Monitoring and Mining Data Streams
Abstract
This work is preliminary since funding for this piece was only received 6 months ago. We have so far managed to complete a revision of the Borealis code base making it much more usable for general applications. We have worked closely with personnel from USARIEM to identify special processing needs for PAN's in a WPSM setting and have redesigned major portions of the Borealis code base to reflect this. It takes a novel position with respect to dealing with failure in a sensor network. The Aurora stream processing engine was a very valuable exercise to gain understanding about the basic technical questions that must be addressed by any stream processing engine. These issues included memory management, tuple scheduling, and load control. Aurora was designed to run on a single server, and its primary optimization goal was low-latency processing. This set of assumptions was chosen because it gave us the best opportunity to study the fundamentals, and it was useful for a large class of monitoring applications.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 2004
- Accession Number
- ADA431589
Entities
People
- Stanley B. Zdonik
Organizations
- Brown University