A Fault Injection Approach to Evaluate Soft-Error Dependability of System Calls. Amarnath, R., Bhat, S. N., Munk, P., & Thaden, E. In 2018 IEEE International Symposium on Software Reliability Engineering Workshops (ISSREW), pages 71–76, 2018. IEEE.
bibtex   
@InProceedings{amarnath18fault,
  author       = {Amarnath, Rakshith and Bhat, Shashank Nagesh and Munk, Peter and Thaden, Eike},
  title        = {A Fault Injection Approach to Evaluate Soft-Error Dependability of System Calls},
  booktitle    = {2018 IEEE International Symposium on Software Reliability Engineering Workshops (ISSREW)},
  year         = {2018},
  pages        = {71--76},
  organization = {IEEE},
  comment         = {* automotive context, specifically with regard to new POSIX-based AUTOSAR
  standard
* motivation: dependability for critical systems using COTS-grade hardware
* QEMU-based fault injection framework
  * i.e., hook into QEMU: after decoding guest instructions
  * bit flips in x86 registers (use Intel i5 for experiments)
  * during the execution Linux (4.10) system calls
  * classify effects on application level
    * classes
      * benign
      * detected and corrected
      * silent data corruption
      * timeout
      * application crash
      * kernel crash
      * system crash
      * QEMU crash
    * classes with respect to
      * system call
        * register injected in
* results can guide software-implemented hardware fault-tolerance (SIHFT)
  measures},
  file         = {:amarnath18fault - A Fault Injection Approach to Evaluate Soft-Error Dependability of System Calls.pdf:PDF},
  groups       = {fault injection},
  timestamp    = {2019-02-22},
}

Downloads: 0