UPPAAL in practice: quantitative verification of a RapidIO network

  • Authors:

  • Jiansheng Xing;Bart D. Theelen;Rom Langerak;Jaco Van De Pol;Jan Tretmans;J. P. M. Voeten

  • Affiliations:

  • University of Twente, Faculty of EEMCS, Formal Methods and Tools, Enschede, The Netherlands and Embedded Systems Institute, Eindhoven, The Netherlands;Embedded Systems Institute, Eindhoven, The Netherlands;University of Twente, Faculty of EEMCS, Formal Methods and Tools, Enschede, The Netherlands;University of Twente, Faculty of EEMCS, Formal Methods and Tools, Enschede, The Netherlands;Embedded Systems Institute, Eindhoven, The Netherlands;Embedded Systems Institute, Eindhoven, The Netherlands and Eindhoven University of Technology, Faculty of Electrical Engineering, Information and Communication Systems group, Eindhoven, The Nether ...

  • Venue:
  • ISoLA'10 Proceedings of the 4th international conference on Leveraging applications of formal methods, verification, and validation - Volume Part II
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Packet switched networks are widely used for interconnecting distributed computing platforms. RapidIO (Rapid Input/Output) is an industry standard for packet switched networks to interconnect multiple processor boards. Key performance metrics for these platforms include average-case and worst-case packet transfer latencies. We focus on verifying such quantitative properties for a RapidIO based multiprocessor platform that executes a motion control application. A performance model is available in the Parallel Object-Oriented Specification Language (POOSL) that allows for simulation based estimation results. It is however required to determine the exact worst-case latency as the application is time-critical. A model checking approach has been proposed in our previous work which transforms the POOSL model into an UPPAAL model. However, such an approach only works for a fairly small system. We extend the transformation approach with various heuristics to reduce the underlying state space, thereby providing an effective approximation approach that scales to industrial problems of a reasonable complexity.