Before I begin, I must confess that this is not my original idea. That distinction goes to Gerhard Beck, a friend and colleague whose disregard of pieties in the enterprise architecture field has provoked a great deal of thought and argument (in the best sense of that word).

The 19th century military theorist Carl von Clausewitz asserted that no plan survives first contact with the enemy. By the same token, no architecture survives first contact with the developers. Events intervene–requirements change, components do not work as anticipated, new capabilities come to market–all these things collude to make the as-built system vary from the as-designed system to one degree or another. And that does not even take into account how rarely developers actually consult the architecture while constructing a system. And no one goes back to update the architecture document to reflect the as-built system. That kind of update costs money and is not seen as a value-added activity by program managers who are trying to keep costs and schedules under control.

So systems never reflect the architecture. The moment development begins, the architecture is a historical artifact. The only architecture that matters is the one on the network. And that one is not static; it is always changing.

Traditional architecture frameworks do not recognize this fundamental fact. All traditional architecture frameworks derive from the original Zachman framework. Conceived in a time when system development projects were major, expensive undertakings, the Zachman framework does not suit all modern architecture needs as well as it once did. It’s a fine framework for new-start systems, but it just is not responsive enough to rapidly changing technology and business needs for the modern IT ecosystem. In an age of Agile and DevOps, traditional architecture frameworks are at a disadvantage.

So, if traditional architecture frameworks are not helpful in understanding the operation of a modern, dynamic enterprise, what is?

Most large enterprises have deployed host-based monitoring agents as part of their cybersecurity strategy. These agents gather a wealth of important architectural data and forward it to a central server. Where it is devoutly ignored by the cybersecurity experts who are trying to protect the enterprise. Do not get me wrong–that is not a criticism of the cyberwarriors out there. I am merely pointing out that they have other fish to fry. There is a wealth of data about the real. operational architecture that can be reused to understand what is really happening in the enterprise. Understanding that ecosystem is where enterprise architects need to focus their efforts.