Memory efficient state-space analysis in software model-checking

  • Authors:
  • Anshuman Mukherjee;Zahir Tari;Peter Bertok

  • Affiliations:
  • RMIT University, Melbourne, VIC, Australia;RMIT University, Melbourne, VIC, Australia;RMIT University, Melbourne, VIC, Australia

  • Venue:
  • ACSC '10 Proceedings of the Thirty-Third Australasian Conferenc on Computer Science - Volume 102
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Formal methods have an unprecedented ability to endorse the correctness of a system. In spite of that, it has been limited to safety-critical and mission-critical systems owing to significant time and memory costs involved. Lately, our ever increasing dependency on software in all walks of our life has necessitated using formal methods for a wider range of softwares. In this paper, we propose an algorithm to make this possible by reducing the memory requirement for model checking, a widely used formal method. A modelchecker stores all explored states in memory to ensure termination. The proposed algorithm slash memory costs by storing these states in compressed form. In compressed form, a state is stored as how different it is from its previous state. Our experiments report a memory reduction of 95% with only doubling of computation delay. Aforesaid reduction allows model checking in a machine with only a fraction of memory needed otherwise. Consequently the advantage is twofold, 1) enormous savings as only a small physical memory is required and 2) as more states can now be stored in a memory of same size, the chances of complete state-space analysis is exceedingly high.