Towards High Performance Spatio-temporal Data Management Systems

The volume of spatio-temporal data is growing at a rapid pace. This is driven by several factors, including the widespread adoption of GPS-enabled mobile devices and the proliferation of RFID-tagged objects in sensor networks. Besides the volume, such spatio-temporal data is characterized by high "velocity", with its high rate of time-stamped location updates. The rise of spatio-temporal "Big data" has led to the emergence of many novel location-oriented applications. These applications often have complex use-cases and service-level requirements. Efficient management of the spatio-temporal data is critical to meet these requirements. This poses some challenges and unique research questions, for instance: i) how to support the high rate of location updates, while at the same time supporting many concurrent historical, present and predictive queries; ii) what kind of database storage organization is suitable for such workload; iii) what are the implications for the spatio-temporal index; and iv) what kind of novel spatio-temporal queries are to be supported. Technological trends involving increasingly large main memory sizes and core counts offer opportunities to address some of these issues. We have addressed a few issues pertinent to high performance commercial Location-Based Services (LBS) by exploiting in-memory database techniques. We propose an in-memory storage organization for high insert performance and introduce a novel spatio-temporal index. With extensive evaluation, we demonstrate that our system supports high insert and query throughputs and it outperforms the leading LBS system by a significant margin. As part our future research we are building a spatio-temporal data management system in the context of a cluster of machines in the Cloud. We are also investigating the possibility of supporting trajectory-based join queries.