High-volume data streams exchanged in XML form are becoming crucial to a wide range of applications in the scientific, government and business domains. However, taking advantage of them raises significant technical challenges and opportunities. First, stream queries require computations to be performed in memory that is limited with comparison to the arbitrarily large data stream. Second, the rapid increase of network bandwidths surpasses the increase of CPU-to-memory and CPU-to-disk transfer rates. The result is that high bandwidth networks will be delivering stream data faster than they can be processed by conventional approaches. Efficient on-the-fly data stream processing is necessary.
This project will develop a comprehensive approach to querying data streams in XML form, covering the spectrum from formal foundations all the way to systems. The expected outcome of the project will be a qualitatively new architecture, technology and query processor for XML data streams, with impact on a wide range of applications. The approach will be tested on seismic data provided by the ROADNet project. However, we expect that the research will have broader impact on voluminous XML data streams.
