Research on massive tile data management based on Hadoop

Abstract

Owing to the continuous development of GIS technology, massive GIS data is born at the right moment. In terms of the storage of massive GIS data, to guarantee the robustness of the system, the cloud storage has been gradually used. This paper focuses on the storage of massive tile data, by the use of Hadoop technology along with SequenceFile to merge multiple tile data into a big file to address the issue that Hadoop is unsuitable for storing multiple small files. During the process of writing the tile data into a big file, the Hilbert space-filling curve is also used to map the tile data from 2-dimensional to 1-dimensional in orde to use one-dimensional indexing to significantly improve the efficiency of later data retrieval.