To measure the achievement of predefined Enterprise Architecture Management (EAM) goals, it is essential to empower business users to define organization-specific Key Performance Indicators (KPIs). However, to support tool-based calculation of such KPIs, a formal model-based query language is required for their definition and calculation.
In this paper we first examine existing general-purpose query languages regarding their suitability for the definition of business-user-specific KPIs in a collaborative environment. Thereafter, we justify the demand for a domain-specific query language ensuring a balance between the strengths of existing query languages and the size and purpose of the EAM domain. Based on this, we outline important design details and a prototypical implementation of such a language in a EAM tool.
Finally, our language design is being evaluated by the implementation of suggested EAM KPIs from literature on the one hand, and by the development of a prototype for the use in an EU project on the other hand.
As motivated in Section 2, a collaborative environment requires the support for the definition and computation of KPIs by business users. Since Microsoft's Excel suffers from some major drawbacks in this context (e.g. missing collaboration support), we decided to integrate a quantification mechanism into a collaborative EAM tool, which, however, requires an appropriate domain-specific query language to define and compute KPIs. Because of our aim to implement a domain-specific query language, which provides a minimal, but sufficient expressiveness, appropriate readability for business users as well as optimizability, we designed a new language instead of integrating an existing one.
The language (Section 3 covers its basic design as well as its supported operators) is inspired by the context of defining and computing business-user-oriented KPIs in a collaborative environment on the one hand, and by existing general-purpose languages (e.g. SQL, OCL, LINQ) on the other hand. Its prototypical implementation in our research EAM tool, covered in Section 4, enables derived attributes of entities as well as embedded expressions. Moreover, since integrated in a web-based tool, our language supports the definition of user-specific, rudimentary visualizations by the generation of HTML markup.
To evaluate the language (see Section 5), we implemented the KPIs of the EAM KPI Catalog as well as the SmartNet Navigator, computing and visualizing a project's status based on related tasks and meetings.
Although the evaluation of our language highlights the capability of defining, computing and visualizing KPIs in a collaborative environment, it is not yet clear, if the language is practicable for business users, which have at least some experience with Microsoft's Excel. Therefore, further evaluation has to be done in order to improve the language and its syntax, especially regarding its suitability for business users.
Apart from this, further new questions emerged during our research, which should be tackled in future research: