Data management for urgent computing environments
Urgent computing cyberinfrastructure provides scientists with on-demand, event driven computing capabilities in distributed computing environments. However, current urgent computing cyberinfrastructure does not support on-demand, priority-based allocation of data resources and services, such as data storage or data transfer services. Data services and resources are important for the emerging class of urgent computing work flows because these work flows are typically data intensive, require storage to temporarily store or stage data between work flow tasks, and integrate a wide variety of data capabilities. Urgent data services and resources are necessary so that these urgent work flows do not bottleneck on inappropriately managed or provisioned data resources. Proper management of urgent data resources requires infrastructure to provision storage and network resources to fulfill the Quality of Service requirements of urgent computing work flows, manage priority-based access to many distributed resources in Grid computing environments, and integrate these data services with existing urgent computing mechanisms. This dissertation describes the current urgent computing infrastructure, emerging urgent computing work flows, development of an Urgent Data Management Framework (UDMF) that provides the necessary capabilities for urgent computing work flows, and an evaluation of UDMF's capabilities
Thesis, Dissertation, English, [2009]