A Trace-Based Simulation of Pointer Swizzling Techniques

Authors:
Mark L. McAuliffe;Marvin H. Solomon
Affiliations:
-;-
Venue:
ICDE '95 Proceedings of the Eleventh International Conference on Data Engineering
Year:
1995

Citing 0
Cited 6

HAC: hybrid adaptive caching for distributed storage systems

Proceedings of the sixteenth ACM symposium on Operating systems principles
Adaptable pointer swizzling strategies in object bases: design, realization, and quantitative analysis

The VLDB Journal — The International Journal on Very Large Data Bases - Persistent object systems
QuickStore: a high performance mapped object store

The VLDB Journal — The International Journal on Very Large Data Bases
Providing Persistent Objects in Distributed Systems

ECOOP '99 Proceedings of the 13th European Conference on Object-Oriented Programming
DIPS: an efficient pointer swizzling strategy for incremental uncaching environments

Journal of Systems and Software
Address translation strategies in the texas persistent store

COOTS'99 Proceedings of the 5th conference on USENIX Conference on Object-Oriented Technologies & Systems - Volume 5

Quantified Score

Hi-index	0.00

Visualization

Abstract

Persistent object-oriented applications that traverse large object graphs can improve their performance by caching objects in main memory while they are being used. While caching offers large performance benefits, the techniques used to locate these cached objects in memory can still impede the application's performance. We present the results of a trace-based simulation study of pointer swizzling techniques (techniques for reducing the cost of access to cached objects). We used traces derived from actual persistent programs to find a class of swizzling techniques that performs well, yet permits changes to the contents of in-memory object caches over the lifetime of an application. Our study demonstrates the superiority of a class of techniques known as "indirect swizzling" for a variety of workloads and system configurations.