Virtualizing Data Access for Generalized Geophysical Processing Systems

Lindquist, K.G., F.L. Vernon, A. Rajasekar, T.S. Hansen, B. Ludaescher, J. Orcutt, H.W. Braun, Y. Bock, J. Eakins, C.S. Keen

We have extended previous work in generalized real-time processing systems by rewriting the packet-handling infrastructure for diverse packet types, and by adding prototype support for transparent, virtualized data access. The redesigned packet-handling infrastructure has allowed us to add prototype support for the Campbell datalogger and for real-time GPS, alongside the remote-camera image data and several others we have previously incorporated. This framework will provide support for the intended addition of meteorology (temp, humidity, wind-speed), stream height, chemistry, etc., thus creating a generalized middleware layer for environmental data collection. We have also connected the Antelope-based real-time processing system to the data-grid technology provided by the San Diego Supercomputer Center's Storage Resource Broker (SRB). These connections include prototype stubs to the Datascope Relational Database Management System and to the Antelope orbserver. These extensions will allow collaborative access to distributed, heterogeneous sensor data from multiple signal domains, providing single-sign-on authentication, data replication, and data scaling to petabytes and hundreds of millions of files. Prototype world-wide-web interfaces to the near-real-time sensors and data-logger data streams demonstrate one of several possible access mechanisms for querying this consolidated data. Finally, we have begun adding active-rule based programming to federate real-time ORB systems.