A distributed data component for the Open Modeling Interface

As the volume of collected data continues to increase in the environmental sciences, so too does the need for effective means for accessing those data. We have developed an Open Modeling Interface (OpenMI) data component that retrieves input data for model components from environmental information systems and delivers output data to those systems. The adoption of standards for both model component input-output interfaces and web services make it possible for the component to be reconfigured for use with different linked models and various online systems. The data component employs three techniques tailored to the unique design of the OpenMI that enable efficient operation: caching, prefetching, and buffering, making it capable of scaling to large numbers of simultaneous simulations executing on a computational grid. We present the design of the component, an evaluation of its performance, and a case study demonstrating how it can be incorporated into modeling studies.