Modified reflective petri net for performance evaluation of policy-driven ASBS

Abstract

The adaption behavior of policy-driven ASBS is very flexible, while the performance of system is remarkably affected by the autonomy of Web services and dynamics of environments. In order to facilitate the design of efficient adaptive policy, it is important to have mechanisms to evaluate system performance. The performance evaluation of policy-driven ASBS is generally difficult due to the complexity of adaptive policies, particularly, when involving unpredictable dynamic environments. In this paper, we proposed modified Reflective Petri Net (mRPN) to analyze the effect of adaptive policy on system performance in different environments. The business behavior and adaptive behavior are modeled separately in mRPN for easy specify and independent analysis, also the model provides performance evaluate ability without changing the foundation of Petri Net. Through an example, we illustrate the use of mRPN in policy-driven ASBS modeling and performance evaluation procedure. © 2013 Springer Science+Business Media.