A Hybrid Fault Tolerance Scheme in Grid Computing

Deepak Srivastava, Ritu Garg

Abstract


An important goal in Grid environment is to construct foolproof fault tolerant system. Fault tolerance service is a mechanism that recovers the system following the occurrence of faults. In this paper, we suggest an improved hybrid checkpoint fault tolerance technique. In such scheme two types of checkpoints, either Cooperative checkpointing with Store-checkpoint (SCPs) or Cooperative checkpointing with Compare-checkpoint (CCPs) are inserted between store-and-compare-checkpoint (CSCPs). SCP and CCP are used to provide reduction in execution time by dynamically adjusting checkpoint interval and in Cooperative checkpoint, less likely checkpoints can be skipped and hence by combining these approaches overall performance is improved. Finally, this paper presents a simulation-based experimental result showing comparisons of performance of other technique too.

Keywords—Grid computing, Checkpoint technique, Cooperative checkpoint, Simulation.


Full Text:

PDF


DOI: https://doi.org/10.26483/ijarcs.v2i6.929

Refbacks

  • There are currently no refbacks.




Copyright (c) 2016 International Journal of Advanced Research in Computer Science