Multidimensional Cloud data index based on KD tree and R-Tree

Multidimensional Cloud data index based on KD tree and R-Tree

He Yi Yue Yangfan Yun Chunrezhouvi

For the cloud storage system, most of the data are stored based on the key value to the Key,value model, the multidimensional query needs to scan the whole dataset completely, and the query efficiency is low, a multidimensional index structure (called Kd-r Index) based on KD tree and R tree is proposed. The KD-R index adopts the double index model, establishes the multidimensional global index based on KD tree on the global server, and constructs the R-Tree multidimensional local index in the local data node. Based on the performance loss model, the R-tree node with lower index cost is published to the global KD tree to optimize multidimensional query performance. Experimental results show that compared with the global distributed R-Tree index, the KD-R index can improve the performance of multidimensional query effectively, and the KD-R index has high availability in the case of server node failure.

Multidimensional Cloud data index based on KD tree and R-Tree