An enterprise architecture framework (EA framework) defines how to create and use an enterprise architecture. An architecture framework provides principles and practices for creating and using the architecture description of a system. It structures architects' thinking by dividing the architecture description into domains, layers, or views, and offers models – typically matrices and diagrams – for documenting each view. This allows for making systemic design decisions on all the components of the system and making long-term decisions around new design requirements, sustainability, and support.[2]
Enterprise architecture regards the enterprise as a large and complex system or system of systems.[3] To manage the scale and complexity of this system, an architectural framework provides tools and approaches that help architects abstract from the level of detail at which builders work, to bring enterprise design tasks into focus and produce valuable architecture description documentation.
The components of an architecture framework provide structured guidance that is divided into three main areas:[4]
The earliest rudiments of the step-wise planning methodology currently advocated by The Open Group Architecture Framework (TOGAF) and other EA frameworks can be traced back to the article of Marshall K. Evans and Lou R. Hague titled "Master Plan for Information Systems"[7] published in 1962 in Harvard Business Review.[8]
Since the 1970s people working in IS/IT have looked for ways to engage business people – to enable business roles and processes - and to influence investment in business information systems and technologies – with a view to the wide and long term benefits of the enterprise. Many of the aims, principles, concepts and methods now employed in EA frameworks were established in the 1980s, and can be found in IS and IT architecture frameworks published in that decade and the next.[9]
By 1980, IBM's Business Systems Planning (BSP) was promoted as a method for analyzing and designing an organization's information architecture, with the following goals:
In 1982, when working for IBM and with BSP, John Zachman outlined his framework for enterprise-level "Information Systems Architecture". Then and in later papers, Zachman used the word enterprise as a synonym for business. "Although many popular information systems planning methodologies, design approaches, and various tools and techniques do not preclude or are not inconsistent with enterprise-level analysis, few of them explicitly address or attempt to define enterprise architectures."[10] However, in this article the term "Enterprise Architecture" was mentioned only once without any specific definition and all subsequent works of Zachman used the term "Information Systems Architecture".[11][12]
In 1986, the PRISM architecture framework was developed as a result of the research project sponsored by a group of companies, including IBM, which was seemingly the first published EA framework.[13]
In 1987, John Zachman, who was a marketing specialist at IBM, published the paper, A Framework for Information Systems Architecture.[11] The paper provided a classification scheme for artifacts that describe (at several levels of abstraction) the what, how, where, who, when and why of information systems. Given IBM already employed BSP, Zachman had no need to provide planning process. The paper did not mention enterprise architecture.
In 1989, the National Institute of Standards and Technology (NIST) published the NIST Enterprise Architecture Model.[14] This was a five-layer reference model that illustrates the interrelationship of business, information system, and technology domains. It was promoted within the U.S. federal government. It was not an EA framework as we see it now, but it helped to establish the notion of dividing EA into architecture domains or layers. The NIST Enterprise Architecture Model seemingly was the first publication that consistently used the term "Enterprise Architecture".[13]
In 1990, the term "Enterprise Architecture" was formally defined for the first time as an architecture that "defines and interrelates data, hardware, software, and communications resources, as well as the supporting organization required to maintain the overall physical structure required by the architecture".[13][15]
In 1992, a paper by Zachman and Sowa[12] started thus "John Zachman introduced a framework for information systems architecture (ISA) that has been widely adopted by systems analysts and database designers." The term enterprise architecture did not appear. The paper was about using the ISA framework to describe, “...the overall information system and how it relates to the enterprise and its surrounding environment.” The word enterprise was used as a synonym for business.
In 1993, Stephen Spewak's book Enterprise Architecture Planning (EAP) defined a process for defining architectures for the use of information in support of the business and the plan for implementing those architectures. The business mission is the primary driver. Then the data required to satisfy the mission. Then the applications built to store and provide that data. Finally the technology to implement the applications. Enterprise Architecture Planning is a data-centric approach to architecture planning. An aim is to improve data quality, access to data, adaptability to changing requirements, data interoperability and sharing, and cost containment. EAP has its roots in IBM's Business Systems Planning (BSP).[13]
In 1994, the Open Group selected TAFIM from the US DoD as a basis for development of TOGAF, where architecture meant IT architecture. TOGAF started out taking a strategic and enterprise-wide, but technology-oriented, view. It emerged from the desire to rationalize a messy IT estate. Right up to version 7, TOGAF was still focused on defining and using a Technical Reference Model (or foundation architecture) to define the platform services required from the technologies that an entire enterprise uses to support business applications.[9]
In 1996, the US IT Management Reform Act, more commonly known as the Clinger-Cohen Act, repeatedly directed that a US federal government agency's investment in IT must be mapped to identifiable business benefits. In addition, it made the agency CIO responsible for, “...developing, maintaining and facilitating the implementation of a sound and integrated IT architecture for the executive agency.”
By 1997, Zachman had renamed and refocused his ISA framework as an EA framework; it remained a classification scheme for descriptive artifacts, not a process for planning systems or changes to systems.
In 1998, The Federal CIO Council began developing the Federal Enterprise Architecture Framework (FEAF) in accordance with the priorities enunciated in Clinger-Cohen and issued it in 1999. FEAF was a process much like TOGAF's ADM, in which “The architecture team generates a sequencing plan for the transition of systems, applications, and associated business practices predicated upon a detailed gap analysis [between baseline and target architectures].”
