Relational Database Support for Spatio-Temporal Data

• Author(s) / Creator(s)

  • Mallett, Daniel

• Technical report TR04-21. As computing technology evolves, the need to manage very large collections of complex data objects continues to arise. Database systems have been continually adapting to support new data-types, for instance, support for spatial data management is now standard in most Relational Database Management Systems (RDBMSs). With the proliferation of mobile devices capable of accurately reporting their position in time and space, there exists an urgent need to efficiently manage very large databases of moving object data inside a RDBMS. Currently, RDBMS support for such spatio-temporal data is limited and inadequate for non-trivial datasets, and most existing spatio-temporal access methods cannot be readily integrated into a RDBMS. This technical report proposes an approach for spatio-temporal storage, indexing and query support that can be fully integrated within any off-the-shelf RDBMS. A cost model which allows one to fine tune the proposed approach in order to minimize the number of disk accesses at query time is also presented. An implementation of the approach in Oracle 9i is compared with other alternatives to map the problem into a RDBMS. Extensive experiments show that the cost model is reliable and that the proposed approach significantly outperforms other options for managing spatio-temporal data inside a RDBMS. | TRID-ID TR04-21

• Date created

  2004

• Subjects / Keywords

  • RDBMS
  • Relational database management systems
  • Relational databases

• Type of Item

  Report

• DOI

  https://doi.org/10.7939/R3JH3D34C

• License

  Attribution 3.0 International