Parallel map projection of vector-based big spatial data: Coupling cloud computing with graphics processing units

Abstract

The objective of this article is to present a framework that couples cloud and high-performance computing for the parallel map projection of vector-based big spatial data. The past few years have witnessed a tremendous growth of a variety of high-volume spatial data—i.e., big spatial data. Map projection is often needed, for example, when we apply these big spatial data into large-scale spatial analysis and modeling approaches that require a common coordinate system. However, due to the size of these data and algorithmic complexity of map projections, the transformation of big spatial data between alternative projections represents a pressing computational challenge. Recent advancement in cloud computing and high-performance computing offers a potential means of addressing this computational challenge. The parallel map projection framework presented in this study is based on a layered architecture that couples capabilities of cloud computing and high-performance computing accelerated by Graphics Processing Units. We use large LiDAR data as an example of vector-based big spatial data to investigate the utility of the parallel map projection framework. As experimental results reveal, the framework provides considerable acceleration for re-projecting vector-based big spatial data. Coupling high-performance and cloud computing, which complement to each other, is a suggested solution for the efficient processing and analysis of big spatial data.