A general approach to data-intensive computing using the Meandre component-based framework

Data deluge is the norm for research as the volumes of data grow every day. Another complication is the growing number of places and ways that the data needs to be accessed. This leads to a common requirement to integrate a variety from data source and providers to work with the informational resources. This paper describes a general approach to data-intensive computing using a component-based framework. Using the Meandre component-based framework to provided the basis for testing the hypothesis; that a general approach to data-intensive computing integration could be accomplished by creating a family of specialized dataflow components that interface with a given API specification and/or protocols. This inherent functional ability allows the dataflow developers to logically create applications, web services, and/or to directly integrate services provided by other applications or servers. The Fedora Commons Repository Experiment provides two cases that demonstrate how the component-based framework for data-intensive computing can map functional API specifications to discrete functional components that may be used in any number of data flows. This paper presents an overview of the Meandre's anatomy and some functional features it provides for components, it describes the experiment and examines the results that affirm that the approach is viable with the potential to be broadly applied to other API specifications that can then be pooled and combined with other component collections by flow developers to create integrated, interactive, and interoperable application and/or services.