Cost-based join processing scheme in a hybrid RDBMS and hive system

摘要

This paper introduces Honeybee, a hybrid Hive and RDBMS system, and describes experimental results related with the performance of its query processing. According to our experimental results, in the case of join queries, unlike other queries, query processing delay becomes a problem due to an increase of the size of output data. To solve this problem, we define a new scheme IRP. We also show how the previous scheme IHP and the IRP can be combined together by carefully choosing a threshold value to further improve the join query processing performance.