An Analytic Model For Performance Evaluation Of Concurrency Control Algorithms

Abstract

An analytic model is proposed to explore performance of a concurrency control algorithm (CCA) of access method. Three dominant factors can indicate whether a CCA is better (or worse) than another in a multi-user environment. These factors reflect the methods for search and concurrency control of a CCA. To make a criterion for evaluating a CCA in term of these three factors, an analytic model is derived to formularize each value of each factor for the worst case. By the model, we can evaluate the performance of each CCA and identify which CCA has the best performance. To illustrate the applicability of the model, four CCAs for R-trees are used as examples to explain how to compute the values of these three factors of a CCA through the analytic model.