions accordingly to meet certain desirable objectives. Traditional software adaptation techniques and recent “models@runtime” approaches usually require an a priori model for a system’s dynamic behavior. Oftentimes the model is difficult to define and labor-intensive to maintain, and tends to get out of date due to adaptation and architecture decay. We propose an alternative approach that does not require defining the system’s behavior model beforehand, but instead involves mining software component interactions from system execution traces to build a probabilistic usage model, which is in turn used to analyze, plan, and execute adaptations. In this article, we demonstrate how such an approach can be realized and effectively used to address a variety of adaptation concerns. In particular, we describe the details of one application of this approach for safely applying dynamic changes to a running software system without creating inconsistencies. We also provide an overview of two other applications of the approach, identifying potentially malicious (abnormal) behavior for self-protection, and improving deployment of software components in a distributed setting for performance self-optimization. Finally, we report on our experiments with engineering self-management features in an emergency deployment system using the proposed mining approach.

An important challenge in dynamic adaptation of a software system is to prevent inconsistencies (failures) and disruptions in its operations during and after change. Several prior techniques have solved this problem with various tradeoffs. All of them, however, assume the availability of detailed component dependency models. This paper presents a complementary technique that solves this problem in settings where such models are either not available, difficult to build, or outdated due to the evolution of the software. Our approach first mines the execution history of a software system to infer a stochastic component dependency model, representing the probabilistic sequence of interactions among the system's components. We then demonstrate how this model could be used at runtime to infer the "best time" for adaptation of the system's components. We have thoroughly evaluated this research on a multi-user real world software system and under varying conditions.