Drafting Enterprise Models

Introduction

Enterprise Modeling refers to the practices to build and use Enterprise Models to represent multiple aspects of an enterprise. The purpose of enterprise models is to document, communicate, analyze organizational structures, business processes, information systems, and infrastructure with the aim of improving these elements (Buschle, Ullberg, Franke, Lagerström, & Sommestad, 2010; Iacob & Jonkers, 2006; Kurpjuweit & Winter, 2007; Lankhorst, 2013). While enterprise models can be small, simple, and well-focused, most are typically the opposite; they are large, complex, and cover a large number of areas and lines of business. Such extensive and complex models are useful for enterprise analysis purposes because they enable holistic approaches for understanding an organization, combining both business and technological aspects (Lagerström, Franke, Johnson, & Ullberg, 2009; Lankhorst, 2013).

However, this complexity has strong implications for the construction of the models. The construction process starts with the selection of a framework of reference that defines the structure of the model, and which is typically based on a metamodel. Then, modelers have to discover and classify multiple sources of enterprise information (e.g., people, documents, meetings) which can be located anywhere in the company, or may even be external (e.g., regulations, policy documents). Afterwards, modelers perform different kinds of observations on sources of enterprise information (e.g., interviews, reviews, meeting minutes), extract information, and introduce these into the model.

It is at this point that problems start to appear. First of all, there is the issue of compatibility between the facts and the metamodel selected as the underlying structure for the enterprise model. If the metamodel does not offer an adequate metatype to represent a certain fact, that fact will either have to be omitted from the model, or it will have to be adjusted in order to match the metatypes existing in the metamodel. For example, standard metamodels commonly have a metatype to represent infrastructure devices, but do not include a specific metatype to represent load balancers. Thus, models sacrifice a bit of semantics when they represent load balancers using the simpler infrastructure device concept. The second problem has to do with the quality of the sources and the observations, which may have different levels of precision and reliability. For example, it is possible to find sources with contradictory information, opposite points of view, obsolete information, and conflicts of interest which can generate inconsistency in the model.

A natural way to handle the construction of enterprise models is to use drafts, that is, to a model in phases without having committed to a definitive metamodel from the beginning. Furthermore, drafts may lack complete information and even be imperfect (imprecise, inconsistent, vague, uncertain) (Florez, Sanchez, & Villalobos, 2014b; Florez, Sánchez, & Villalobos, 2013). The problem is that current enterprise modeling environments, as well as general purpose modeling frameworks and tools, lack the capabilities to support what was just described: typically there is little or no support for flexible or changing metamodels, nor is there support for imperfection in the models, or to trace the modeling processes.

Models have to be perfectly structured during the entirety of the modeling process.

The objective of this chapter is to propose mechanisms to support the construction of drafts for enterprise modeling, while addressing problems highlighted in existing tools. This proposal is based on three core elements: 1) a strategy to support flexible metamodels which makes it possible to easily change the metamodels even after the modeling process has started; 2) the necessary elements to model imperfection caused by the information gathering process and the quality of the sources; and 3) the support for documenting the construction and evolution of the models. This documentation, which presents information about the sources, observations, and facts used by modelers to build the drafts, is valuable to build the definitive versions of the models. The whole proposal described has been implemented in iArchiMate, which is a modeling tool for the creation of enterprise model drafts based on ArchiMate (The Open Group, 2012).