Fault tolerance using token ring checkpointing in distributed mobile computing systems

Abstract

Mobile Computing represents a new paradigm that aims to provide continuous network connectivity to users regardless of their location. The vast computing potential of mobile distributed systems is often hampered by their susceptibility to failures. These failures lead to faults. The goal of designing the fault tolerance sys tem is to ensure that the system as a whole continues to function correctly, even in the presence of faults. Therefore, many techniques have been developed to add reliability and high vailability to distributed systems. Checkpointing is an attractive approach for transparently adding faulttolerance to distributed applications. Checkpointing is primarily used to avoid losing all the useful processing done before a fault has occurred. Checkpointing consists of intermittently saving the state of a program in a reliable storage medium. Mobile computing introduces new challenges in designing checkpointing algorithms that are used to make distributed applications fault-tolerant. Compared to traditional distributed environments, wireless networks are typically slower, providing lower throughput and latency. Additional constraints of lack of stable storage, low bandwidth, limited battery life and disconnections drastically hamper the performance of such algorithms. Recently, more attention has been paid to providing checkpointing protocols for mobile systems. Some of these protocols have been adopted from the traditional distributed environment, while others have been created from scratch for mobile systems.The goal of this research work is to design and implement an efficient non-blocking checkpointing algorithm suitable for distributed mobile systems. The proposed work uses the token ring concept in the checkpointing process. This helps the checkpointing process to minimize the amount of dependent details passed through the network and terminate the process. The token is considered as a checkpoint request message. The checkpoint request is passed through the network.