A Principle-based Goal-oriented Requirements Language (GRL) for Enterprise Architecture

Abstract

Business processes, goals, and tasks of individuals in an Enterprise must constantly be aligned with several regulations, standards, policies and EA principles imposed internally by organizations or externally by governments. Due to the complexity of these documents and their constant changes, it is not possible for organizations and individuals to ensure compliance with the descriptive and textual version of these documents. To overcome this, we propose a principle-based GRL framework which help to formally model these interactions with an extension of GRL and analyze the impact of the principles, regulations, standards and policies on the overall goals of the organization. In this paper, we discuss our proposed framework and then focus mainly on providing a GRL pro le for Enterprise Architecture principles.

Introduction

Enterprise Architecture (EA) principles are an organization's basic philosophies that guide the development of its architecture [ 14, 11 ], and aim to bridge the gap between high-level strategic intentions and concrete design decisions [ 9 ]. EA principles establish a context for architecture design decisions by translating business criteria into a language and speci cation that technology managers can understand and use, delimiting decisions about the architecture [ 5, 9 ].

The development of an organization so that it meets its goals is a joint effort of involved stakeholders and employees. Each individual contributes to the development of the enterprise by ful lling his own tasks and satisfying his own goals. To achieve organizational goals, individual goals and tasks have to be driven by processes to ensure compliance with internal and external constraints (e.g., strategy, EA principles, policies, regulations, requirements). Much work has been done to model policies, regulations and standards with goal modeling notations [ 7, 15 ]. However, in the context of EA, principles are informal in nature, and similar to regulations, quite vague and complex. Therefore, reasoning in general, and compliance in particular, becomes a di cult task to ful ll. By formalizing EA principles in a goal modeling framework and mapping their structure to GRL intentional elements, we aim to create an opening for further reasoning. Business Motivation Model (BMM) [ 1 ], ARMOR [ 13 ], and the motivational extension to ArchiMate [ 6 ] try to formalize such principles in terms of goals and rationales. As Quartel et al. [ 13 ] state, BMM mainly focuses on business plans, their elements, and relationships between the motivations and the business elements. It does not function as a requirements modeling language. ARMOR, which is based on goal-oriented requirements engineering methods, aims to ll the gaps for modeling motivations. However, ARMOR or motivational extension of ArchiMate do not focus on modeling principles. In addition, they lacks many of the features existing in the current goal modeling notations such as i* or GRL [ 4 ]. ARMOR is not scalable and does not include automatic analysis mechanism and tool support. GRL, on the other hand, tries to overcome the scalability issues with tool support, such as the one found in jUCMNav [ 3 ].

GRL, which is part of the User Requirements Notation (URN) [ 10 ], aims to elicit, analyze, and describe stakeholder requirements as goals. GRL supports di erent evaluation mechanisms that enable modelers to analyze the trade-o s among (often con icting) goals of stakeholders. It also helps document the rationale behind such requirements. GRL can be extended by the help of Metadata and URN links concepts of URN. Metadata are name-value pairs used to annotate (e.g., for stereotyping) any model element whereas URN links are used to de ne typed links between any pair of model elements. With the use of the metadata, it is possible to annotate GRL intentional elements with stereotypes. Ghanavati et al. [ 8 ] discussed how to model regulations with an extension of GRL called Legal-GRL. Here, we aim to do the same for modeling principles and rationales in EA by introducing a GRL Pro le for modeling EA principles.

In this paper, we introduce a principle-based GRL Pro le. To that end, we rst introduce a formalism for modeling the EA principles, adopting the principle structure of Greefhorst and Proper [ 9 ]. Next, we extend GRL with EA principles concepts and map EA principles elements to GRL intentional elements. 2 2.1

Proposed Framework

Principles based GRL

Our principle-based GRL framework, shown in Fig. 1., has three layers, being:

Organization: The organization layer contains two GRL models. The rst GRL model belongs to the organization and contains its higher levels goals. This model «contributes»to the GRL of individuals. Each individual has his own goal to achieve. By satisfying their own goals he «contributes»to the organization's goals. A decision made by an individual can be captured (at run-time or a posteriori ) in a \decision text". This element has a «traces»link with the individual's GRL model.

EA Principles: The EA Principles layer contains the textual representation of an EA principle (see Table 1) and a GRL representation of the same principle. The EA Principles GRL «complies»with the organization's goals and has the role to «restrict»the individual's goals and tasks (GRL model).

Knowledge base: In this layer, we capture the decisions taken based on the individual's GRL (the link «decides») and the previous EA Principles GRL's (the link «generates»). In the future, we aim to construct a knowledge-base system including past cases to suggest templates at run-time for individual's GRL. However, that is out of the scope of this paper.

