So Hyun Ryu

Collaborative object-oriented visualization environment

In this paper, we present a Collaborative Object-oriented Visualization Environment (COVE) which provides a flexible and extensible framework for collaborative visualization. COVE integrates collaborative and parallel computing environments based on a distributed object model. It is built as a collection of concurrent objects: collaborative and application objects which interact with one another to construct collaborative parallel computing environments. The former enables COVE to execute various collaborative functions, while the latter allows it to execute fast parallel visualization in various modes. Also, flexibility and extensibility are provided by plugging the proper application objects into COVE at run-time, and making them interact with one another through collaboration objects. For our experiment, three visualization modes for volume rendering are designed and implemented to support the fast and flexible analysis of volume data in a collaborative environment.

Workflow management system based on service oriented components for grid applications

To efficiently develop and deploy parallel and distributed Grid applications, we have developed the Workflow based grId portal for problem Solving Environment(WISE). However, the Grid technology has become standardized related to Web-services for Service-oriented architecture and the workflow engine of WISE is insufficient to meet the requirements of a service-oriented environment. Therefore, we present a new workflow management system, called the Workflow management system based on Service-oriented components for Grid Applications(WSGA). It provides an efficient execution of programs for computational intensive problems using advanced patterns, dynamic resource allocation, pattern oriented resource allocation, and configures Web-service as an activity of workflow in Grid computing environment. In this paper, we propose the WSGA architecture design based on service-oriented components, and functions for using Web-services and increasing system performance. Also, we show an implementation method, and report the system performance evaluation of the WSGA architecture design.

XML-Based Workflow Description Language for Grid Applications

In Grid computing PSE is essential for developing parallel and distributed applications in various computation intensive areas. The workflow-based PSE supports the coordinated execution of multiple application tasks on distributed Grid resources by enabling users to describe a workflow by composing many existing grid services or new functions. In this paper, we propose a new workflow model which consists of powerful workflow patterns efficiently used to represent various parallelisms inherent in parallel and distributed applications. It also provides an integrated scheme which represents data and control flows simultaneously. Also, we present an XML based Grid workflow description language(GWDL) which efficiently reflects our proposed workflow model.

DOVE: A Virtual Programming Environment for High Performance Parallel Computing

In this paper, we present a new parallel programming environment, which is called DOVE(Distributed Object-oriented Virtual computing Environment), based on distributed object model. A parallel program is built as a collection of concurrent and autonomous objects interacting with one another via method invocation. It appears to a user logically as a single virtual computer for a set of heterogeneous hosts connected by a network as if objects in remote site reside in one virtual computer. The main goal of DOVE is to provide users with easy-to-use programming environment while supporting efficient parallelisms encapsulated and distributed among objects. Efficient parallelisms are supported by various types of method invocation and multiple method invocation to the object group. For the performance evaluation purpose of DOVE, we have developed two parallel applications both on DOVE and PVM. Our experiment shows that DOVE has better performance than PVM and provides an efficient and easy-to-use parallel programming environment for a set of heterogeneous and clustered computers.

Interactive visualization pipeline architecture using work-flow management system on grid for CFD analysis

Since the potential of the impact of complex problem like CFD on design of material that is difficult to test actually has been established, CFD analysis that consist of two parts – simulation part and visualization part is considered as important work. But many researcher think it is difficult to analyze CFD problem because of following two reasons. First reason is that CFD analysis usually needs numerous processing time and much computing power. and Second is that the tasks of simulation and visualization are usually done independently – numerical simulation then visualization with stand-alone visualizing system. we propose solution of two problem. First problem is solved by using parallel/distributed programming on problem solving environment on grid environment – WISE system. and Second problem is solved by interactive visualization pipeline architecture that enables user to conducting simulation and visualization at the same time automatically applying pipeline concept. Therefore we presented interactive visualization pipeline architecture and implemented using WISE system.