Processors, Pipelines, and Protocols for Advanced Modeling Networks

Abstract

Predictive capabilities arise from our understanding of natural processes and our ability to construct models that accurately reproduce these processes. Although our modeling state-of-the-art is primarily limited by existing computational capabilities, other technical areas will soon present obstacles to the development and deployment of future predictive capabilities. Advancement of our modeling capabilities will require not only faster processors, but new processing algorithms, high-speed data pipelines, and a common software engineering framework that allows networking of diverse models that represent the many components of Earth's climate and weather system. Development and integration of these new capabilities will pose serious challenges to the Information Systems (IS) technology community. Designers of future IS infrastructures must deal with issues that include performance, reliability, interoperability, portability of data and software, and ultimately, the full integration of various ES model systems into a unified ES modeling network.