Scalable distributed concolic testing: a case study on a flash storage platform

  • Authors:
  • Yunho Kim;Moonzoo Kim;Nam Dang

  • Affiliations:
  • CS Dept., KAIST, Daejeon, South Korea;CS Dept., KAIST, Daejeon, South Korea;CS Dept., KAIST, Daejeon, South Korea

  • Venue:
  • ICTAC'10 Proceedings of the 7th International colloquium conference on Theoretical aspects of computing
  • Year:
  • 2010

Quantified Score

Hi-index 0.03

Visualization

Abstract

Flash memory has become a virtually indispensable component for mobile devices in today's information society. However, conventional testing methods often fail to detect hidden bugs in flash file systems due to the difficulties involved in creating effective test cases. In contrast, the approach of model checking guarantees a complete analysis, but only on a limited scale. In the previous work, the authors applied concolic testing to the multi-sector read operation of a Samsung flash storage platform as a trade-off between the aforementioned two methods. This paper describes our continuing efforts to develop an effective and efficient verification framework for flash file systems. We developed a scalable distributed concolic algorithm that utilizes a large number of computing nodes. This new concolic algorithm can alleviate the limitations of the concolic approach caused by heavy computational cost. We applied the distributed concolic technique to the multi-sector read operation of a Samsung flash storage platform and compared the empirical results with results obtained with the original concolic algorithm.