Research Of Oracle Performance's Prognostication & Analyse

Prognostic Model is most important in the fild of Oracle performance tunning. As using Prognostic Model,the resonsibilities of Oracle DBA are to get the existed statistical data and to prognosticate the future requires in varous field of Oracle. The paper's author have analysised the temporal and spacial distribution of 10 whcih was done on data file of system and he use the funcion of ARCOFU proccess to have getten the trendency of 10 information. Afterward, the autor hasten forther imformaion and got the prognostication when to and how to modulate the 10 to maintain the excellence performance of the system.Meanwhile, Oracle DBA have the capablity to anaysis the ditails of the Oracle data cache (Keep Pool, DEFAL'LT Pool and RECYCLE pool) and to accurately measure the performance of those cache pools in a span of time. Based on the existed data, Oracle DBA can accurately forecaste when to add RAM to those data cache in order to maintain the current performance of the system.Using Prognostic Model, Oracle DBA have the chance to use the data in the table collected by the performance-data-collecting system to divide information based the previously various paremeters, Generally speaking, classification based on decision tree or neural network and prediction based on regression model are used widely. But they are not usually satisfied with the need of practical problem. Decision tree is often applied for predicting the class label of data objects and regression model is usually applied in the course of linear problem. With the constantly expanding applied field of data mining, new models and arithmetic continually appear and are used for analyzing and solvingproblem as the complementarities of old and mature models and arithmetic. This article brings forward a new prognostic model, which is based on N_rank transfer probability coming from the theory of random process and restricted by minimum confidence and adjustment matrix. Using this model, ultimately Oracle performance prognostic result can be found.Making Gs system be researchful object and applying Prognostic Model, the author has prognosticated the future performance data based on the real historical data. Distilling effective data from dbselect database and building data mining database; According to the idea of prognostication, computing data in order to satisfy the need of prognostication and using minconf to judge, finally getting the result. Because of the objective factor, the result must be adjust by adjustment matrix. According to prognostic result, it is possible that oracle resource are arranged in reason in the course of resource plan and eventually Oracle performance can be satisfied with system's requirement.The paper has instructional purport for Oracle resource' s adequate exertion.