Back to Search
Start Over
Approach to Accelerating Dissolved Vector Buffer Generation in Distributed In-Memory Cluster Architecture.
- Source :
-
ISPRS International Journal of Geo-Information . Jan2018, Vol. 7 Issue 1, p26. 20p. - Publication Year :
- 2018
-
Abstract
- The buffer generation algorithm is a fundamental function in GIS, identifying areas of a given distance surrounding geographic features. Past research largely focused on buffer generation algorithms generated in a stand-alone environment. Moreover, dissolved buffer generation is dataand computing-intensive. In this scenario, the improvement in the stand-alone environment is limited when considering large-scale mass vector data. Nevertheless, recent parallel dissolved vector buffer algorithms suffer from scalability problems, leaving room for further optimization. At present, the prevailing in-memory cluster-computing framework--Spark--provides promising efficiency for computing-intensive analysis; however, it has seldom been researched for buffer analysis. On this basis, we propose a cluster-computing-oriented parallel dissolved vector buffer generating algorithm, called the HPBM, that contains a Hilbert-space-filling-curve-based data partition method, a data skew and cross-boundary objects processing strategy, and a depth-given tree-like merging method. Experiments are conducted in both stand-alone and cluster environments using real-world vector data that include points and roads. Compared with some existing parallel buffer algorithms, as well as various popular GIS software, the HPBM achieves a performance gain of more than 50%. [ABSTRACT FROM AUTHOR]
- Subjects :
- *GEOGRAPHIC information systems
*COMPUTER workstation clusters
*HILBERT space
Subjects
Details
- Language :
- English
- ISSN :
- 22209964
- Volume :
- 7
- Issue :
- 1
- Database :
- Academic Search Index
- Journal :
- ISPRS International Journal of Geo-Information
- Publication Type :
- Academic Journal
- Accession number :
- 127700501
- Full Text :
- https://doi.org/10.3390/ijgi7010026