R-Tree for Phase Change Memory

Computer Science and Information Systems

Elkhan Jabarov¹  Gyu Sang Choi²  Byung-Won On³ 

[1] Computer and Radio Communications Department, Korea University;Department of Information & Communication Engineering, Yeungnam University;Department of Statistics and Computer Science, Kunsan National University

关键词: spatial database;    spatial data;    PCM;    R -Tree;    spatial tree;    endurance;    indexing algorithm;   

DOI  :  10.2298/CSIS160620008J

学科分类：社会科学、人文和艺术（综合）

来源: Computer Science and Information Systems

【 摘 要 】

Nowadays, many applications use spatial data for instance-location information, so storing spatial data is important.We suggest using R-Tree over PCM. Our objective is to design a PCM-sensitive R-Tree that can store spatial data as well as improve the endurance problem. Initially, we examine how R-Tree causes endurance problems in PCM, and we then optimize it for PCM. We propose doubling the leaf node size, writing a split node to a blank node, updating parent nodes only once and not merging the nodes after deletion when the minimum fill factor requirement does not meet. Based on our experimental results while using benchmark dataset, the number of write operations to PCM in average decreased by 56 times by using the proposed R -Tree. Moreover, the proposed R-Tree scheme improves the performance in terms of processing time in average 23% compared to R-Tree.

【 授权许可】

CC BY-NC-ND