Modern enterprise landscapes based on heterogeneous application architectures consist of numerous software products, integrating middleware and many further software and technical layers. Owing to their heterogeneous nature, these landscapes are innervated by different architectural approaches, structures, mechanisms and system behaviors. On top of the landscapes, business scenarios perform the specific business tasks and obligations. The scenarios make use of diverse applications or functionalities and data from diverse information sources. Flawless execution regarding correctness and punctuality is often essential for the survivability of companies. Landscape integrity on functional and data level is therefore an important quality which needs to be monitored, retained and regained in the case of inconsistencies occurring.
Although enterprise landscapes play a vital role for most corporate processes, management, control and maintenance issues are not sufficiently taken into consideration. Research which concentrates on heterogeneous application landscapes as an integral whole, that means, including all individual parts, mutual dependencies and activities on all layers, has been neglected.
Minor incidents or changes can still cause serious repercussions and may reduce a company’s productivity. A loss of data or misuse can also harm its image tremendously. From this it follows, that any change must be taken seriously and dependencies within the application landscape must be evaluated carefully beforehand.
Our approach was to find mechanisms which improve the control and transparency of heterogeneous application landscapes. Therefore, maintenance and business scenarios were investigated by means of an attempt to answer the questions what information is required and when intervention is needed in order to support their executions.
As a result, the thesis deals with control instances combined with dedicated knowledge databases. This combination provides more details about current and past states of landscape artifacts and activities. More precise information will lead to better controllability and maintainability of complex environments. The information is also of interest for several other purposes, such as optimizations, redesign or development.
Landscape controllers are definitely a step ahead in coping with the constantly growing size of distributed and heterogeneous application architectures. Their use will have a significant impact on qualities, like reliability, transparency, traceability and evolvability, too.