Efficient, flexible, and typed group communications in Java

Abstract

Group communication is a crucial feature for high-performance and Grid computing. While previous works and libraries proposed such a characteristic (e.g. MPI, or object-oriented frameworks), the use of groups imposed specific constraints on programmers -for instance the use of dedicated interfaces to trigger group communications.We aim at a more flexible mechanism. More specifically, this paper proposes a scheme where, given a Java class, one can initiate group communications using the standard public methods of the class together with the classical dot notation; in that way, group communications remains typed. Furthermore, groups are automatically constructed to handle the result of collective operations, providing an elegant and effective way to program gather operations. This flexibility also allows to handle results that are groups of remotely accessible objects, and to use a group as a means to dispatch different parameters to different group members (for instance in a cyclic manner). Furthemore, hierarchical groups can be easily and dynamically constructed; an important feature to achieve the use of several clusters in Grid computing.Performance measures demonstrate the viability of the approach. The challenge is to provide easy to use, efficient, and dynamic group management for objects dynamically distributed on the Grid.

Grid’BnB: A Parallel Branch and Bound Framework for Grids

“…Method calls sent to active objects are asynchronous with transparent future objects and synchronization is handled by a mechanism known as wait-by-necessity.ProActive provides typed group communication, an important feature for high-performance and grid computing. The group communication [7] extends the ProActive elementary mechanism for asynchronous remote method invocation and automatic futures.…”

Programming, Composing, Deploying for the Grid

“…The framework is called ProActive, a Java-based middleware (programming model and environment) for object and component oriented parallel, mobile and distributed computing. As this article will show, ProActive is relevant for grid computing due to its secure deployment and monitoring aspects [1,2], its efficient and typed collective communications [3], and component-based programming facilities [4] thanks to an implementation of the Fractal component model [5,6], taking advantage of its hierarchical approach to component programming.…”

“…The instruction vg.f() completes when f() has been called on all members. Other features are available regarding group communications: parameter dispatching using groups (through the definition of scatter groups), hierarchical groups, dynamic group manipulation (add, remove of members), group synchronization and barriers (waitOne, waitAll, waitAndGet); see [3] for further details and implementation techniques.…”

“…abstract away from the hardware and software runtime configuration by introducing and manipulating in the program virtual processes where the activities of the application will be subsequently deployed, 2. provide external information regarding all real processes that must be launched and the way to do it (it can be through remote shells or job submission to clusters or grids), and define the mapping of virtual processes onto real processes, 3. provide a mean to visualize, complete or modify the deployment once the application has started, 4. provide an infrastructure where Grid security is expressed outside the application code, outside the firewall of security domains, and in both cases in a high-level and flexible language.…”

A parallel object-oriented application for 3D electromagnetism

“…The result group is transparently built at invocation time, with a future for each elementary reply. It will be dynamically Other features are available regarding group communications: parameter dispatching using groups (through the definition of scatter groups), hierarchical groups, dynamic group manipulation (add, remove of members), group synchronization and barriers (waitOne, waitAll, waitAndGet); see [1] for further details and implementation techniques.…”