ITR: Supporting Dynamic and Scalable Distributed System Infrastructures

University Of California-Davis, Davis CA

Investigators

Abstract

Contemporary distributed systems are increasingly extensible, heterogenous and dynamically evolving. Further, the execution environment (including the hosts and the network) is dynamic in its underlying structure and capacity. This research project aims to develop and analyze an adaptable distributed programming environment (including a programming model and a runtime system) for effectively utilizing such complex systems. The research project will develop a programming environment in which programs and its components can freely migrate across different hosts. The distribution of program components across different hosts can be controlled by binding specific mobility paradigms with the components. Further, the mobility paradigms can be changed dynamically, both by an application and/or by the runtime system. Such a programming model entails an execution environment in which distribution of program components can be controlled dynamically in order to adapt to changes in load, resource availability, resource distribution, component distribution and computational capabilities of hosts. The research will also develop runtime system techniques for creation and management of mobile components, resource allocation and security. It will develop adaptive scheduling schemes that allow hosts to precisely control allocation of local and distributed resources to non real-time and real-time distributed components.