Predicting Future Page Access by Analysing Object Relationships


Authors: Nils Knafla
Date: December 1997
Type: Technical Report, ECS-CSG-35-97
Institution: Department of Computer Science, Edinburgh University

Abstract

In this report we present a new approach to predicting page access probability by considering the structure of the relationships between objects. Database objects and their relationships to other objects are modelled by a discrete-time Markov Chain. We present two approaches to compute the page access probability: (a) Using Hitting Times and Absorption Probabilities and (b) Using the Chapman-Kolmogorov Equations. If the probability of a page is higher than a threshold defined by cost/benefit parameters then the page is a candidate for prefetching. To determine the prefetching threshold we consider various cost parameters to compare the benefit of a correct prefetch with the cost of an incorrect prefetch.


Nils Knafla, <nk@dcs.ed.ac.uk>