Abstract: Traditional query processors generate full, accurate query results, either in batch or in pipelined fashion. Researchers reviewed that this strict model is too rigid for queries over distributed data sources. For executing such query, space is required in main memory to accommodate the data for query processing. When main memory gets full then some data must be moved to disk for further processing as part of flushing. An optimal strategy should be used to remove the victim data from memory to the disk. Victim data must be selected on the basis of its contribution in future query processing. In this study various join algorithms and flushing techniques used in online environment is reviewed.