An End-Around Approach for Efficient Join Query Processing

Abstract

This paper introduced a method for producing immediate and result in multi-join query, in homogeneous and heterogeneous environment. In recent years Adaptive or Non Blocking join algorithms have attracted a lot of attention in streaming applications, where data is provided from autonomous data sources in heterogeneous network environments. This algorithms are better as compared to traditional algorithms is that they can generates join results as early as the first input tuples are on hand hence it improves pipelining, smooth out join result production and also masking source or network delays. As response time of the queries places a very important role in adaptive join, the join algorithm like Hash Join, Sort Merge Join are become unacceptable for this environment because they require preprocessing before generating the join result. Hence, in adaptive join technique only possible algorithm is Nested loop join. In Nested Loop Join, every single record of the outer relation is compared with every single record of the inner relation. The no. of comparisons done by the nested loop join can be reduced by making improvement in Block Nested loop Join. In proposed End-Around Block Nested loop join outer and inner table’s comparison is done in parallel and whenever a row in first location didn’t find a match then row from first location removed and placed at rear end as like in a queue, the matched row removed from inner relation and added to result set. Whenever, New tuple arrive is then pushed into rear end and process is continuing with new incoming tuples in streaming environments.