Sciweavers

CASCON
2001

An analytical model for buffer hit rate prediction

13 years 5 months ago
An analytical model for buffer hit rate prediction
Of the many tuning parameters available in a database management system (DBMS), one of the most crucial to performance is the buffer pool size. Choosing an appropriate size, however, can be a difficult task. In this paper we present an analytical modeling approach to predicting the buffer pool hit rate that can be used to simplify the process of buffer pool sizing. Since the buffer replacement algorithm determines the buffer hit rate, we model the replacement algorithm which, in the case of DB2/UDB, is a variation of the GCLOCK algorithm. A Markov Chain model of GCLOCK is used to estimate the hit rate for a buffer pool. We evaluate the accuracy of the model's estimates with experiments carried out on DB2/UDB with the TPC-C benchmark. The model is validated for both single and multiple buffer pool cases.
Yongli Xi, Patrick Martin, Wendy Powley
Added 31 Oct 2010
Updated 31 Oct 2010
Type Conference
Year 2001
Where CASCON
Authors Yongli Xi, Patrick Martin, Wendy Powley
Comments (0)