An Adaptive Replanning Mechanism for Dependable Service-Based Systems

A promising way to guarantee dependability of service-based systems (SBSs) is replanning service bindings dynamically with the environment changing. As embedded into system execution, such replanning process will directly affect the overall performance of SBSs. While various replanning techniques have been proposed up to now, it is still challenging that how to provide cost-effective replanning methods for high-quality SBSs. In this paper, we propose a practical solution by improving the flexibility of replanning and establish an adaptive replanning mechanism for that. This new mechanism places more emphasis on the cause-effect relationship among system execution states, solution space changes, replanning strategies and their potential effects in system adaptation, and focuses on improving the following three leading aspects, i.e. replanning trigger, search scope determination and service selection, which affect the actual cost and effect of replanning. The main idea and key algorithms for implementing this adaptive replanning mechanism are presented. Besides this, empirical study based on a load rate querying service is used to illustrate and evaluate our approach.