Memory Flushing in Multijoin Query for Query Optimization

The enhancement of internet technology and advances in heterogeneous databases produced the requirement of database style querying over data from sources distributed around the world. When executing an online join query space is required in main memory to accommodate the data required to process the query. Memory of a specified amount is allotted to the query plan. When the main memory becomes full some data is required to move to disk to make room for new incoming tuples. Moving data from memory to disk is called Flushing. An optimal flushing strategy should be used to select the victim data so that the data which will least contribute in the future results would be flushed. All the previous algorithms are focused on flushing the memory when it becomes full but in this case some incoming data may be lost during the duration of flushing. In this paper we will propose a novel algorithm EarlyFlush to flush the memory before it becomes full to prevent the data loss and increase the throughput.