On Native Location-Optimized Data Systems

With the ubiquity of location detection devices and location services and the massive amounts of the location data being produced, there is dire need for developing highly scalable location data systems. Currently, location data is usually supported as an afterthought in existing data systems, and hence these systems are not optimized natively for location data handling. This short paper reflects a lightning talk presented by the author. It calls for designing systems with location data being a first-class citizen, and highlights several important aspects that characterize location-optimized data systems. These include deterministic performance, supporting write-optimization, update tolerance, having adaptive location indexes and adaptive data stores, and having tight integration with the graph, document, relational and other multi data models for optimized performance.