Data originated from the components of a system are produced quickly in large amount, and reveal a lot of information about the systems status. Because of that the analysis of these data is high priority. The processing of these data can be greatly helped by the known information of the system or the connections of the components in the system. Therefore, this information can be hard to find manually.
A possible solution could be to store the system model and its main connections, which can be evaluated to get a quick answer for the asked question about the system. I am looking for a solution in my thesis for this problem. In the system I built I am using on-tology as a possible structure to store the model. In my thesis, I show how to create the ontology from a system model, how to store and access it, and how to connect to the relevant measurements.
In my thesis I show this workflow on the case study of the performance measurement a virtual infrastructure and the measured results of its test. To store the ontology, I pro-pose graph database and time series database for the measurement results, selected after a technology evaluation phase. The connection between the two types of data is made under R scripts, where questions can be answered on the basis of the ontology, without previous exact topological connections from the measured system.
At the end of my thesis I show how to process and analyse the time series which are returned by asking questions from the ontology. Through the analysis I can make state-ments about the systems status, and forecast the future values of the components.
This solution makes the system analysis easier and faster for the data analysts. In the case study helps to find the bottlenecks of the system and inspect the connections be-tween the components of the system.