The purpose of the chapter is to provide clarity on what a Federated Enterprise Architecture (FEA) is and what the benefits as well as risks are in contrast to a non-federated enterprise architecture. The chapter draws upon organizational theory, federalist theory, and case studies to explicate what constitutes a federated model and the expected federated EA benefits. There are a number of challenges with the concept of a FEA. Two are focused on in this chapter: the meaning of federated EA and associated benefits and risks. The first is the use of the term “federated,” which occurs rather frequently in ICT literature, such as “federated search” or “federated database design,” and in the context of IT governance, “federal model” in Drs. Weill and Ross’s book IT Governance. The term also appears in the non-ICT context such as “federated insurance.” However, the term “federated” is frequently not defined and when defined speaks to a decentralization concept. This distinction is relevant to the understanding and success of a federated EA implementation. In reviewing federalist theory, there is a clear difference between decentralization and federalism. It is argued that the so-called federal or federated “model,” as described, is not federated but is a form of decentralization. The second challenge within the EA discipline is the lack of benefits attributed to a FEA. In the few sources that exist for FEA benefits are either not stated or the stated benefits could equally apply to a non-FEA. It is argued that scalability is the singular key benefit that FEA provides over a non-FEA, and the following non-FEA benefits are enhanced: 1) agility and IT innovation, 2) process consolidation and business process standardization and discipline, and 3) interoperability. However, while there are clear benefits to FEA, there are inherent risks.
TopIntroduction
Within the discipline of Enterprise Architecture (EA) there is an emergence of an EA style or archetype – Federated Enterprise Architecture (FEA) or hybrid (Allega, 2009; Burke & Tuft, 2012; DoD CIO, 2007; Drecun, 2003; Fernandez, Zhao, & Wijegunaratne, 2003; Ross & Beath, 2005; Roy, 2006; Wilson, 2012; Zachman, 2006). The advancement of this new form of EA has been argued as a response to today’s complex organizational structures such as networked Multinational Corporations (MNC), M-form, and large federal government entities (i.e., Cabinet level Departments) (Allega, 2009; Zachman, 2006) who through competitive pressures and/or scale are attempting to optimize centralization and decentralization structures and management controls to improve organizational performance (Chandler & Mazlish, 2005; Prahalad & Doz, 1981).
There is pragmatic interest in a FEA, based on a recent informal survey taken by Gartner during a Webinar of 74 responses 11% indicated they are pursuing a federated EA approach (Burton, 2010). There is also in interest in FEA by the public sector. The United States’ Department of Defense (DoD CIO, 2007) and the Department of Health and Human Services (United States, 2010) are pursuing a FEA and the national governments of Canada (Ministry of Information and Communications Technology, 2006), Jordan (Roy, 2006), and Uganda (Rwangoga & Baryayetunga, 2007) have proposed the use of a FEA as a component of their respective E-Government programs. Curiously enough, at least in these public sector examples, justification for selecting an FEA approach rather than a classical or non-FEA is omitted. Given that the empirical evidence for the benefits of classical EA is wanting (Boucharas, van Steenbergen, Jansen, & Brinkkemper, 2010). What empirical or theoretical evidence is there to support the adoption of an FEA over a non-FEA? Are there additional risks and do they outweigh the benefits? If the scope is extended to include IT Governance as a compliment to enterprise architecture (Sambamurthy & Zmud, 1999; Weill & Ross, 2004; Weill & Ross, 2005; Willson & Pollard, 2009), additional key insight is revealed.
Weill and Ross (2004) in their book IT Governance observed, for 74 government and not-for-profit organizations, most used a federal approach (as described similar to federated) for IT Governance input (to include IT principles, IT architecture, IT infrastructure), but not for the corresponding decision execution. Furthermore, of the six archetypes identified, they state “The federal model is undoubtedly the most difficult archetype for decision making because enterprise leaders have different concerns from business unit leaders” (p. 61). Returning to the private sector, of the top three IT governance performance strategies only two include a federal archetype and only for one of the five decision criteria – Business Application Needs (Weill & Ross, 2004), a narrow application of a federal strategy. In a separate study of 356 multibusiness firms, Tanriverdi (2006), focused on the performance effects of IT synergies and found that of three IT governance modes (centralized, decentralized and hybrid), the hybrid mode was statistically the least effective, supporting prior observations by Brown (1998; 1999). As described, the hybrid IT mode combined central and decentralized decision-making and shares characteristics of a federal model or federated structure. Taken all together, not a resounding affirmation for a FEA approach – what then is the allure?