MIDDLEWARE-BASED LOAD BALANCING FOR COMMUNICATING JAVA OBJECTS
DOI:
https://doi.org/10.47839/ijc.2.3.226Keywords:
Dynamic load balancing, load metrics, communication patternsAbstract
In the context of heterogeneous networks, like clusters of workstations, the design of programming and execution environments aims to automatically adapt execution to fluctuations that may appear in the execution of distributed and parallel Java applications. ADAJ, Adaptive Distributed Applications in Java, addresses this problem, dealing with both parallelism and distribution features. Ease of programming is achieved through an object and method parallelism paradigm. The trade-off between transparency of such a parallelism expression and efficiency is solved by application redeployment, meant to maintain a good performance level. This is the purpose of the load balancing in ADAJ, a dynamic and transparent tool at the middleware level, which exploits information issued from observation of the application, in order to consider both object activity and communication patterns. Communications generate attraction relations between objects and this article presents the evaluation of the load balancing mechanism for a type of asynchronous applications in which the communication aspect is important.References
A. Bouchi, R. Olejnik, and B. Toursel. A New Estimation Method for Distributed Java Object Activity. In IPDPS 2002 - Workshop on Java for Parallel and Distributed Computing, Fort Lauderdale, USA, 2002.
M. Phillippsen and M. Zenger. JavaParty - Transparent Remote Objects in Java. In ACM 1997 Workshop on Java for Science and Engineering Computation, Las Vegas, USA, June 1997.
A. Corradi, L. Leonardi, and F. Zambonelli. High-Level Directives to Drive the Allocation of Parallel Object-Oriented Applications. In Proceedings of HIPS’97, Amsterdam, Pays Bas, 1997.
K. Schloegel, G. Karypis, and V. Kumar. Graph Partititioning for High Performance Scientific Simulations. Technical Report: TR 00-018, Dept. of Computer Science and Engineering, University of Minnesota, 2000. To be included in CRPC Parallel Computing Handbook.
J. Arabe, A. Beguelin, B Lowekamp, E. Seligman, M. Starkey, and P. Stephan. Dome: Parallel programming in a heterogenous multi-user environment. Technical report, Carnegie Mellon University, Avril 1995.
M. Banatre, Y. Belhamissi, V. Issarny, I. Puaut, and J.P. Routeau. Adaptive Placement of Method Executions within a Customizable Distributed Object-Based Runtime System – Design, Implementation, and Performance. ISSN 1350-2042 TR 64, IRISA and CRIN-Nancy, 1994.
P. C Fishburn. A survey of multiattribute/ multicriteria evaluation theories. In S. Zionts, editor, Multicriteria problem solving, pages 181–224. Springer Verlag, Berlin, 1978.
V. Felea. Exploiting Runtime Information in Load Balancing Strategies. In P. Kacsuk and D. Kranzlmuller and Z. Nemeth and J. Volkert, editor, Distributed and Parallel Systems – Clu-ster and Grid Computing, pages 21–29, Linz, Austria, 2002. Kluwer Academic Publishers.
CWI - The National Research Institute for Mathematics and Computer Science in the Netherlands. Test Set for IVP Solvers. http://www.cwi.nl/ftp/IVPtestset/descrip.htm.
Downloads
Published
How to Cite
Issue
Section
License
International Journal of Computing is an open access journal. Authors who publish with this journal agree to the following terms:• Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
• Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
• Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.