An empirical study of transaction throughput thrashing across multiple relational DBMSes
作者：
Highlights：
•
• We propose a novel structural causal model to explain the origins of DBMS thrashing.
• We present a space of hypotheses that can refine and support the proposed model.
• We extend a recent DBMS-centric research infrastructure for thrashing experiments.
• We propose a novel analysis protocol that can be applied to obtained thrashing data.
• We suggest engineering implications to DBMS developers and DBAs for better performance.
摘要
Highlights •We propose a novel structural causal model to explain the origins of DBMS thrashing.•We present a space of hypotheses that can refine and support the proposed model.•We extend a recent DBMS-centric research infrastructure for thrashing experiments.•We propose a novel analysis protocol that can be applied to obtained thrashing data.•We suggest engineering implications to DBMS developers and DBAs for better performance.
论文关键词：DBMS Thrashing,Transaction,Throughput,Factors,Structural Causal Model,Empirical Study
论文评审过程：Received 17 November 2015, Revised 8 August 2016, Accepted 21 December 2016, Available online 27 December 2016, Version of Record 27 April 2017.
论文官网地址：https://doi.org/10.1016/j.is.2016.12.004
