TY - JOUR
T1 - Scalable spatio-temporal continuous query processing for location-aware services
AU - Xiong, Xiaopeng
AU - Mokbel, Mohamed F.
AU - Aref, Walid G.
AU - Hambrusch, Susanne E.
AU - Prabhakar, Sunil
PY - 2004
Y1 - 2004
N2 - Real-time spatio-temporal query processing needs to effectively handle a large number of moving objects and continuous spatio-temporal queries. In this paper, we use shared execution as a mechanism to support scalability in location-aware servers. Our main idea is to maintain a query table that stores information about continuous spatio-temporal queries. Then, answering spatio-temporal queries is abstracted as a spatial join among the moving objects and queries. Three query join policies are proposed aiming to minimize the cost of the join operation under the shared execution paradigm, namely the Clock-triggered Join Policy, the Incremental Join Policy, and the Hot Join Policy. We introduce the concept of a No-Action Region that is used in conjunction with the hot join policy. We propose algorithms that calculate the No-Action region for objects and queries. Experimental performance demonstrates that the No-Action region is more efficient than other approaches when used along with the hot join policy. Experiments also demonstrate that the hot join policy outperforms the clock-triggered join policy and the incremental join policy in terms of both I/O and CPU costs.
AB - Real-time spatio-temporal query processing needs to effectively handle a large number of moving objects and continuous spatio-temporal queries. In this paper, we use shared execution as a mechanism to support scalability in location-aware servers. Our main idea is to maintain a query table that stores information about continuous spatio-temporal queries. Then, answering spatio-temporal queries is abstracted as a spatial join among the moving objects and queries. Three query join policies are proposed aiming to minimize the cost of the join operation under the shared execution paradigm, namely the Clock-triggered Join Policy, the Incremental Join Policy, and the Hot Join Policy. We introduce the concept of a No-Action Region that is used in conjunction with the hot join policy. We propose algorithms that calculate the No-Action region for objects and queries. Experimental performance demonstrates that the No-Action region is more efficient than other approaches when used along with the hot join policy. Experiments also demonstrate that the hot join policy outperforms the clock-triggered join policy and the incremental join policy in terms of both I/O and CPU costs.
UR - http://www.scopus.com/inward/record.url?scp=5444220068&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=5444220068&partnerID=8YFLogxK
M3 - Conference article
AN - SCOPUS:5444220068
SN - 1099-3371
VL - 16
SP - 317
EP - 326
JO - Proceedings of the International Conference on Scientific and Statistical Database Management, SSDBM
JF - Proceedings of the International Conference on Scientific and Statistical Database Management, SSDBM
T2 - Proceedings - 16th International Conference on Scientific and Statistical Databse Management, SSDBM 2004
Y2 - 21 June 2004 through 23 June 2004
ER -