LSShare: an efficient multiple query optimization system in the cloud

Xing Ge, Bin Yao, Minyi Guo, Changliang Xu, Jingyu Zhou, Chentao Wu, Guangtao Xue

Abstract: Multiple query optimization (MQO) in the cloud has become a promising research direction due to the popularity of cloud computing, which runs massive data analysis queries (jobs) routinely. These CPU/IO intensive analysis queries are complex and time-consuming but share common components. It is challenging to detect, share and reuse the common components among thousands of SQL-like queries. Previous solutions to MQO, heuristic or genetic based, are not appropriate for the large growing query set situation. In this paper, we develop a sharing system called LSShare using our proposed Lineage-Signature approach. By LSShare, we can efficiently solve the MQO problem in a recurring query set situation in the cloud. Our system has been prototyped in a distributed system built for massive data analysis based on Alibaba’s cloud computing platform ( http://www.alibaba.com/ ). Experimental results on real data sets demonstrate the efficiency and effectiveness of the proposed approach.