THE STUDAY A ORIENTED APPLICATION DATA ACCESS PATTERN ANALYSIS AND PREDICTION

Abstract

An innovative approach for assessing and forecasting object-oriented application runtime behavior in relation to data access patterns across their domain objects is presented in this study. Three different stochastic model implementations are used for the study and forecasts. Markov Chains, Importance Analysis, and Bayesian Inference provide the foundation of the models. The solution handles all required modifications to the target applications under examination in an entirely automated manner, eliminating the need for developer participation. Implementing the TPC-W and oo7 benchmarks validates the findings. Monte Carlo simulations have been used to simulate the oo7 benchmark as a stochas-tic process. It can be shown that our method yields accurate findings in relation to associated with collecting are minimal, between 5% and 9%. A wide range of object-oriented applications may be implemented using the system because to its sufficient flexibility.