ACM Transactions on Database Systems (TODS)
Redundancy in spatial databases
SIGMOD '89 Proceedings of the 1989 ACM SIGMOD international conference on Management of data
The R*-tree: an efficient and robust access method for points and rectangles
SIGMOD '90 Proceedings of the 1990 ACM SIGMOD international conference on Management of data
Reactive data structures for geographic information systems
Reactive data structures for geographic information systems
A class of data structures for associative searching
PODS '84 Proceedings of the 3rd ACM SIGACT-SIGMOD symposium on Principles of database systems
Safely and Efficiently Updating References During On-line Reorganization
VLDB '98 Proceedings of the 24rd International Conference on Very Large Data Bases
Generating Network-Based Moving Objects
SSDBM '00 Proceedings of the 12th International Conference on Scientific and Statistical Database Management
Research issues in automatic database clustering
ACM SIGMOD Record
| Hi-index | 0.00 |
An important goal for a spatial database system is to minimize the I/O-cost of queries and other operations. One essential component to achieve this objective is the buffer manager. The placement of the spatial objects on disk pages is another important factor; a reasonable clustering of the objects helps to minimize the I/O-cost of queries. However, it is a difficult task to define and maintain an efficient clustering. In this paper, a cluster manager is proposed, which re-clusters spatial objects dynamically. The reorganization is performed using the pages kept in main memory by the buffer manager. Therefore, no additional disk accesses are required for non-static spatial databases. For deciding which spatial objects should be stored together, the requests for the objects are recorded and analyzed. An experimental performance evaluation compares the impact of several parameter settings on different spatial que-ries. It is demonstrated that significant performance gains can be achieved by using the cluster manager.