In 2001, the US Chief CIO council published A practical guide to Federal Enterprise Architecture, which starts, “An enterprise architecture (EA) establishes the Agency-wide roadmap to achieve an Agency's mission through optimal performance of its core business processes within an efficient information technology (IT) environment." At that point, the processes in TOGAF, FEAF, EAP and BSP were clearly related.
In 2002/3, in its Enterprise Edition, TOGAF 8 shifted focus from the technology architecture layer to the higher business, data and application layers. It introduced structured analysis, after information technology engineering, which features, for example, mappings of organization units to business functions and data entities to business functions. Today, business functions are often called business capabilities. And many enterprise architects regard their business function/capability hierarchy/map as the fundamental Enterprise Architecture artifact. They relate data entities, use cases, applications and technologies to the functions/capabilities.
In 2006, the popular book Enterprise Architecture As Strategy[16] reported the results of work by MIT's Center for Information System Research. This book emphasises the need for enterprise architects to focus on core business processes ("Companies excel because they've [decided] which processes they must execute well, and have implemented the IT systems to digitise those processes.") and to engage business managers with the benefits that strategic cross-organisational process integration and/or standardisation could provide.
A 2008 research project for the development of professional certificates in enterprise and solution architecture by the British Computer Society (BCS) showed that enterprise architecture has always been inseparable from information system architecture, which is natural, since business people need information to make decisions and carry out business processes.[9]
In 2011, the TOGAF 9.1. specification says: "Business planning at the strategy level provides the initial direction to enterprise architecture."[17] Normally, the business principles, business goals, and strategic drivers of the organization are defined elsewhere.[9] In other words, Enterprise Architecture is not a business strategy, planning or management methodology. Enterprise Architecture strives to align business information systems technology with given business strategy, goals and drivers. The TOGAF 9.1 specification clarified, that, "A complete enterprise architecture description should contain all four architecture domains (business, data, application, technology), but the realities of resource and time constraints often mean there is not enough time, funding, or resources to build a top-down, all-inclusive architecture description encompassing all four architecture domains, even if the enterprise scope is [...] less than the full extent of the overall enterprise."[18]
In 2013, TOGAF[19] is the most popular Architecture framework (judged by published certification numbers) that some assume it defines EA.[9] However, some still use the term Enterprise Architecture as a synonym for Business Architecture, rather than covering all four architecture domains - business, data, applications and technology.
Since Stephen Spewak's Enterprise Architecture Planning (EAP) in 1993 – and perhaps before then – it has been normal to divide enterprises architecture into four architecture domains.
Note that the applications architecture is about the choice of and relationships between applications in the enterprise's application portfolio, not about the internal architecture of a single application (which is often called application architecture).
Many EA frameworks combine data and application domains into a single (digitized) information system layer, sitting below the business (usually a human activity system) and above the technology (the platform IT infrastructure).
For many years, it has been common to regard the architecture domains as layers, with the idea that each layer contains components that execute processes and offer services to the layer above. This way of looking at the architecture domains was evident in TOGAF v1 (1996), which encapsulated the technology component layer behind the platform services defined in the "Technical Reference Model" - very much according to the philosophy of TAFIM and POSIX.
The view of architecture domains as layers can be presented thus:
Each layer delegates work to the layer below. In each layer, the components, the processes and the services can be defined at a coarse-grained level and decomposed into finer-grained components, processes and services. The graphic shows a variation on this theme.
In addition to three major framework components discussed above.
An ideal EA framework should feature:
Most modern EA frameworks (e.g. TOGAF, ASSIMPLER, EAF) include most of the above. Zachman has always focused on architecture description advice.
The application and technology domains (not to be confused with business domains) are characterized by domain capabilities and domain services. The capabilities are supported by the services. The application services are also referred to in service-oriented architecture (SOA). The technical services are typically supported by software products.
The data view starts with the data classes which can be decomposed into data subjects which can be further decomposed into data entities. The basic data model type which is most commonly used is called merda (master entity relationship diagrams assessment, see entity-relationship model). The Class, subject and entity forms a hierarchical view of data. Enterprises may have millions of instances of data entities.
The Enterprise Architecture Reference Traditional Model offers a clear distinction between the architecture domains (business, information/data, application/integration and technical/infrastructure). These domains can be further divided into Sub domain disciplines. An example of the EA domain and subdomains is in the image on the right.
Many enterprise architecture teams consist of Individuals with Skills aligned with the Enterprise Architecture Domains and sub-domain disciplines. Here are some examples: enterprise business architect, enterprise documentational architect, enterprise application architect, enterprise infrastructure architect, enterprise information architect, etc.
An example of the list of reference architecture patterns in the application and information architecture domains are available at Architectural pattern (computer science).
A view model is a framework that defines the set of views or approaches used in systems analysis, systems design, or the construction of an enterprise architecture.
Since the early 1990s, there have been a number of efforts to define standard approaches for describing and analyzing system architectures. Many of the recent Enterprise Architecture frameworks have some kind of set of views defined, but these sets are not always called view models.
Perhaps the best-known standard in the field of software architecture and system architecture started life as IEEE 1471, an IEEE Standard for describing the architecture of a software-intensive system approved in 2000.
In its latest version, the standard is published as ISO/IEC/IEEE 42010:2011. The standard defines an architecture framework as conventions, principles and practices for the description of architectures established within a specific domain of application and/or community of stakeholders, and proposes an architecture framework is specified by:
Architecture frameworks conforming to the standard can include additional methods, tools, definitions, and practices beyond those specified.
Nowadays there are now countless EA frameworks, many more than in the following listing.
Enterprise architecture frameworks that are released as open source:
