This project, acquiring and deploying infrastructure for several data-intensive stream processing applications, supports the following core projects:
-ETL: High-performance data warehouse load and maintenance strategies using innovative memory-resident pipelined processing techniques; -D-CAPE: Continuous stream processing technologies supporting on-line plan adaptation and optimization strategies coordinating run-time query workload re-distribution versus disk spilling; -RAINDROP: Scalable Xquery processing over XML streams offering novel advances in automata-style processing, in stream-specific multi-query optimization, and in XML-specific load shedding.
With expectations of break-throughs in scalable stream handling, the work impacts applications that require real-time stream monitoring, such as monitoring for changes in environmental conditions (e.g., smoke in buildings, medical conditions that require attention or medication, and critical military government, etc.). The project also deals with the challenges and opportunities raised by portable computers, wireless communications, and high bandwidth, environments where high volumes (in the order of tera-bytes) of stream data sources may be present anywhere and at any instant.
Broader Impact: The infrastructure enhances the integration of research and education. WPI is known for its project-oriented undergraduate education, where all undergraduates must conduct a year-long senior level research project, the Major Qualifying Project (MQP). The cluster services these students as well as the faculty, and graduate students.
