Applying Adaptive Role-Based Model to Self-Adaptive System Constructing Problems: A Case Study

摘要

When computational process occurs in uncertain environments, the environmental changes often cause unpredictable influence on the computation. Self-adaptive software is one of the possible ways to tackle and handle this motivation. It evaluates its own behavior and surrounding environment, and adapts itself when the evaluation indicates that it is not accomplishing what the system is intended to do, or when better functionality is possible. Since one of the objectives of role models is to support object adaptation to the environment through role binding mechanism, it is expected that role modeling can be conveniently used to build self-adaptive systems. This paper provides some essential features or mechanisms of self-adaptive systems which are intended to be captured by adaptive role models. We propose Self-Epsilon model, an extension of our developed role-based model Epsilon, to combine and integrate all those features with role-based mechanisms into a single model. This paper also shows how the model can be applied to practically design and implement self-adaptive systems. We illustrate our approach through a case study of a simplified system of coordinated traffic jam monitoring used in some related work. The evaluation shows the reasonable and acceptable scalability of our model implementation in the framework.