Consider the allegory of Sara Winchester. By the Victorian era, murderous capabilities of Winchester rifles amassed her fortune. Despite affluence, deep regret impaired Ms. Winchester. A physic medium consulted, "Build a great mansion and the slain will forgive." A four-decade project ensued. The project began and ended without a blueprint. A bonanza of work ensued for carpenters, plumbers and every tradesman imaginable. Cutting-edge, innovative technology defined the project. Breakthrough advances in heating, plumbing, lighting and elaborate detail are ever-present in Sara's reparation to the dead.
Not surprisingly, the lack of architectural planning led to some interesting "features." The 160-room mansion has multiple staircases to nowhere. Many fine Tiffany windows bedazzle the visitor, despite being set in the floor. The Winchester Mystery House is an archetypical example of a battle fought without a plan.
IT leads business through the battles needed to thrive. However, has IT planned these many pitched battles? Is IT campaigning toward improved profit through reusable, consistent architecture? In candid moments IT confesses, "Heck no." Has a bonanza of work and rework ensued? Many a general ledger would say yes.
Contemplating IA from within the walls of one's own Winchester mystery house is a common experience. Often, there is no IA plan. Is there a slogan? Perhaps the motto is buy, implement, depreciate, run for 5 more years and then repeat. Perhaps the modus operandi implements best-of-breed technologies on a project-by-project basis. There are many creeds, and perhaps the catchphrase of the day does not include IA. This does not prohibit inserting it. Regardless of the lack of IA, set a new course.
To begin, let IA tollgates exist in every phase of the software development lifecycle (SDLC). Projects are the key. Consider the ANSI Standard definition of a project: "A project is a temporary endeavor undertaken to create a unique product or service."1 By definition, something will appear for the first time because of a project. Add IA at inception. However, carefully construct the definition of IA to avoid calamity. After all, Sara Winchester was an architect.
Definitions of IA Proliferate
The creation of blueprints is a salient definitional thread. Begin with a concept, a model and a plan. Plan simply. Create and improve architecture with each project. Show that each project has thought about and has improved organizational architecture. Leave blueprints after each project as artifacts of success. Projects in the initiation phase will be pleased to have these artifacts available, especially as the last project funded. However, beware of excess. Overburdening any one project with a global IA implementation task invites failure. Hit the essential IA points at outset. Migrate incrementally toward an enterprise architectural blueprint. Let the blueprint consist of straightforward building blocks.
Begin with three building blocks: data, applications and systems architectures. Select the basic blueprinting tools for each block. Some familiar coins for data architecture are entity definitions, entity relationship diagram and entity-to-business-function mappings. Establish the foundations of application architectures on listings of all applications, matrix classifications to their business functions, appointment of Subject Matter (Business) and Domain (technical) experts, and analytical tools focused on usability and user experience. Deliver Services by mapping a canonical model across process, data, and applications. Data domains with defined ownership drive higher reuse as logic binds data, users and hardware. Systems architecture definitions arise from understanding physical infrastructure and distribution patterns of data and applications, and their business domains. Ask, "What is on this physical infrastructure? How is it backed up, recovered, replicated, moved and presented to the enterprise." IT is very familiar with these architectural methods and tools.
True, every company has an IA. The concept of blueprinting the enterprise IA as data, applications and systems building blocks is mature. Alas, not all efforts are equal, and sadly, mature systems decline. Nevertheless, each fundamental block has champions. For example, the Data Management Association (DAMA) is now creating a detailed Data Management Body of Knowledge (DMBOK). The DMBOK, a work in progress, offers a clear framework to understand the functional areas of data architecture. Generally, defining each architectural building block is essential. However, the union of all three blocks is indispensable and requires program coordination. Consider.
The E-Commerce Crown Jewel Project begins. The data architect delivers normalization, banishing data redundancy and waste. However, the application architect laments slow response times due to a five-way data join frequently called and summarily stuffed across the enterprise service bus. All this causes grief for the systems architect, now fending off demands for new hardware to speed up the bus. None of the three holistically observes the terrain. Forgetting that balance is crucial, they serve one domain. However, the whole is more than the sum of its parts. Endeavor to proportion all three building blocks by integrating blueprints. Enter the information architect offering guidance to the portfolio of projects. Most welcome the direction.
Architectural tools and good intentions stimulate technologists. However, the inception of an IA journey begins a descent to fear for some in the organization. Why fear?
Be the first to comment on this post using the section below.