Multiple range query optimization with distributed cache indexing
Proceedings of the 2006 ACM/IEEE conference on Supercomputing
DiST: fully decentralized indexing for querying distributed multidimensional datasets
IPDPS'06 Proceedings of the 20th international conference on Parallel and distributed processing
Implementation of x-tree with 3d spatial index and fuzzy secondary index
FQAS'11 Proceedings of the 9th international conference on Flexible Query Answering Systems
Location-Aware Caching for Semantic-Based Image Queries in Mobile AD HOC Networks
International Journal of Multimedia Data Engineering & Management
| Hi-index | 0.00 |
Scientific applications that query into very large multi-dimensionaldatasets are becoming more common. Thesedatasets are growing in size every day, and are becomingtruly enormous, making it infeasible to index individualdata elements. We have instead been experimenting withchunking the datasets to index them, grouping data elementsinto small chunks of a fixed, but dataset-specific, sizeto take advantage of spatial locality. While spatial indexingstructures based on R-trees perform reasonably well forthe rectangular bounding boxes of such chunked datasets,other indexing structures based on KDB-trees, such as Hybridtrees, have been shown to perform very well for pointdata. In this paper, we investigate how all these indexingstructures perform for multidimensional scientific datasets,and compare their features and performance with that ofSH-trees, an extension of Hybrid trees, for indexing multi-dimensionalrectangles. Our experimental results show thatthe algorithms for building and searching SH-trees outperformthose for R-trees, R*-trees, and X-trees for both realapplication and synthetic datasets and queries. We showthat the SH-tree algorithms perform well for both low andhigh dimensional data, and that they scale well to high dimensionsboth for building and searching the trees.