Fuzzy rule based checkpointing arrangement for fault tolerance in Mobile Grids

Mobile Grids (MoGs) are an emerging sphere in distributed computing and are expected to become a crucial part of computational Grids in future. Apart from allowing a convenient user access to the Grid, mobile hosts can act as valuable computing resources in the Grid. However, mobile hosts have limitations of resources such as stable storage, wireless bandwidth, battery power etc. and are also subject to mobility and frequent disconnections. Therefore, a fault tolerance scheme is essential if the computing power of mobile hosts has to be used effectively in the grids. This paper presents a checkpointing based fault tolerance technique for Mobile Grid computing systems. A Fuzzy Rule based system (FRS) is used to construct an efficient checkpointing arrangement for MoGs. With FRS, superior nodes, i.e., having sufficient resources, are selected as Checkpoint Storage Nodes (CSNs) and other mobile hosts transfer their checkpointed data to one of the selected neighboring CSN. The objective of FRS is to increase the probability of recovery of checkpointed data subsequent to a failure; thereby allowing a distributed application to complete its execution successfully on the MoG. Simulations show FRS's better recovery probabilities in comparison to a random checkpointing arrangement.