We propose a novel framework, called DRASTIC (Database FRAmework for SpatioTemporal Indexing in Clusters), for increasing the performance of storage and indexing when storing and accessing spatiotemporal datasets.
Spatiotemporal data is exceedingly important today, since a large number of modern devices (e.g., phones, tablets, sensor motes) are today equipped with GPS chips and have an internal clock. Massive amounts of spatiotemporal data are thus created on a daily basis in many contexts. However, manipulating and querying these large spatiotemporal datasets is today impractical, due to the inherent limitations of current database systems, which were not developed to store such data.
We propose a fundamentally new way of storing and querying spatiotemporal data using a new mathematical model and a dedicated indexing and storage system. Our mathematical modelling approach to the indexing of spatiotemporal datasets will lead to better efficiency when analysing such dataset, enabling real-time processing of large spatiotemporal data which is unfeasible on today’s data management platforms, and will in turn result in improved spatiotemporal data mining capabilities.
Collaboration: Prof. Philippe Cudré-Mauroux, University of Fribourg