PARMI: a publish/subscribe based asynchronous RMI framework for cluster computing

Authors:
Heejin Son;Xiaolin Li
Affiliations:
Scalable Software Systems Laboratory, Computer Science Department, Oklahoma State University, Stillwater, OK;Scalable Software Systems Laboratory, Computer Science Department, Oklahoma State University, Stillwater, OK
Venue:
HPCC'07 Proceedings of the Third international conference on High Performance Computing and Communications
Year:
2007

Citing 6
Cited 1

The Information Bus: an architecture for extensible distributed systems

SOSP '93 Proceedings of the fourteenth ACM symposium on Operating systems principles
Designing and Building Parallel Programs: Concepts and Tools for Parallel Software Engineering

Designing and Building Parallel Programs: Concepts and Tools for Parallel Software Engineering
Optimizing Remote Method Invocation in Java

DEXA '02 Proceedings of the 13th International Workshop on Database and Expert Systems Applications
Coupling Dynamic Load Balancing with Asynchronism in Iterative Algorithms on the Computational Grid

IPDPS '03 Proceedings of the 17th International Symposium on Parallel and Distributed Processing
The many faces of publish/subscribe

ACM Computing Surveys (CSUR)
Distributed Programming with Typed Events

IEEE Software

Publish/subscribe as a model for scientific workflow interoperability

Proceedings of the 4th Workshop on Workflows in Support of Large-Scale Science

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents a publish/subscribe based asynchronous remote method invocation framework (PARMI) aiming to improve performance and programming flexibility. PARMI enables high-performance communication among heterogeneous distributed processors. Based on publish/subscribe paradigm, PARMI realizes asynchronous communication and computation by decoupling objects in space and time. The design of PARMI is motivated by the needs of scientific applications that adopt asynchronous iterative algorithms. An example scientific application based on the Jacobi iteration numerical method is developed to verify our model and evaluate the system performance. Extensive experimental results on up to 60 processors demonstrate the significant communication speedup using asynchronous computation and communication technique based on the PARMI framework compared to a baseline scheme using synchronous iteration and communication.