Abstract: To add high availability and reliability to mobile networks, checkpoint based rollback recovery techniques are widely applicable. Checkpointing methods for traditional distributed systems cannot be applied directly to the mobile networks. The main focus in a single process checkpointing protocol is on ?nding optimal checkpoint interval to minimize the loss due to any fault, but in a distributed environment the main focus is on ?nding out and saving a global consistent state of the system. The challenge in ?nding a global consistent state is that interprocess communication creates dependencies that must be factored, otherwise the global checkpoint becomes useless. Mobile ad hoc networks throw up a plethora of challenges in tracking interprocess dependences including how to reliably save checkpoints in face of transience and node failures, where to save the checkpoints and how to reconstruct the stable global state from the nodes which are available after the fault. This paper present the simulated performance analysis of couple of uncoordinated checkpointing protocols for wireless ad-hoc networks on the basis of different size clusters, different number of recovery messages, different numbers of failures, and different recovery time.
Keywords: checkpointing, uncoordinated, recovery, simulation, algorithm, cluster.