Window query processing for joining data streams with relations

  • Authors:

  • Kristine Towne;Qiang Zhu;Calisto Zuzarte;Wen-Chi Hou

  • Affiliations:

  • The University of Michigan, Dearborn, MI;The University of Michigan, Dearborn, MI;IBM Toronto Laboratory, Markham, Ontario, Canada;Southern Illinois University, Carbondale, IL

  • Venue:
  • CASCON '07 Proceedings of the 2007 conference of the center for advanced studies on Collaborative research
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Query processing for data streams raises challenges that cannot be directly handled by existing database management systems (DBMS). Most related work in the literature mainly focuses on developing techniques for a dedicated data stream management system (DSMS). These systems typically either do not permit joining data streams with conventional relations or simply convert relations to streams before joining. In this paper, we present techniques to process queries that join data streams with relations, without treating relations as special streams. We focus on a typical type of such queries, called star-streaming joins. We process these queries based on the semantics of (sliding) window joins over data streams and apply a load shedding approximation when system resources are limited. A recently proposed window join approximation based on importance semantics for data streams is extended in this paper to maximize the total importance of the approximation result of a star-streaming join. Both online and offline approximation algorithms are discussed. Our experimental results demonstrate that the presented techniques are quite promising in processing star-streaming joins to achieve the maximum total importance of their approximation results.