Peta-scale Singel System Image

My work in this area is to investigate the applicability of single system image (SSI) approach for peta-scale computers. Peta-scale computers with thousands of times more computational power will be available to scientists by 2009. The overwhelming size and complexity of a peta-scale system requires a computing environment that addresses the scalability of the file system, a network that facilitates communications across the scale of processors, and an aggressive approach for dealing with operating system noise. The research approach undertaken in this project is to simultaneously scale Linux to 100,000 processors and employ the Open Single System Image (OpenSSI) project as a baseline to provide a balanced solution for managing system resources. OpenSSI is an open source project maintained by Bruce Walker (Co-PI of this proposal) which provides a single root file system and single process space across distributed resources. OpenSSI is a full implementation of the single system image and has a more fault tolerant peer-to-peer communication system than other implementations available.

Java Embedded Mircokernel

My work in this area is to develop a framework for a Java Embedded Micro-kernel (JEM ) that is capable of dynamically loading service modules at runtime onto devices with limited system resources, such as intelligent network cards or PDAs. The JEM system allows for a range of Java-based modules to be developed and utilized to, for example, dramatically improve the communications performance of Java applications, move operating system or application level services down into the network layer, or provide other specialized services.

Virtual Research Environment: Sakai VRE Demonstrator

This project aims to address the requirement for a single point of access to a comprehensive set of Grid and collaboration services in a VRE. more detail here

The UK e-Science OGSA Testbed project

This is a one-year project funded by EPSRC via the e-Science programme. ACET is managing the project, which also involves Daresbury Laboratory, Manchester, Reading, and Westminster universities. The objective of the project is to install and test the middleware based on the Open Grid Services Architecture (OGSA) that was recently released, and then deploy a number of applications on the testbed. more detail here

Web/Grid Portal

My work in this area is to identify basic Grid services and to develop a set of generic tools and libraries to ease the development of a Grid enabled portal. This work has led to the creation of a prototype Grid service portal, DSGPortal . As part of this investigation, we have also implemented a perl-based Web portal for the Grid enabled Liquid Crystal Modelling and Visualization simulation . Web/Grid Portal work is a research project in collaboration with CLRC e-science center to identify and to develop Grid services' portal for the UK HPC community.

Network Performance Evaluation

My work in this area includes performance data collection, visualization, and analysis, benchmarking methology, and development of benchmarking tools. In addition, I have collaborated with network vendors in developing and testing network device drivers. Past research includes performance evaluation of high speed networks (i.e., Gigabit Ethernet, Fast Ethernet, Virtual Interface Architecture, and InfiniBand .

Cluster Computing

My work in this area is to explore hardware and software technology needed to build an affordable, high performance computing platform using LINUX PCs (see KSU Fianna Cluster) . The model used is the Beowulf-Class clustering. The focus of this research is benchmarking various aspects of system performance. Independently, I have investigated several cluster management tools and batch scheduling systems.

Computational Steering and Visualization

My work in this area was to develop an environment for visualization and interactive computational steering on a network of distributed workstations. The target application of this environment is a 3-dimensional liquid crystal material in a confined geometry. The model used is based on the Landau-de Gennes theory. The properties of liquid crystals is of importance, for example, for the design of liquid displays. The target application is characteristic of a class of problems which would benefit significantly from computational steering, since it has multiple solutions with different symmetry structures and exhibits bifurca tion or turning-point behavior for certain values of the determining parameters. Hence this is a path following or continuation problem.

Generic Tools for Distributed Computing

I have designed and prototyped two libraries, Bag of Jobs and SMLib, as toolkits to support message passing and shared memory programming paradigms. The libraries reside on top of PVM/MPI system at the library layer. They serve as complements to PVM/MPI with high-level message-passing and shared memory functions.

Bag of Jobs is based on master-slave model. It addresses the issues of scheduling, synchronization, and mechanisms to aid programmers in writing their application programs based on master-slave model. This library handles the master-slave communication patterns thus shielding programmers from the details of network communications layer.

SMLib provides an environment for programmers to write their application programs using the shared memory paradigm. SMLib is an object based shared memory library. It uses distributed dynamic manager scheme to manage shared data objects. Strict memory consistency and releasememory consistency were implemented to maintain memory coherence.