Spatial Data Partitioning Method Based on Spatio-Temporal and Semantic Features

As the volume of spatial data is increasing, distributed storage has become an effective storage management strategy for large amounts of data. The core of distributed storage is the data partitioning method, which partitions massive data onto different nodes evenly. Based on Geohash and Hilbert curve, this paper proposes a spatial data partitioning method of spatial data, which considers the characteristics of spatiotemporal and semantics, and then stores the partitioned spatio-temporal data onto each node to balance the data volume of each node. Experiments show that the spatial data partitioning method proposed in this paper can effectively take into account the balance of data volume, spatio-temporal proximity and semantic similarity. Compared with traditional data partitioning methods, the proposed method achieves the balance of data onto various nodes, and its data blocks have better spatio-temporal proximity and semantic similarity at the same time.