Query cost estimation for read intensive flash memory based database systems

Abstract

Flash memory database systems are relatively new area of research interest in computing community. Most database techniques, already developed and proposed, are based on the assumption that the database is hard disk resident, due to this fact the cost measurement of query algorithms is highly dependent on the mechanical nature of hard disks. But the read, write, erase behavior of flash memories are radically different from hard disks, so that new measures of query costs that are tuned by the characteristics of flash memories should be introduced. In this paper we propose a new query cost measurement procedure for read intensive queries like join. We then compare and contrast already developed join algorithms with the new cost measurement.