Grid Services-based Data Stream for LEAD
|Full Title:||Grid Services-based Data Stream for LEAD|
|Date & Time:||25 Nov 2003 at 09:00|
|Event Info:||Grid applications such as the newly funded Linked Environments for Atmospheric Discovery (LEAD) project are rich with data streams generated by scientific instruments and sensors. These data streams form a collective resource of distributed time-based data to which clients need and want timely access. Our work, undertaken in the context of LEAD, undertakes to make these distributed data streams a recognized data resource in the architectural framework of GGF OGSI grid services.
In this talk we will introduce the main research objectives of LEAD, a collaboration between computer scientists, atmospheric scientists, and educators from 8 institutions. One objective of LEAD is to inject relevant real-time environmental data from sensors and NEXRAD radars into ongoing weather forecast models, so as to improve forecast accuracy. The task requires, among other tasks, synchronization between the streaming data and running models. This talk will introduce a new architecture that integrates a data streams system into the GGF OGSI grid services architecture. This is done by leveraging OGSA-DAI, the grid services based interface to databases developed in the UK. We point out where the research challenges exist.
Access to time-based data streams can be thought of as access to a distributed database where a request is a long-running database query that executes continuously. Our prototype, called dQUOB, implements this abstraction. We show in this talk how continuous query systems such as dQUOB provide intuitive access to a data stream resource.
Beth Plale is an Assistant Professor in the Computer Science Department at Indiana University. Prior to joining Indiana University, Dr. Plale held a Postdoc in the Center for Experimental Research and Computer Systems, College of Computing, Georgia Tech. Dr. Plale's Ph.D. is in computer science from State University of New York Binghamton. Her research interests are in distributed systems, middleware, data management, streaming data, and database query processing. Dr. Plale is a member of IEEE and ACM.