A proposal to minimize the cost of processing big geospatial data in public cloud providers. Issue 3 (5th May 2021)

• Record Type: Journal Article
• Authors: Bachiega, João; Holanda, Maristela; Araujo, Aleteia P. F.
• Is Part Of: Transactions in GIS. Volume 25:Issue 3(2021)
• Page Start: 1599
• Page End: 1624
• Publication Date: 2021-05-05
• Subjects: Geographic information systems -- Periodicals; 910.285
• DOI: 10.1111/tgis.12754
• Languages: English
• ISSNs: 1361-1682

Abstract

Abstract: Spatial data represent abstractions of real‐world entities and can be obtained in various ways. They have properties that differentiate them from other types of data, such as a complex structure and dynamism. In recent years with the increasing volume of spatial data, referred to as "big geospatial data, " some tools have been developed to process such data efficiently, such as SpatialHadoop. The use of appropriate data indices based on the data set to be processed, as well as queries and operations to be performed, is essential for the optimal performance of these applications. In particular, since public cloud providers' charges are based on the resources used, it is imperative to optimize application execution in order to avoid unnecessary expense. This article proposes the use of six conditions that seek to minimize the cost of processing big geospatial data in public cloud providers. The tests performed demonstrate that the use of these conditions and the choice of the lowest‐cost provider can reduce the total processing cost by up to 41%.