Kroll, Lars (2011) Performance Monitoring for a Web-based Information System (Bachelor thesis), Kiel University, Kiel, Germany, 40 pp
Thesis.pdf - Published Version
Download (1062Kb) | Preview
For the omni-present web-based information systems to be as unobtrusive as possible, when providing their services, it is necessary that they meet certain performance requirements. These could be, for example, Average Time to Answer (ASA) or Time Service Factor (TSF) metrics as specified in a Service Level Agreement (SLA). If a system has requirements on its performance, as many web-based information systems do, it will usually be necessary to verify compliance with these using performance testing. During performance tests, however, bottlenecks can only be identified when sufficient data is available. To reduce the overhead of gathering system metrics during performance testing, low-overhead monitoring frameworks like Kieker have been introduced.
In order to ensure compliance with performance requirements over the service life of a system, these monitoring frameworks can be used continuously. In that way they become a part of the software system. Therefore we argue that performance monitoring should be integrated into the development cycle.
In this thesis we evaluate four run-time monitoring approaches with regard to their usage for performance monitoring. COMPAS JEEM, Magpie, the Rainbow Project and Kieker are presented and analysed based on the effort of instrumentation, benefit of their data and the run-time overhead associated with using them.
Following the evaluation we present guidelines for the integration of continuous performance monitoring into software development. The guidelines are organised in a way that can be used for a large number of software development processes. They cover the most important decisions to be made while developing a continuously monitored software system. Additionally, we demonstrate and evaluate these guidelines on an example, by applying them to the Thales Product Information System (TPIS) using the Kieker framework.
|Document Type:||Thesis (Bachelor thesis)|
|Thesis Advisors:||Hasselbring, Wilhelm, Thiele, Stefan and Waller, Jan|
|Research affiliation:||Kiel University > Faculty of Engineering > Department of Computer Science > Software Engineering|
|Date Deposited:||26 Sep 2012 11:34|
|Last Modified:||22 Oct 2012 08:25|
Actions (login required)