Integrating Distributed Data Systems Using Ontologies, Web Services and Standards: An MMI Case Study

The Marine Metadata Interoperability (MMI) project promotes the exchange, integration and use of marine data through enhanced data publishing, discovery, documentation and accessibility. One of the goals of the MMI project for 2005 is to create a web application that can query distributed and heterogeneous data repositories using ontologies to solve the semantic heterogeneities, SOAP web services as transport protocols, and content standards such as those promoted by the Dublin Core Metadata Initiative. The MMI demonstration began by making available, in one portal, two heterogeneous and distributed data systems built by the Monterey Bay Aquarium Research Institute (MBARI). The two systems were the Shore-Side Data System (SSDS) and the Autonomous Ocean Sampling Network (AOSN). SSDS is a data management system designed to systematically collect and catalog both data streaming from deployed observatory instruments, and data contained in external data files. The AOSN data system facilitates data collection, storage, retrieval, discovery, and public access for the intensive, multi-institutional Monterey Bay field program in 2003. The systems use different data models, and different names identify similar data fields. We will present the process we followed to demonstrate an interoperable solution, and lessons learned during the course of the demonstration. The process included development of interoperable solutions for communication protocols, metadata content standards, and the vocabularies used to exchange the standard content. Simple interfaces were defined and iteratively improved, and vocabulary lists from each system (addressing parameters, instruments, and units of measurement) were exported and mapped. Similar processes have been advocated by MMI for a wide variety of interoperability challenges, and this demonstration represented the first experience using real world systems and data. From these lessons, we will improve a larger demonstration project, as well as the processes and tools we support to enable interoperable data management solutions.