Efficient synthesis of robust models for stochastic systems

Radu Calinescu,Milan Češka,Simos Gerasimou,Marta Kwiatkowska,Nicola Paoletti

doi:10.1016/j.jss.2018.05.013

Abstract

We describe a tool-supported method for the efficient synthesis of parametric continuous-time Markov chains (pCTMC) that correspond to robust designs of a system under development. The pCTMCs generated by our RObust DEsign Synthesis (RODES) method are resilient to changes in the system’s operational profile, satisfy strict reliability, performance and other quality constraints, and are Pareto-optimal or nearly Pareto-optimal with respect to a set of quality optimisation criteria. By integrating sensitivity analysis at designer-specified tolerance levels and Pareto optimality, RODES produces designs that are potentially slightly suboptimal in return for less sensitivity—an acceptable trade-off in engineering practice. We demonstrate the effectiveness of our method and the efficiency of its GPU-accelerated tool support across multiple application domains by using RODES to design a producer-consumer system, a replicated file system and a workstation cluster system.

Highlights

Robustness is a key characteristic of both natural [1] and human-made [2] systems
We propose a tool-supported method for the efficient synthesis of parametric continuous-time Markov chains that correspond to robust system under development (SUD) designs
Methods that rigorously evaluate how the transition probabilities affect the satisfiability of temporal properties have been studied in the context of parameter synthesis

Summary

Introduction

Robustness is a key characteristic of both natural [1] and human-made [2] systems. Systems that cannot tolerate change are prone to frequent failures and require regular maintenance. As such, engineering disciplines like mechanical and electrical engineering treat robustness as a first-class citizen by designing their systems based on established tolerance standards Despite significant advances in software performance and reliability engineering [5, 6, 7, 8, 9, 10], the quality attributes of software systems are typically analysed for point estimates of stochastic system parameters such as component service rates or failure probabilities. Even the techniques that assess the sensitivity of quality attributes to parameter changes Even the techniques that assess the sensitivity of quality attributes to parameter changes (e.g. [11, 12, 13, 14, 15]) focus on the analysis of a given design at a time instead of systematically designing robustness into the system under development (SUD)

Methods

Results

Conclusion