Middleware infrastructure for parallel and distributed programming models in heterogeneous systems

Abstract

In this paper, we introduce a middleware infrastructure that provides software services for developing and deploying highperformance parallel programming models and distributed applications on clusters and networked heterogeneous systems. This middleware infrastructure utilizes distributed agents residing on the participating machines and communicating with one another to perform the required functions. An intensive study of the parallel programming models in Java has helped identify the common requirements for a runtime support environment, which we used to define the middleware functionality. A Java-based prototype, based on this architecture, has been developed along with a Java Object-Passing Interface (JOPI) class library. Since this system is written completely in Java, it is portable and allows executing programs in parallel across multiple heterogeneous platforms. With the middleware infrastructure, users need not deal with the mechanisms of deploying and loading user classes on the heterogeneous system. Moreover, details of scheduling, controlling, monitoring, and executing user jobs are hidden, while the management of system resources is made transparent to the user. Such uniform services are essential for facilitating the development and deployment of scalable high-performance Java applications on clusters and heterogeneous systems. An initial deployment of a parallel Java programming model over a heterogeneous, distributed system shows good performance results. In addition, a framework for the agents' startup mechanism and organization is introduced to provide scalable deployment and communication among the agents. Index Terms-Distributed systems middleware, parallel programming models, parallel and distributed Java, cluster, heterogeneous systems, distributed agents.

A Framework for Distributed Dynamic Load Balancing in Heterogeneous Cluster

“…MA support load balancing in parallel and distributed computing [5,9] , e.g., Traveller [31] using resource broker. It implements parallel application such as L. U. Factorization and sorting.…”

“…Furtermore, the agent uses a path through a network based upon known infrastructure characteristics (QoS). An agent optimizes its transmission between Agent hosts (AHs) [4] with the help of several migration strategies described in [5] .…”

“…Other areas where MAs are seen as offering potential advantages-wireless or mobile computing [3,4] dynamic deployment of code, thin clients or resource limited devices, personal assistants and MA-based parallel processing [5,9] . The idea of using MA in load balancing has been floating around for sometime in homogeneous telecommunication networks because traditional load balancing approaches are implemented based on message passing paradigm [1,10] .…”

Load Balancing on Open Networks: A Mobile Agent Approach

“…A framework for load balancing using MA named EALBMA (Efficient and Adaptive Load Balancing based on MA) [28] has been made in which a novel algorithm for updating load information partially based on MA called ULIMA.MA support load balancing in parallel and distributed computing [7,8] , e.g., Traveller [22] using resource broker. It implements parallel application such as L. U. Factorization and sorting.…”

“…MA supports a variety of webbased distributed applications namely: systems and distributed information management [3] and information retrieval [4] . Other areas where MAs are seen as offering potential advantages-wireless or mobile computing [5,6] dynamic deployment of code, thin clients or resource-limited devices, personal assistants, and MA-based parallel processing [7,8] . Traditional load balancing approaches on distributed web servers are implemented based on message passing paradigm [1,24] .…”