| Achieving fault tolerance is an inevitable problem in distributed systems, with it becoming more challenging in decentralized, heterogeneous, and dynamic-environment systems such as a Grid. When deploying applications requires time-criticality, how to allocate resources for jobs in a fault-tolerant manner is an important issue for the delivery of the services.;The Water Threat Management project is a research to find solutions for the contamination incidents problems in urban water distribution systems, and it involves the development of the cyberinfrastructure in a Grid environment. To handle such urgent events properly, the deployment of the system demands real-time processing without the failure.;Our approach of integrating a fault-tolerant framework into a Water Threat Management system provides fault tolerance at the "queuing stage" rather than the "job-execution stage" by scheduling jobs in fault-tolerant ways. This includes the development of the batch queuing system in the Cyberaide Shell project. In addition, we present a dynamic workflow in the Water Threat Management system that can reduce the queue wait time in the changing environment. |
