A Hybrid B+tree Hash Index for Efficiency Improvement in a NAND Flash Memory
Hui-Jeong Ju, Sung-Je Cho · Advanced science and technology letters · 2015
Recently due to the development of wireless communication tech- niques, transaction processing within a mobile computing environment has been transferring to a memory based database from a disk based database. A memory based database, unlike a disk based database, uses up only a small capacity of memory and operates on a low-power environment; thus uses a NAND flash memory for fast transaction processing most of the time. This NAND flash memory has its strengths in its light weight, high-speed approaching velocity, and its durability. However due to the NAND flash memory's structure, it takes up lots of time in self-placement modifications and thus creates problems in performance deterioration due to its slow processing speed of writing, deleting, and modifying operations. In order to mend this performance deterioration problem, a model that reduced the maximum amount of problems by using split function of the overflow bucket extended hash index has been introduced. The problem of this model is that problems still exist regarding performance deteri- oration since the split function of the overflow bucket takes form of a sequential file processing. Thus in this said study, it seeks to introduce a hybrid B+tree hash model that has modified the preexisting problems of the existing model by designing a B+tree for the overflow bucket's sequential file processing. Also it has proven that the suggested model excels the performance of the preexisting model through a performance review.