Adaptive adjustment of DBMS parameters based on influence diagram
Sun De-quan · Jisuanji gongcheng yu sheji · 2007
Adjusting the parameters or of DBMS manually in response to the changing workload is a difficult task.A probabilistic,graphical model known as influence diagram is introduced to adaptively adjust the parameters of DBMS.It divided into five steps: De-signing the structure of the model,collecting the experimental data,discretizing the data,learning the model's parameters,evaluating and using the model.The experiment using this methodology to simultaneously adjust two knobs from redo log subsystem of Oracle sho-wed that the influence diagram could effectively solve such problems.