WITH rapid advances in electronics miniaturization, wireless communication and positioning technologies, the acquisition and transmission of spatio-temporal data using mobile devices are becoming pervasive. This fuels the demand for location-based services (LBS).

In location-aware or location-based services (LBS), managing and processing spatio-temporal data are the central problems. Existing RDBMS have not been built to meet spatio-temporal requirements such as fast and frequent updates (due to agility of moving objects), moving queries vs moving objects, and sharing of information among moving objects. In this project, we intend to design various techniques for providing effective and efficient database support for LBS.

We have built an operable system (version 1) for moving objects on the top of a popular relational database system MySQL. By making use of the well tested and robust B+-tree available in the MySQL, our implementation enables this relational database engine to effectively store moving object information after transformation. The indexes provide the most effective means in reducing the search cost and the overall query processing cost. The B-link tree provides a good degree of concurrency for the B+-tree, and unlike the R-tree based indexes, our indexes do not incur high update overhead (and hence lock contention)! Query processing strategies are subsequently developed around these indexes. User-specified spatial-temporal queries are converted into standard SQL statements which are efficiently processed in the relational engine. Most importantly, all these have been achieved neatly and cost effectively without altering too much of MySQL core.

funded by A*STAR.