Use this URL to cite or link to this record in EThOS:
Title: An agent-based visualisation system
Author: Roard, Nicolas
Awarding Body: Swansea University
Current Institution: Swansea University
Date of Award: 2007
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
This thesis explores the concepts of visual supercomputing, where complex distributed systems are used toward interactive visualisation of large datasets. Such complex systems inherently trigger management and optimisation problems; in recent years the concepts of autonomic computing have arisen to address those issues. Distributed visualisation systems are a very challenging area to apply autonomic computing ideas as such systems are both latency and compute sensitive, while most autonomic computing implementations usually concentrate on one or the other but not both concurrently. A major contribution of this thesis is to provide a case study demonstrating the application of autonomic computing concepts to a computation intensive, real-time distributed visualisation system. The first part of the thesis proposes the realisation of a layered multi-agent system to enable autonomic visualisation. The implementation of a generic multi-agent system providing reflective features is described. This architecture is then used to create a flexible distributed graphic pipeline, oriented toward real-time visualisation of volume datasets. Performance evaluation of the pipeline is presented. The second part of the thesis explores the reflective nature of the system and presents high level architectures based on software agents, or visualisation strategies, that take advantage of the flexibility of the system to provide generic features. Autonomic capabilities are presented, with fault recovery and automatic resource configuration. Performance evaluation, simulation and prediction of the system are presented, exploring different use cases and optimisation scenarios. A performance exploration tool, Delphe, is described, which uses real-time data of the system to let users explore its performance.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available