| | 9 OCTOBER 2018At the federal, or enterprise level, EA will focus on core, foundational elements of the architecture, including security, information management, master data and systems of record (to borrow from the Pace Layering view). These almost become architectural "policies", which allow EA to focus on a much smaller set of standards and patterns, i.e., what matters most. Whereas at the state and local levels, EA governance may allow any technology that is best suited for the businesses' needs, so long as no federal or state policies are violated. Sure, affordability, complexity, interoperability, etc. are still important measures, but perhaps, not the over arching focuses they once were.The digital enterprise requires more speed and agility. Providing more prescriptive guidelines, instead of prohibitive rules, ensures a balance between business flexibility and enterprise compliance.However, as governance shifts, visibility to the entire technology ecosystem becomes more critical. EA must have the tools and ability to see what is being used, where and by whom. Communication about technology and data decisions needs to be shared across business units and - a federated model will require more communication and collaboration across teams. At Cardinal Health, EA produces several dashboards to measure and monitor "complexity" for each business unit, such as application counts, health, total cost of ownership and a variety of other measures filtered by unit and business process. CIOs have incentives to drive annual improvements in their complexity scores, for example: reduce cost, lowerapp counts, improve health, etc. EA helps drive the improvements, but ultimately, the IT solution teams are accountable. In turn, EA focuses on the rules of the road and the architectural policies and leaves everything else up to them.As cloud tools change every day, there is not enough time nor capacity to evaluate every single tool and data set. Rather, EA should focus on the decision trees to provide guidance to the teams making the technology decisions. As an example, these decision trees could help determine when to use cloud-native vs. cloud-agnostic solutions ­ balancing operating costs with switching costs if cloud-portability is a requirement.Adopting this type of model provides a nice balance of governance and control with flexibility and speed ­ shifting where and how decisions are made to where they matter most. This shift allows EA to focus less on reviewing and approving "all" technology decisions, and spending more time on the patterns and guidelines of how technology should work within the environment. That said, EA cannot abdicate its role in helping select and recommend technologies. Decisions at the state and local levels are made with the business units or teams, and not for them. Ideally, EA provides teams more flexibility and control over their technology decisions, while also shifting accountability for the impact of these decisions to the teams making them. Communication about technology and data decisions needs to be shared across business units and - a federated model will require more communication and collaboration across teams Jeff Greer
< Page 8 | Page 10 >