Causality join query processing for data stream by spatio-temporal sliding window

시공간 슬라이딩윈도우기법을 이용한 데이터스트림의 인과관계 결합질의처리방법

  • 권오제 (부산대학교 컴퓨터공학과) ;
  • 이기준 (부산대학교 정보컴퓨터공학부)
  • Published : 2008.07.31


Data stream collected from sensors contain a large amount of useful information including causality relationships. The causality join query for data stream is to retrieve a set of pairs (cause, effect) from streams of data. A part of causality pairs may however be lost from the query result, due to the delay from sensors to a data stream management system, and the limited size of sliding windows. In this paper, we first investigate spatial, temporal, and spatio-temporal aspects of the causality join query for data stream. Second, we propose several strategies for sliding window management based on these observations. The accuracy of the proposed strategies is studied by intensive experiments, and the result shows that we improve the accuracy of causality join query in data stream from simple FIFO strategy.