Building a high performance ArcGIS cluster and its application to climate data processing

1. Introduction

ArcGIS Desktop provides an integrated GIS, combining object-oriented and traditional file-based data models with a set of tools to create and work with geographic data.

Currently, ArcGIS Desktop software contains functions for mapping, data manipulation, data management, as well as a geoprocessing framework, which allows users to choose among the geoprocessing functions in a variety of ways to build extensions and toolboxes.

From the early efforts by Openshaw and Abrahart 2000, high performance computing (HPC) has been adopted to address computational and data intensive geospatial problems.

The integration of HPC and ArcGIS would create new capability and new possibilities of tackling challenging research questions. Herein, we report our effort of establishing a parallel ArcGIS environment on a computing cluster. First, we present the major components of our high performance computing environment. We then describe the configuration of ArcGIS in the cluster. Finally, we present a case study to demonstrate the parallel ArcGIS capability and performance. We hope those ideas can inspire further developments and deployment of ArcGIS on HPCs.