THE WEB SERVICE BASED APPROACH FOR DATA DISTRIBUTION AT THE IRIS DMC

In 2010 the IRIS Data Management Center (DMC) began developing web service interfaces for most of our internal, primarily seismological, repositories. The design and evolution of these services has been guided by a few core principles: offer data in forms useful for both a broad audience and domain scientists, create interfaces directly usable by researchers, rely on basic tenants of RESTful service design, and initiate each service design with a common set of conventions. We have found that services created in this fashion ensure usability and consistency in a wide variety of computing environments and are readily integrated into new and existing software systems. By using a very common subset of HTTP, for which general support is ubiquitous, we ensure broad usability on a wide variety of platforms. Through interfaces design choices, documentation and interactive tools for illustrating the services, researchers are able to use the services with minimal technical expertise. By offering data in simple text and images along with formats specific to seismology, we intend to reduce barriers for use.

The suite of services has proven to be very robust and flexible and has allowed our data center’s distribution to continue growing exponentially. Not only are the services offered publicly, the IRIS DMC uses them internally in a Service Oriented Architecture to handle nearly all data extraction requests since 2014. Our general design and interface specification for 3 core services were adopted by the International Federation of Digital Seismograph Networks (FDSN) as standards for seismological data centers. Leveraging these standardized services running at more than 13 international data centers, we have developed the IRIS Federator to allow users to discover and access seismological data across centers. Our web service approach has also been generalized and applied to data centers and repositories in other geoscience disciplines through the GeoWS project, a building block in NSF’s EarthCube initiative. We believe the general concepts and lessons learned are applicable to distribution from any geoscience data repository and beyond.