Coordinated Checkpointing Algorithms: A Comparative Analysis

Abstract

In distributed system collection of independent computers appear to the users as a single computer. The property of fault tolerance is that it enables a system to continue operating properly if the failure occurs in the event of some of the components. Minimum process coordinated checkpointing is an appropriate approach to initiate fault tolerance in mobile distributed systems. Checkpoint-based rollback recovery restores the system state to the most recent reliable set of set of Checkpoints whenever a failure occurs. Checkpointing can be coordinated, uncoordinated and quasi-synchronous. Coordinated checkpoint process coordinates their checkpoints to form a system-wide consistent state. The approach is domino-free. Coordinated checkpointing can be blocking and nonblocking. Either of all the processes in the distributed system may need to checkpoint or only processes may be required to a minimum number of checkpoint. Reducing the number of processes to checkpoint may introduce blocking. The nonblocking checkpointing set of rules introduce overhead of piggybacking a few information for nonintrusiveness. The Performance analysis of two algorithms based on three parameters viz. packet delivery ratio, end to end delay and bytes overhead is done by implementing in Network simulator-2. The comparative results are shown in graphs and it is devised that algo of Awasthi is better than Kumar’s algo in terms of all the parameters taken.

Keywords: Distributed system, fault tolerance, checkpoint, coordinated checkpointing, checkpoint interval, Consistent global state

 Cite this Article

Rekha Rao, Jawahar Thakur. Coordinated Checkpointing Algorithms: A Comparative Analysis.Current Trends in Information Technology, 2015; 5(3): 14–20p.