Countermeasures Optimization in Multiple Fault-Injection Context. Boespflug, E., Ene, C., Mounier, L., & Potet, M. In 2020 Workshop on Fault Detection and Tolerance in Cryptography (FDTC), pages 26–34. IEEE.
bibtex   
@InProceedings{boespflug20countermeasures,
  author       = {Boespflug, Etienne and Ene, Cristian and Mounier, Laurent and Potet, Marie-Laure},
  booktitle    = {2020 Workshop on Fault Detection and Tolerance in Cryptography (FDTC)},
  date         = {2020},
  title        = {Countermeasures Optimization in Multiple Fault-Injection Context},
  organization = {IEEE},
  pages        = {26--34},
  comment      = {* context: fault injection as a (security) attack vector
* understand *countermeasures check points* (CCPs) as important
  countermeasure

  * i.e., where program checks safety conditions at runtime

* focus on optimization

  * rather than on low overhead, as usual in literature

* formal descriptions

  * programs, execution, modification, injection and detection

* multi-fault injection

* can be extended, e.g., by duplicating CCPs

  * i.e., one for true, one for false branch
  * e.g.::

    condition = calculate_condition()
    if condition:
      if not condition:
        injection_detected()
      do_something()
    else
      if condition:
        injection_detected()

  * \# now, I think, if branches are attacked, there is a safety net

* code is being instrumented to contain CCPs
* algorithm to optimize (i.e., reduce) number of CCPs
* experiment

  * using Lazart

    * input: LLVM code, fault model, security property
    * dynamic symbolic execution

  * for single faults as example, 20% - 80% of CCPs can be
    removed (optimized)},
  file         = {:boespflug20countermeasures - Countermeasures Optimization in Multiple Fault-Injection Context.pdf:PDF},
  groups       = {fault tolerance},
  timestamp    = {2021-03-16},
}

Downloads: 0