Automatic Addition of Fault-Tolerance

It is difficult (if not impossible) to anticipate all types of faults that may perturb a program at the design time. Hence, as we encounter new types of faults, we need to add corresponding levels of fault-tolerance to the program at hand. An automatic approach for the addition of fault-tolerance to existing programs facilitates the adaptation of programs to new environmental conditions. Also, the automatically synthesized program is correct by construction and, hence, there is no need for its correctness proof. Moreover, since we begin with an existing fault-intolerant program, the derived fault-tolerant program reuses it. Therefore, it would be possible to add fault-tolerance even to programs for which the entire specification is not available. Third, in this approach, the concerns of the functionality of a program and its fault-tolerance are separated. This separation is known to help in simplifying the reuse of the techniques used in manually adding fault-tolerance. We expect that the same advantage will apply in the automated addition of fault-tolerance. 

To learn more about this line of research, please refer to my PhD dissertation.

Collaborator: Sandeep Kulkarni

Short bio            CV                               Research                Teaching              

Awards               Publications           For prospective PhD students!