An effective data clustering method based on expected update time in flash memory environment

Abstract

Flash memory has its unique characteristics: The write operation is much more costly than the read operation, and in-place updating is not allowed. In flash memory environment, in order to reduce the cost of copying valid pages during an erase operation, hot data clustering methods have been proposed. They try to store data with high write frequency together into the same block. In this paper, we first analyze the fundamental problem of existing hot data clustering methods. Based on this analysis, we propose an effective method for data clustering in flash memory environment. The proposed method tries to store data having similar expected update times together in the same block, thereby reducing the cost of copying valid pages significantly. For performance evaluation, we conduct extensive experiments. The results show that our method achieves speed-up by up to 1.7 times compared with existing one.