Hybrid High Performance Networking and High Performance Computing with Java Client/Server Enhanced Technology
Jun Ni, Ph.D. M.E.
Research Scientist, Research Technology/Academic Technology
Adjunct Asistant Professor, Department of Computer Science
The University of Iowa, Iowa City, IA 52242
Problem:
With the recent development of high performance networking (HPN) and high performance computing (HPC) with enterprise Java (J2EE) programming, many scientific and engineering problems can be solved on the net based on the proposed distributed computer system with parallel computing.
The current research focuses on the scalable hybrid HPN-HPC system (HHHS), enhanced by Java client/server technology in Web application and parallel computing.
The Approach:
The system is introduced based on the hybrid structure of computers, composed of various models, including single-server single-client (SSSC), single-server multi-clients (SSMC), multi-servers single-client (MSSC), and multi-servers multi-clients (MSMC). The system can be used with either internet computing as a component of GRID, or intranet computing as a local computing facility. LINUX based server clusters are needed to perform massive and parallel computing, as well as to queue scientific information in high performance databases. Each server cluster contains multiple servers and each server performs specific computing tasks. Some servers are used to communicate to other servers and clients using Java JINI, Java servlets or JSP. Others are dedicated to internal parallel computing with Java programming. This is performed directly through Java's socket communication or Java Native Interface (JNI) to access existing message passing libraries such as MPI and PETSc in C//C++.
The client clusters act not only as an interface to communicate to servers by queuing or submitting computing jobs, but also performs customized sub-computing tasks using Java applets, such as client data generation, specific problem initialization, and post-processing of data visualization and data storage, using customized Java programs and applets.
The success of the system is built upon the leading edge technologies used in large-scale high performance networking on the internet with TCP/IC communication, high speed local network systems, internal communication of client/server clusters, system configuration and topology, and optimal searching and gathering of TCP/IPs.
The experiment of HHHS is conducted using two LINUX clusters. One is used as the server cluster and the other is used as the client cluster. The server cluster contains a web server with Java servlet and JSP engines, a database sever, and computation servers, on which J2EE, Java server programs and other parallel libraries are installed. The client cluster is composed of several client nodes. Each client node has a web browser, Java2 SDK, and applets which can be used to pre-processing initialization and post-processing visualization.
Two applications are designed to test the current HHHS. One is client/server parallel computing of transport phenomena during 3D diffusion of heat and mass transfer. The other is parallel computing and simulation of 3-D Lyapunov space in chaotic phenomena. The former is intended to test massive and parallel computing performed by the server with visualization by the client, while the latter is designed to test computing performed by the client with data collection and leaping by the server.
Impact, Importance, Interest, Audience:
The current research is very important to many audiences who are interested in high performance computing with Java on the net, as well as scientists and engineers who are interested in client/server distributed computing. The proposed research will impact advanced applications of computing on HPN-HPC systems, for use in business or research and education.
Visual Presentation:
The poster will be a clear and concise layout using color charts to present the computer structures, tables to present the performance benchmarks, and pictures to present the computation results. It will also include an outline summarizing system configuration, models, accomplishments, and future studies.