An improved Hilbert curve for parallel spatial data partitioning
Lingkui Meng, Changqing Huang, Chunyu Zhao, Zhi-Yong Lin · Geo-spatial Information Science · 2007
A novel Hilbert-curve is introduced for parallel spatial data partitioning, with consideration of the huge-amount property of spatial information and the variable-length characteristic of vector data items. Based on the improved Hilbert curve, the algorithm can be designed to achieve almost-uniform spatial data partitioning among multiple disks in parallel spatial databases. Thus, the phenomenon of data imbalance can be significantly avoided and search and query efficiency can be enhanced.