Undersökning och effektivisering av nyttjandegrad i mjukvarutestmiljö : En fallstudie vid Försäkringskassan IT

Sammanfattning: Background. Testing of a software system is a critical part of the software development process, as the need for efficient and high-quality software systems is of great importance. By using so-called test environments, the testing process can be automated and more effective, which in turn contributes to lower time and development costs. Objectives. This thesis aims to investigate the degree of usage of a test environment by introducing a model that measures the degree of usage for a test environment group. This thesis also aims to introduce suggestions for new measuring points to monitor the degree of usage for test environments. This thesis also presents suggestions for improvement measures that can be taken to increase the degree of usage for a test environment. Methods. The method that has been applied to carry out this thesis is a case study. The study began with a literature review followed by observations and semistructured interviews to identify factors that affect the measurement of the degree of usage. Also, archive research in form of log analysis has been carried out to analyze and retrieve data from machine logs. The suggested model has been evaluated through a static evaluation and the research has been carried out in collaboration with Försäkringskassan IT. Results. The results of this thesis are presented through a model that which is based on Goal Question Metric (GQM) approach. The model measures the degree of usage in a test environment. The model is based on reading data from machine logs, where the data reading befalls both during the night and day. The data reading can then be visualized by two (2) curves that present both an average usage (approximately 30 days) and a daily usage. These curves help to create an overall picture of the usage of a test environment, but also when during a day mostly usage happens. The visualization can then present through predefined limits on what is considered as good usage (the limits are different for different types of tests). To increase the measurement of the degree of usage, the model can be supplemented with measurement data from, among other things, the number of logins in the test environment, incoming and outgoing network traffic. Based on the analysis, measures such as sharing integration testing environments, and prioritize teams/projects that are working towards the next release can be taken to increase the degree of usage of a test environment. Conclusions. The degree of usage of a test environment can be measured by reading data from machine logs. In combination with measurement data from CPU, memory, and network traffic, the usage of a test environment can be measured and monitored, which in turn contributes to a higher degree of usage and efficiency around the testing process.