Live from Brainstorm - Integrating Architecture into Development - James Taylor's Decision Management

« Live from Brainstorm - Enabling the agile enterprise | Main | Live from Brainstorm - Making the transition to services engineering »

September 19, 2006
Live from Brainstorm - Integrating Architecture into Development

I am attending the Brainstorm BPM/SOA/Rules event in Washington DC this week and blogging as I go.

Second session was Richard Burk from the OMB (Office Management Budget)on Integrating Architecture Into the Development of Cross-Agency Lines of Business. Dick leads the development of a business and technology framework for e-gov.He is focused on getting different agencies to use common elements of enterprise architecture across the large IT budgets involved in government. He noted that he has a Board of Directors of 535 (House+Senate) which explains a lot :-) Seriously though, he went on to explain how priorities tend to be set politically and how this has implications for architecture as these priorities drive 10,000 programs that can easily be isolated and so create problems. Most current issues in fact cut across agencies and involve cooperation between agencies, state and local government and commercial bodies. Rarely does an initiative or bill manage this, hence the OMB architecture work on Enterprise Architecture to make these things work together. He recommended "Enterprise Architecture as Strategy" by Ross, Weill and Robertson as an overview of this approach.

One of the first issues is one of semantics about what everyone is talking about. He introduced the Federal Enterprise Architecture reference models - Performance Reference Model (what am I trying to achieve), Business Reference Model (what do I do and for whom), Service Component Reference Model (what services do I need), Data Reference Model (what data for these services)and Technical Reference Model (technologies to enable all of this). A functional view of the federal government based around the Zachmann Framework (albeit slightly simpler). He explained some architecture principles that drive overall choices for federal government programs - these were great, nice and high-level but good guidance e.g. "the federal government is a single, unified enterprise" or "the federal government focuses on citizens". Seemed very "well, duh" but I liked them as it amazes me how often this kind of stuff is not written down or agreed. Most companies could take a lesson from the Feds here. Dick gave some great examples of how a simple statement drives behavior by drilling into the "federal architecture is mission-driven" and showing how it leads to things like only monitoring improved (not as-is) processes for instance. He also talked about how architects must focus on the line of business - what is your business?

To "eat the elephant" they drive down into various service areas - business services that support an agency inside and citizen services that face out (each agency is impacted by some citizen services) as well as cross-service facilities like records management or security. These can then be applied iteratively to the agencies and their lines of business and work both top-down (adopting standards) and bottom-up (driven by needs in specific areas) to derive an Architecture Portfolio. Specific investments are made in the context of these architecture outlines to create an Investment Portfolio. Planning results in a Transition Strategy and all of this must have end to end governance. Separating the architecture from the budget process and getting it up front to guide everything and focus investments was key and took a while, especially given the desire to hang on to money allocated to specific initiatives. One of the nice things he talked about was various aspects of initiatives to see how well they are going. They were:

Completeness (how much of it is done)
Usage (who uses it, for what, what percentage)
Results (what return on this investment)

I liked this as it seems to me to be a great set of measures for any initiative. This year he has a lot of cross-agency initiatives (24) and they are developing a catalog to show how these support the services each agency uses while giving each agency 6 months to apply it and figure out how to request the money they want for NEXT year. OMB has to verify that the investments match to the architecture so that the budget process is linked with the architecture. This stops the architecture from becoming an interesting experiment and makes it part of how agencies build and use budgets. Very cool.

This did not really fit with any of my categories but I thought the legacy modernization was the best, given how effective his approach seemed likely to be as a way to manage large scale legacy updates.

Technorati Tags: architecture, government, legacy modernization, portfolio management