Loraine Lawson spoke with Leon A. Kappelman, editor of The SIM Guide to Enterprise Architecture, which was written by the Society for Information Management’s Enterprise Architecture Working Group, about why enterprise architecture can help save modern companies from collapsing under their own complexity.
Lawson: In the press release for the book, you're quoted as saying enterprise architecture is the key to managing objectives such as business/IT alignment, agility, even legacy replacement and data integration. I wonder why you say that. Why is EA critical, particularly in regards to integration?
Kappelman: If you can't describe it, if you don't know what it is - I would say you can't manage it and you can't build it. And in order to do integration or alignment or any of those things, you have to get groups of people to know what it is in a shared way.
Architecture is the key to doing that. We've been using it for thousands of years to do exactly that. Humans did not build the pyramids by getting 10,000 slaves together and saying, "Go get some rocks and start piling them up." No, they had a vision. And yeah, the people cutting the rocks only had a small piece of it, just like the person writing a piece of code or a bridge between two programs so they can integrate data only have a small piece, but someone - some larger group - has a shared vision that that really is what we're trying to accomplish. And now we can divide that up into the division of labor that will help us create that.
“The working group defines enterprise architecture and its definition that we could all live with is, “The holistic set of descriptive representations about the enterprise over time.”
- Leon A. Kappelman
- Editor of The SIM Guide to Enterprise Architecture
Lawson: How does the book define enterprise architecture as a role, as a position within a company or organization?
Kappelman: The working group defines enterprise architecture and its definition that we could all live with is, “The holistic set of descriptive representations about the enterprise over time.” But the simple version of that is it's the models of the enterprise.
Let me back into it with starting with the alignment objective. I think IT's obsession, if I can call it that, with alignment is actually a symptom of the underlying problem in IT and really in the enterprise today. And IT is really just a reflection of the whole. It's kind of like - I'm a little overweight, so if I get on the scale every day and check my weight, it's not going to help me change my weight. It gives me a measure of the fact that I'm a little overweight, but unless I change my attitudes and my behaviors, nothing is going to change on that measurement of my overweightness. This annual obsession of, "Oh, I must align, I must align" is kind of like standing on the scale without changing our behaviors.
So what behaviors need to change? Well, it's back to that translation thing. How do we communicate? How do we get everybody on the same page? You know, how do we create that shared vision? And enterprise architecture is an important part of the answer to that question.
Lawson: Do you see it as largely an IT function?
Kappelman: No, not at all. It's about IT as part of it because IT is part of the enterprise, but it's about much more than just IT. Actually I see it as an opportunity for IT.
Jeanne Ross wrote the foreword to our book and she actually starts off by saying that IT will report to enterprise architecture. I don't necessarily see it that way because I think enterprise architecture becomes more a staff function to senior management to help them clarify the vision so that everyone can understand it. In other words, it's really about creating a shared language so that everyone can get on the same page, because if we can't get the people on the same page, then we're not going to align anything to anything; everybody is throwing rocks in a pile. You know, there's no vision. Yeah, you could slap it all together like we do today and throw lots of middleware at it and integration and set up data warehouses and spend fortunes on cleaning up the data that should have been - or at least today, could be architected to at least standardized data models enterprise-wide, but it's like building a skyscraper out of 10,000 log cabins. You might be able to do it but if a big storm comes around or an earthquake shakes the ground, it may not hold up so well.
To ShareThis, click on a service below: