GoFast: Graph-based optimization for efficient and scalable query evaluation
作者：
Highlights：
• We formalize a logical model describing the query execution of RDF systems.
• We present the essential statistics collected for each graph fragment.
• We detail a cost model comparing execution plans based on the disk and network costs.
• We present a study of the problem allowing to choose the optimal execution plan.
• We prove the problem’s complexity and provide a branch and bound solution.
• We present a new strategy to eliminate upstream not-relevant data fragments.
摘要
•We formalize a logical model describing the query execution of RDF systems.•We present the essential statistics collected for each graph fragment.•We detail a cost model comparing execution plans based on the disk and network costs.•We present a study of the problem allowing to choose the optimal execution plan.•We prove the problem’s complexity and provide a branch and bound solution.•We present a new strategy to eliminate upstream not-relevant data fragments.
论文关键词：Optimization,RDF,SPARQL,Cardinality estimation,Cost model
论文评审过程：Received 22 July 2020, Revised 16 December 2020, Accepted 7 February 2021, Available online 17 February 2021, Version of Record 26 February 2021.
论文官网地址：https://doi.org/10.1016/j.is.2021.101738
