Abstract

In future, WSN (Wireless Sensor Network) which consists of n number of tiny wireless embedded systems will be an indispensable part of our daily life. Fault tolerance is an imported technique which has to be implemented in sensor networks to avoid catastrophic events in the network; failure in the network may lead to wastage of resources. Fault tolerance is the property that enables a system to continue operate properly in the event of failure of (or one or more faults within) some of its components. An improved version of checkpointing called on-demand checkpoint and rollback recovery scheme is implemented in this paper. The cluster head collects data from its member nodes and aggregates the collected data. To improve reliability and reduce recovery latency, our method uses coordinating checkpoint, non blocking checkpoint to make the checkpoint taken consistent in nature. In the proposed scheme, backup nodes monitor and checkpoint the current state of the cluster head on demand. Experimental comparisons with CTP (Collection Tree Protocol) and demand checkpoint with CTP shows time and Packet Delivery Ratio values are identical saying checkpointing does not affect the WSN performance due to inclusion of checkpoint. The proposed scheme is designed and implemented in Contiki OS using sensor network simulator COOJA.