Processamento distribuído da junção espacial de múltiplas bases de dados: multi-way spatial join

• Autor(a) principal: Cunha, Anderson Rogério
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFG
• Texto Completo: http://repositorio.bc.ufg.br/tede/handle/tede/3836
• topic: Processamento distribuído; Junção espacial; Multi-way spatial join; R-Tree; Distributed processing; Spatial join; CIENCIA DA COMPUTACAO::SISTEMAS DE COMPUTACAO

Spatial join is one of the spatial operations of higher computational cost. Its complexity increases significantly when it involves multiple databases (multi-way spatial join). Traditional processing strategies of multi-way spatial join apply combinations of binary join algorithms on centralized computing environments. For complex queries, this approach requires much computational power, making it often unfeasible in centralized environments. This work proposes the Distributed Synchronous Traversal algorithm (DST), whose goal is to enable the distributed processing of multi-way spatial joins on a cluster of computers. The DST algorithm is based on Synchronous Traversal algorithm and processes the multiway spatial join in a single synchronous descent upon R-Trees levels of the database entries (the final outcome is built incrementally, without creating temporary databases). To the best of our knowledge, there are no other proposals in the literature that deal with this problem in a distributed fashion and on a peer-to-peer architecture. Many challenges had to be overcome, such as the definition of data structures that enabled the mapping of the semantics of queries of multi-way spatial join and coordination of the required distributed processing. DST proved to be satisfactorily parallelizable and scalable process real datasets in experiments performed in clusters of 1, 2, 4 and 8 servers.