Bo Kyeong Kim , Young Do Joo , Dong Ho Lee

The KIPS Transactions:PartD, Vol. 15, No. 6, pp. 755-766, Dec. 2008
10.3745/KIPSTD.2008.15.6.755,   PDF Download:

Abstract

Recently, NAND flash memory is becoming into the spotlight as a next-generation storage device because of its small size, fast speed, low power consumption, and etc. compared to the hard disk. However, due to the distinct characteristics such as erase-before-write architecture, asymmetric operation speed and unit, disk-based systems and applications may result in severe performance degradation when directly implementing them on NAND flash memory. Especially when a B-tree is implemented on NAND flash memory, intensive overwrite operations may be caused by record inserting, deleting, and reorganizing. These may result in severe performance degradation. Although μ-tree has been proposed in order to overcome this problem, it suffers from frequent node split and rapid increment of its height. In this paper, we propose Log-Structured B-Tree(LSB-Tree) where the corresponding log node to a leaf node is allocated for update operation and then the modified data in the log node is stored at only one write operation. LSB-tree reduces additional write operations by deferring the change of parent nodes. Also, it reduces the write operation by switching a log node to a new leaf node when inserting the data sequentially by the key order. Finally, we show that LSB-tree yields a better performance on NAND flash memory by comparing it to μ-tree through various experiments.

Statistics
Cite this article