Abstract: Data partitioning is a challenge in the distributed execution of multiway spatial join queries. An efficient execution requires a balanced data distribution in the cluster computers as well as a distribution that maintains spatial data colocalization. In this monograph, two spatial data distribution methods were compared: Round-Robin and Proximity Area, and a new one was proposed, called Gain-Loss, based on the R0-tree algorithms. Our experiments in a controlled environment, using synthetic datasets, show that the new method presented a very reduced area overlap between servers in all scenarios and also a regular object balancing. This result indicates a more efficient execution of queries, due to a reduction in the computational resources usage, mainly network usage and processing time.

Keywords: Multiway Spatial Join; Data Distribution; Gain-Loss.

Complete monograph. Copyright © 2018. All rights reserved.

Disclaimer: Although the student carefully wrote the original abstract, and it was revised and improved, English is not him or the advisor' mother language. The original work is written in Portuguese.

Citation: Guilherme Silva Tonon. Métodos de Distribuição de Dados para Processamento Distribuído de Multijunções Espaciais. Monograph. Bacharelado em Ciências da Computação. Universidade Federal de Goiás, Regional Jataí. Jataí, GO, Brasil. 2018. 47p.

Copy citation in bibtex format.