ORGANIZATION EA Principles

Organisation GRL <<contributes>> <<contributes>>

Individual GRL <<traces>>

Decision text < < ittrrssce > >

Past cases

Knowledge base

Template/

Patterns Captured decisions EA Principle Text <<traces>>

EA Principle GRL

<<generate>> There are multiple views on EA principles. We use the structure as used by the Schiphol Group, a Netherlands-based Enterprise [ 12 ](see Table 1). We extend GRL with a set of stereotypes to capture the aspects related to EA principles. This extension helps understanding the principles in GRL, make them di erentiated from the original GRL and helps in compliance analysis with principles. Goals originating from the principle are mapped to softgoals or goals in GRL, annotated as «Principle»goals. The future situation which describes the «FutureState»when applying the principles, is modeled with intentional elements. When future state is achieved, they contribute to the high-level goals of the principles. Added Values which are the side-e ect values achieved by enforcing the principles are modeled as «AddedValue»softgoals. They are linked to «FutureState»with correlation links. «Application»tasks, which represent alternative, should be performed to reach the «FutureState». 2.3

Example

In this example, we use one of the principles de ned by a Schiphol Group. This principle shown in Table 2 is translated from Dutch internal documents of the organization. A partial de nition of the same principle can be found in [ 9 ]. This EA principle is for re-use before package selection and states that \The enterprise shall re-use the already in use applications and infrastructure. If it is not possible to re-use the existing solutions, then a standard solution shall be purchased. If a standard solution is not su cient, custom development is the third choice. Applying this principle results in more e cient use of ICT resources and reduces cost and time." EA principles are vague, complex, informal in nature, and are usually represented in a textual format. Thus, it is di cult for organizations and individuals to align their goals and tasks with principles and manage the changes.

To deal with this, we proposed a principle-based GRL to capture the existing elements in EA principles. As a starting point, we de ned the stereotypes of the principle-based GRL using the EA principle structure of the Schiphol Group. However, we need to improve this pro le by capturing various principle structures and provide concrete/formal steps for modeling with principle-based

ICT department is well aware of the trade-o between package selection, build and re-use. Regular discussions tackle these three options, Similar features are found in more than one system.

There is a catalogue of reusable components, There is a list of criteria for re-usability (e.g., costs and ownership), Optimal use of existing functionality, Reduce time by making clear which functionality is already available and can be re-used, Produces reliable and stable ICT environments, Lower diversity.

The project must balance requirements and functionality that are available o -the-shelf, Purchasing more packages which increase vendor dependency, should be manageable.

Re-use of components unless it is not possible, Use of custombuild component if no other option exists

ICT Applications GRL pro le. We aim to provide a set of rules and constraints for modeling EA principles with GRL and means to verify the correctness and well-formedness of the created GRL models. We will use the GRL capability to include constraints associated with stereotyped and linked elements [ 2 ] and de ne a UML Object Constraint Language (OCL) library. We will also improve the linkset and the analysis algorithms in GRL for compliance with principles. Finally, we want to provide textual templates for EA principles which are aligned with the GRL models.

Acknowledgments. This work has been partially funded by AFR - PDR grant #5810263 and FNR PEARL programme.

References

1. Business Motivation Model - Version 1.1. May 2010 .

2. D. Amyot , J. Horko , D. Gross , and G. Mussbacher . A lightweight GRL pro le for i* modeling . In Proc. of ER'09 Workshops , pages 254 { 264 . Springer-Verlag, 2009 .

3. D. Amyot et al. GRL Modeling and Analysis with jUCMNav . In Proceedings of iStar 2011 , pages 160 { 162 .

4. D. Amyot et al. Evaluating goal models within the goal-oriented requirement language . Int. J. Intell. Syst. , 25 : 841 { 877 , August 2010 .

5. F. J. Armour , S. H. Kaisler , and S. Y. Liu . A Big-Picture Look at Enterprise Architectures. IT Professional , 1 ( 1 ): 35 { 42 , Jan . 1999 .

6. C. Azevedo et al. An Ontology-Based Semantics for the Motivation Extension to ArchiMate . In EDOC' 11 .

7. S. Ghanavati . Legal-URN framework for legal compliance of business processes . http://hdl.handle.net/10393/24028, 2013 . Ph.D. thesis, UO, Canada.

8. S. Ghanavati , D. Amyot , and A. Rifaut . Legal goal-oriented requirement language (Legal-GRL) for modeling regulations . In MiSE @ICSE (To appear), 2014 .

9. D. Greefhorst and H. Proper . Architecture Principles { The Cornerstones of Enterprise Architecture . Enterprise Engineering Series . Springer, Germany, 2011 .

10. ITU-T. Recommendation Z. 151 ( 11 /08): User Requirements Notation (URN) { Language De nition . http://www.itu.int/rec/T-REC-Z. 151 /en, 2008 .

11. A. Lindstro m. On the Syntax and Semantics of Architectural Principles . In HICSS,' 06 .

12. J. Peetoom. [A101 .00] Architectuurprincipes v. 0 .5. Internal report by Schiphol Group (in Dutch), January 2012 .

13. D. Quartel , W. Engelsman , H. Jonkers , and M. van Sinderen . A Goal-Oriented Requirements Modelling Language for Enterprise Architecture . In EDOC '09.

14. G. L. Richardson , B. M. Jackson , and G. W. Dickson . A principles-based enterprise architecture: lessons from Texaco and Star Enterprise . MIS Quarterly , 14 ( 4 ), 1990 .

15. A. Rifaut and E. Dubois . Using goal-oriented requirements engineering for improving the quality of ISO/IEC 15504 based compliance assessment frameworks . In RE'08 , pages 33 { 42 , Spain , 2008 .