This thesis has developed a Dynamic Configuration Management Framework and implemented a prototype using Puppet Enterprise to support evaluation of the suitability of said framework. The field of research in Configuration Management, specifically in the High Performance Computing Application area for Cloud infrastructures is still under developed and this research contributes greatly to that field of research, serving as a stepping stone for future work to be done. The results from the implementation of the prototype demonstrates an overwhelming increase in efficiency and reliability of Configuration Management of 77% to 81%, and 89% to 100% respectively. This assures the framework of two facts. Firstly, the proof that the use of the framework proposed and the Configuration Management tool has a positive effect on efficiency, and secondly, that the reliability of the system Configuration Management increases when compared to a traditional style of Configuration Management, that is - the manual or semi-scripted process. This thesis demonstrates the feasibility on the investment for a Configuration Management tool in use with distributed infrastructure such as the Cloud. The framework proposed also serves as a guideline and process chart for Configuration Management of dynamic packages and High Performance Computing application requirements.
History
Campus location
Australia
Principal supervisor
Jefferson Tan
Year of Award
2013
Department, School or Centre
Information Technology (Monash University Clayton)