It’s a new year and another year passed without a metadata solution. Gosh, a vendor promised me this in 1996 in their next release. Is 14 years too long to wait? Can’t technology put this puzzle together?
With most of my readers being seasoned data professionals, you know what I am talking about. It’s not that my colleagues and I haven’t solved some of our metadata problems. A good bit of our metadata is comprehensive and does meet our business needs. The big pink elephant in the room is the fact that these pieces of metadata aren’t fully integrated and traceable from end-to-end.
End to end integration… just about every metadata vendor promises that. They can show up on your doorstep with a laptop that paints a rosy visual picture of metadata flowing from source to target. Just buy, install and configure their product, and in a few days all of your metadata woes will be solved. We all realize that when that demo becomes real life in our shops, the seamless integration just slips away.
Why is it so hard to integrate metadata? Existing IT applications and environments; business processes and priorities; and corporate policies and infrastructure factor into how an enterprise sees metadata. Designing software to account for all of these factors and deliver results in a consistent manner is difficult. I’ve personally seen great improvement on the metadata tool front, but no star has emerged.
Established organizations must deal with a tangled heap of legacy applications and databases, as well as, newer web and distributed applications. All of these are spread across complex infrastructures from mainframes to Linux servers. Data is transformed into information as it passes through COBOL, Java and other programs. As we try to add these into the big metadata picture, the ability to deliver that unified view becomes more difficult and unobtainable.
I personally believe that a comprehensive, fully integrated metadata management solution will not be a reality in the remaining years of my career. This is where I’ve set my expectations to align closer with reality. What I would like to see in the coming years are metadata solutions that can address at least one major area of focus such as tracing a data element from a data model through an analytical report column. It should be able to work across software from multiple vendors, be easily maintainable, and offer a user interface that is both technical and business friendly.
I think it is acceptable to have multiple metadata solutions in a single shop. What is most important about this approach is that each solution should offer business benefit and not overlap and confuse the use of the complimentary metadata solutions in place. Loose integration is a nice-to-have but not expected.
For metadata to take a valued place in our organizations, it needs to serve a definite purpose. In other words, it just can’t sit there and look pretty. A common complaint about implementing a metadata solution is the fact that no one will pay for it. I suggest the root of this issue is not knowing the real purpose that metadata serves in your organization. Provide valuable metadata that is used daily by your coworkers, and you will find the funds to make it more valuable. Justifying real cost savings through its use is a must.
For now, I’ll keep improving the value of the metadata I maintain; align closer to my customer’s metadata needs; and work with my colleagues on integrating our metadata whenever possible. We may implement some new tools in this coming year. As a realist, I know some will be a success and others may not be. Just like a responsible driver, I’ll be a responsible metadata driver in my role.
I’ve reached 2NF in “Why Be Normal?” How normalized are you?
Find out more about “Why Be Normal?” at http://erwin.com/whybenormal/. Want to know what “We Be Normal!” is all about? Visit https://communities.ca.com/web/ca-modeling-global-user-community/ to get the whole story.
Twitter: @ERwinModeling and #YBNormal