Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Object views: language support for intelligent object caching in parallel and distributed computations

Object views: language support for intelligent object caching in parallel and distributed... Object-based parallel and distributed applications are becoming increasingly popular, driven by the programmability advantages of component technology and a flat shared-object space. However, the flat shared-object space introduces a performance challenge: applications that rely on the transparent coherent caching of objects achieve high performance only on tightly coupled parallel machines. In distributed environments, the overheads of object caching force application designers to choose other solutions. Consequently, most applications sacrifice programmability, relying instead on either the explicit coherence management of cached objects, or on vastly different middleware abstractions such as multicast and events. In this paper, we describe object views — language support for efficient object caching in parallel and distributed computations. Object views specify restrictions on how computation threads can use an object, providing the underlying implementation with information about the potential side effects of object access, and thereby enabling construction of scalable, low-overhead caching protocols customized to application requirements. We present extensions to the Java programming language for expressing object views, and describe the design and implementation of a translator and run-time system for executing view-augmented Java programs on a distributed cluster of workstations. Experimental results based on a shared whiteboard application demonstrate that view-based object caching can achieve performance superior to multicast- and event-based implementations, while retaining essentially a shared object interface. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png ACM SIGPLAN Notices Association for Computing Machinery

Object views: language support for intelligent object caching in parallel and distributed computations

Download PDF
14 pages

Loading next page...
 
/lp/association-for-computing-machinery/object-views-language-support-for-intelligent-object-caching-in-Lkme20bnmW

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Association for Computing Machinery
Copyright
Copyright © 1999 by ACM Inc.
ISSN
0362-1340
DOI
10.1145/320385.320433
Publisher site
See Article on Publisher Site

Abstract

Object-based parallel and distributed applications are becoming increasingly popular, driven by the programmability advantages of component technology and a flat shared-object space. However, the flat shared-object space introduces a performance challenge: applications that rely on the transparent coherent caching of objects achieve high performance only on tightly coupled parallel machines. In distributed environments, the overheads of object caching force application designers to choose other solutions. Consequently, most applications sacrifice programmability, relying instead on either the explicit coherence management of cached objects, or on vastly different middleware abstractions such as multicast and events. In this paper, we describe object views — language support for efficient object caching in parallel and distributed computations. Object views specify restrictions on how computation threads can use an object, providing the underlying implementation with information about the potential side effects of object access, and thereby enabling construction of scalable, low-overhead caching protocols customized to application requirements. We present extensions to the Java programming language for expressing object views, and describe the design and implementation of a translator and run-time system for executing view-augmented Java programs on a distributed cluster of workstations. Experimental results based on a shared whiteboard application demonstrate that view-based object caching can achieve performance superior to multicast- and event-based implementations, while retaining essentially a shared object interface.

Journal

ACM SIGPLAN Notices – Association for Computing Machinery

Published: Oct 1, 1999

There are no references for this article.