Principles of database buffer management
ACM Transactions on Database Systems (TODS)
A database disk buffer management algorithm based on prefetching
Proceedings of the seventh international conference on Information and knowledge management
A framework for modeling buffer replacement strategies
Proceedings of the ninth international conference on Information and knowledge management
SIGMETRICS '02 Proceedings of the 2002 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
Characterization of database access pattern for analytic prediction of buffer hit probability
The VLDB Journal — The International Journal on Very Large Data Bases
2Q: A Low Overhead High Performance Buffer Management Replacement Algorithm
VLDB '94 Proceedings of the 20th International Conference on Very Large Data Bases
Content-sensitive data prefetching
Content-sensitive data prefetching
Approximating the optimal replacement algorithm
Proceedings of the 1st conference on Computing frontiers
General adaptive replacement policies
Proceedings of the 4th international symposium on Memory management
The performance impact of kernel prefetching on buffer cache replacement algorithms
SIGMETRICS '05 Proceedings of the 2005 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
Hi-index | 0.00 |
Since operating systems (OSs) file systems are designed for a wide variety of applications, their performance may become suboptimal when the workload has a large proportion of certain atypical applications, such as a database management system (DBMS). Consequently most DBMS manufacturers have implemented their own file manager relegating the OS file system. This paper describes a novel page replacement strategy (Least Likely to Use) for buffer management in DBMSs, which takes advantage of very valuable information from the DBMS query planner. This strategy was implemented on an experimental DBMS and compared with other replacement strategies (LRU, Q2 and LIRS) which are used in OSs and DBMSs. The experimental results show that the proposed strategy yields an improvement in response time for most types of queries and attains a maximum of 97-284% improvement for some cases.