As a CIO, I am a strong proponent of Enterprise Architecture (EA) and the components of EA articulated by Steven Spewak. Eight years ago, I would sit with the Informatica R&D Chief Architect and describe what I needed to realize our IT architectural vision, as well as the problems I wanted to overcome.
So, what were the problems I wanted addressed?
First, I am a believer of a best of breed strategy. I fundamentally believe the “megavendors” are dictating IT strategy, yet they cannot innovate fast enough – thereby harming IT. To build a best of breed approach, I wanted to build a loosely coupled architecture. In essence, I wanted to abstract the data away from the applications we ran, thereby enabling me to switch vendors if necessary. This would enable me to provide the best solutions to our business as well as maintain negotiating leverage with my vendors. The challenge is that no technology existed to do this cost effectively.
Second, I know integration is one of the unspoken, yet high cost areas of IT. When I talk to my peers, many will cite that 30-40% of their run rate for applications is spent on managing all of the interfaces within their team. Any time someone makes a change, it takes months to implement as multiple groups need to test, validate and coordinate the update. My goal is to make this a minimal part of my run rate and contribute as much to innovation.
Third, historically there had been a morass of integration and data quality technologies. The highest cost of most IT organizations is labor, and I didn’t want my team to learn multiple tools to realize our architectural vision. I wanted a standard platform for everyone to use and not have to use multiple tools, which would require more human capital investment, more software and infrastructure cost and more vendors to manage. Our goal here is to simplify.
Fourth, I have been a strong proponent of master data and managing the master data within an organization. I’m proud to say that our IT organization has been building master data solutions for the past five years. With that said, it’s been a challenge managing data quality across multiple systems. Our goal had been to make data quality a foundational element within our infrastructure so that all applications and data movement would adhere to a consistent set of quality rules. Thus, we would have pervasive data quality within our organization.
Lastly, I was a fan of UML because it improved the quality/fidelity of requirements and traceability from specification to code. This worked well in the application development world, but there was no equivalent for the data world. I wanted a similar paradigm where business users, the true stewards of the data, could define and articulate their needs which would then automatically pass to analysts and developers. This would help with data ownership, time to market and quality of solutions.
Now enter Informatica 9. I don’t claim to be a product guru like other members of our team. What I do know is that it does enable our team to fully realize our architectural vision. Based on my conversations with other CIOs, we believe this is a game-changing platform for IT.