Browsing by Subject "system Grid"

Now showing 1 - 2 of 2

Access status: Open Access ,
Component-based methodology for programming and running scientific applications on the grid
(Data obrony: 2009) Malawski, Maciej
Wydział Elektrotechniki, Automatyki, Informatyki i Elektroniki
Development of modern scientific applications makes issues such as composition and integration a significant challenge. Additionally, the Grid technology induces further problems resulting from the distributed, shared, heterogeneous and unreliable nature of resources it offers. For these reasons, programming and running scientific applications on the Grid remains an important and relevant problem. The proposed methodology to address this problem is based on two principles: to follow a component-based programming model and to use a flexible technology for infrastructure virtualization. The author demonstrates how this can be implemented by combining the unique features of the Common Component Architecture (CCA) model together with the H2O resource sharing platform, and how it can be enhanced by a set of methods and tools. The solutions described in this thesis include high-level scripting and descriptor-based approaches, support for multiple programming languages using Babel, interoperability with GCM model, and deployment on existing infrastructures. The MOCCA component framework serves as a basis for those higher-level solutions. Applications and tests included Monte Carlo simulation, domain decomposition of cellular automata, data mining experiments in the ViroLab virtual laboratory, as well as a set of synthetic benchmarks designed to verify the proposed solutions.
Access status: Open Access ,
Monitoring of grid scientific workflows
(Data obrony: 2009) Baliś, Bartosz
Wydział Elektrotechniki, Automatyki, Informatyki i Elektroniki
Grid scientific workflows are a means of defining in silico experiments in modern infrastructures for e-Science. Monitoring of Grid scientific workflows poses specific challenges due to the dynamicity of the Grid, and the diversity of scenarios where monitoring is important. The main contribution of this dissertation is the identification and analysis of key challenges in monitoring of Grid scientific workflows, elaboration of solutions to those challenges, and validation of those solutions. Four areas where key issues arise are recognized: building an infrastructure for Grid scientific workflows, solving the problem of on-line monitoring support within this infrastructure, monitoring of workflow legacy backends, and development of an information model for recording workflow executions. The results of the research are validated using diverse methodologies. Prototypes of the designed software components were built and used for monitoring of real-life workflows within EU-IST projects. Model-based approach using Queuing Networks and a discrete-event simulation has also been used to validate the performance characteristics of the proposed solutions.