R-tree: A Hardware Implementation

R-tree data structures are widely used in spatial databases to store, manage and manipulate spatial information. As the data volume of such databases is typically very large, the query operation on R-tree data structure has an important impact on the performance of spatial databases. To boost the performance of R-tree search operations, we propose a new parallel R-tree search algorithm, which utilizes an adjacency matrix representation for an R-tree and performs the search using binary arithmetic among matrix elements. When the algorithm is implemented in hardware, we demonstrate through our simulation that the run-time complexity of the new algorithm is only bounded by the height of an R-tree. Furthermore, we find that a program with the proposed algorithm in hardware is 8.27X faster than the same program with its software counterpart in solving an example search problem. In the future, more research will be conducted to adapt the proposed algorithm with divideconquer paradigm to hand...

• Type: Conference
• Year: 2008
• Where: CDES
• Authors: Xiang Xiao, Tuo Shi, Pranav Vaidya, Jaehwan John Lee
• Keywords: Algorithm, CDES 2008, Hardware, R-tree Data Structure, Spatial Databases