< retour aux publications

Effective checkpoint and rollback using hardware/OS collaboration

Auteur(s) : M. Portolan, R. Leveugle

Doc. Source: 22nd IEEE International Symposium on Defect and Fault-Tolerance in VLSI Systems (DFT'07)

Publisher : IEEE

Pages : 370-378

We present a technique that achieves an effective and complete checkpoint and rollback by exploiting collaborations between the Operating System and the cache system in a microprocessor-based environment. That scheme allows overheads to be kept extremely low, while ensuring a great generality.