Distributed resource management and interoperability in computational modeling systems

A Computational Modeling System (CMS) provides scientific investigators with a unified computational environment and easy access to a broad range of modeling tools. The goal of a CMS is to provide computational support that increases the efficiency of scientists in the iterative process of modeling. A CMS consists of a computational modeling environment with distributed resource access support for the environment. The modeling environment is based on a characterization of scientific modeling activities that is focussed on the manner in which scientific concepts are represented, manipulated, and evaluated, in the scientific modeling process. We refer to this formalization as representational structures (or "R-structures"). Abstract Resource Views map R-structures to low-level physical data and tools. The distributed resource access component of the CMS provides a uniform interface that supports the access and manipulation of data and services distributed over a network. It provides a unifying framework for the integration of existing tools, such as data management and mathematical software modules, and a distributed modeling environment. We also provide a simple yet powerful strategy for migrating existing monolithic computational modeling applications to the high-level modeling environment of a CMS. We have developed an operational CMS, Amazonia, which combines the distributed resource access component and a graphical modeling environment. Amazonia supports earth science applications in terms of a specific set of R-structures and a "seamlessly" integrated and extensible collection of computational modules.