Cluster Computing and Performance Measurement

There is a continual demand for greater computational power from computer systems
than is currently possible. Areas requiring great computational speed include numerical
simulation of scientific and engineering problems. Such problems often need huge
quantities of repetitive calculations on large amount of data to give valid results.
Cluster computing offers many advantages as a highly cost-effective and often scalable
approach for high-performance computing in general. To achieve the full potential of
high performance computing systems, centralized configuration services are the starting
point. For a large scale of projects, cluster computing is required where it is supposed
to be optimized for the system topology and management of the project. This paper
presents the consequences of using cluster computing and performance management
and the consequences without this technology. The experimental results of this paper
highlight the affects of the design of this service and provide a comprehensive
performance analysis of the project.