Performance Evaluation of Data Migration Policies for a Distributed Storage System with Dynamic Tiering

Abstract

We have proposed a distributed storage system which autonomously migrates data blocks into a suitable storage node by considering their access frequency and read/write ratio. In this paper, we show that revised data migration policies can rapidly distribute input/output (I/O) hot spot over the storage nodes and can improve I/O performance of the entire system. Our simulation results show that the proposed method can reduce I/O execution time around 26% than the previous one at the maximum.