Abstract—Distributed Hash Table (DHT) is a promising approach for a distributed data management platform in a large-scale network environment. In order to provide an efficient query processing and resource sharing in information discovery services, any DHT-based P2P system essentially needs to build an efficient indexing scheme for supporting users’ desired queries (complex query, i.e., multi-dimensional and/or range query). In this paper, multi-dimensional indexing scheme is built by generating multi-dimensional keys. For this purpose, k-dimensional tree (kd-tree) is proposed to be used in the indexing scheme. This paper focuses on the performance of kd-tree regarding to its splitting threshold value and evaluates it by using three metrics: (1) number of empty nodes on tree (2) number peers with data size zero, and (3) number of wrong labels. According to the simulated results, this paper defines that kd-tree is built at the splitting threshold value “200” and builds a multi-dimensional indexing scheme over Chord. This paper shows that how the proposed indexing scheme with the usage of kd-tree supports the complex query processing over Chord.
Index Terms—Indexing over DHT, complex query processing over DHT, DHT-based P2P query processing
The authors are with the University Computer Studies, Yangon, Myanmar (e-mail: yyimar@gmail.co, unghteinmaw@gmail.com, kmnweucsy@gmail.com).
[PDF]
Cite: Yi Yi Mar, Aung Htein Maw, and Khine Moe Nwe, "Usage of Kd-Tree in DHT-Based Indexing Scheme," International Journal of Future Computer and Communication vol. 2, no. 5, pp. 456-460, 2013.