Enterprise Architecture: A Social Network Perspective David Dreyfus and Bala Iyer Boston University {ddreyfus, bala}@bu.edu Abstract IS Architecture emerges as a result of a sequence of IS project implementations. The architecture that emerges can be viewed as a network of software components linked by their interdependencies. The network influences, and is influenced by, the intra- organizational interdependencies in which it is embedded. IT management can influence the evolution of the network, and, by extension, the evolution of the organization. However, given time and cost constraints, IT management can most directly influence only a few of the components in the network, the architectural control points. In this research we show how a network perspective using research from social network analysis provides a useful abstraction for understanding architecture. We apply modular operators from design theory to enact changes to architecture. Finally, we show that by following a few simple rules, enterprises can improve the fitness of their architecture as the network emerges and the control points shift over time. 1. Introduction Enterprises face a turbulent, competitive environment. To respond, they must continuously innovate and engage in multiple IT implementation projects. IT managers must decide whether to make or buy systems, and whether to in-source or out-source projects. To provide guidance for project-level decision-making, companies develop a designed architecture. However, subsequent projects’ impact on existing applications, data, and technology can, and often does, create a gap between the emergent architecture and the designed architecture. We propose that the ability to define and manage this gap will improve firms’ ability to realize the expected returns on their IT investments. We argue that understanding this emergent architecture allows managers to recognize the impact of potential projects, and make decisions that maximize the chances of individual- project and IT-strategy success. IS architecture has no universally accepted definition in either the research arena or in the practitioner world [1]. Architecture can be viewed strategically [1-5], organizationally [6-10], and technologically [11-14]. Previous studies have summarized that an IS architecture includes a group of shared, tangible IT resources (i.e., hardware, software, data, training, management, etc.) that provide a foundation to enable present and future business applications [5, 8, 10, 15]. Architecture, as implemented through its IT infrastructure, should be flexible, reliable, robust, scalable, and adaptable [7, 8, 15]. It should support the reuse of business components within a firm, while supporting firm responsiveness, innovativeness, and economies of scope [15]. A review of these and other articles clearly illustrates that when the given definition for architecture is translated into action, the concept becomes very complicated. Architecture implementation also involves learning effects. As researchers have explored organizations making changes to architecture, they have identified two strategies: localized exploitation or enterprise-wide integration [16]. In addition, they have identified that changes do not occur in one step: they occur in stages [1]. Architecture also reflects and supports business strategy. Architecture is not just concerned with the allocation of resources at the physical level, but also the support of strategic business goals. The architectural challenge is not just cost minimization in the allocation of task to computational device, but the alignment of the task structure [17], supported by the information system, with the business objectives of the organization. Zachman provides a useful framework that identifies the components of an IT architecture as well as the various perspectives taken during the design and implementation of an architecture [18]. In this view, there is no such thing as a single information architecture; there are many. Separate architectures exist for scope/objectives, business model, information system description, technology model, detailed description, and machine language description. For these six categories, there are also the descriptions for Proceedings of the 39th Hawaii International Conference on System Sciences - 2006 1 0-7695-2507-5/06/$20.00 (C) 2006 IEEE