Performance management of enterprise information systems using a knowledge base

Software Performance Engineering methods are applied to evaluate and improve the performance of Enterprise Information Systems. The application of these methods generates huge amount of heterogeneous and distributed measurement data as well as data of software performance modeling techniques. In order to simplify and automate the application of Software Performance Engineering and to enable meaningful performance analysis these two types of data must be stored and integrated in an efficient way. To meet this challenge the design and development of a Performance Knowledge Base is proposed in this paper. The Performance Knowledge Base is designed as a centralized, harmonized and persistent data store for the storage of all kind of generated performance data and performance knowledge. It combines and integrates measurement data and model data with the help of an integrated data mart. This paper draws a picture of the proposed architecture and the main functional requirements.