Load Shedding for Aggregation Queries over Data Streams

  • Authors:
  • Brian Babcock;Mayur Datar;Rajeev Motwani

  • Affiliations:
  • -;-;-

  • Venue:
  • ICDE '04 Proceedings of the 20th International Conference on Data Engineering
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Systems for processing continuous monitoring queriesover data streams must be adaptive because data streamsare often bursty and data characteristics may vary overtime. In this paper, we focus on one particular type ofadaptivity: the ability to gracefully degrade performancevia "load shedding" (dropping unprocessed tuples to reducesystem load) when the demands placed on the systemcannot be met in full given available resources. Focusingon aggregation queries, we present algorithms that determineat what points in a query plan should load sheddingbe performed and what amount of load should be shed ateach point in order to minimize the degree of inaccuracyintroduced into query answers. We report the results of experimentsthat validate our analytical conclusions.