Zirkelbach, Christian (2015) Performance Monitoring of Database Operations. (Master thesis), Kiel University, Kiel, Germany, 109 pp.

Preview	Text msc_czi_thesis.pdf - Published Version Download (2MB) | Preview
Preview	Slideshow thesis_presentation_czi.pdf - Presentation Download (767kB)

Abstract

The complexity of software projects and related software systems advances and likewise the
quantity of data increases. Therefore within the area of software development, performance
engineering is a popular topic. This is usually connected to performance tuning of a
software system. Performance problems are often related to database operations, as
complex database queries challenge the software engineer to achieve efficiency.

In order to aid the process of analyzing performance problems, in the context of database
communication, it is useful to conduct a performance analysis focused on database statements,
which are executed by the software system. Most tools, which support a performance
analysis, cover just business operations, e.g., operation calls in Java. So seeking a
suitable tool for this purpose is challenging, as we focus on database communication and
software, which is freely-available.

In this thesis, we present an aspect-oriented approach for performance monitoring of
database operations, in order to support software engineers in conducting a performance
analysis. As we developed our monitoring approach in a generic manner, we allow several
monitoring tools to use our generated monitoring records as input data for posterior
analysis and visualization purposes. We verified the versatility of our monitoring approach
through the integration into Kieker Trace Diagnosis, a graphical user interface for displaying
program traces, followed by another tool, ExplorViz. As proof of concept, we investigated
the first usage of our developed software by evaluating the usability through a first
experiment based on a paper-based questionnaire in combination with our developed tool.
In our experiment, we observed positive feedback from our participants, which correlated
with the high average correctness rates of the results. Therefore, the experiment reveals that
our software has a bearing on conducting a performance analysis of database operations.
As we have validated the usability of our developed tool through our conducted experiment,
our software system is worthwhile to be extended in the future.

Actions (login required)

View Item