Enterprise metadata management

Today we are facing an increasingly connected world forcing the software industry to adapt to changes and new requirement almost in real time. One of the big obstacles is complexity. Complexity creates a blurred picture and in fact an incomplete view of facts.

Another major problem is the lack of good tools for visualization. Enterprise metadata becomes secondary and less important. The biggest work is to untangle the whole system to get a clear picture.

In this article we are untangling a large enterprise application of a swiss bank and zoom in tohighlight some interesting pain points.

But first let’s embrace complexity!
Simplicity often is located on the other side of complexity. So let’s dive into it.

Extract and loading the metadata into NEO4J
First step is easy …..extracting the metadata from all the involved systems.

Get the BIG picture
After loading the metadata into Neo4J let’s have a first look at the connected data.
The usually grey metadata becomes color 🙂


Zooming IN :  Finding the root(s)

Finding the root and the dependencies of objects are essential to determine the impact and calculate the costs of software changes.

Change Request : One Column (Account_RefNr) needs to be changed

What is the impact of such a ’small‘ change?
Surprisingly the Column (red) is occurring 9 times in 9 Tables (blue) and is related to 20 View (green).


Zooming IN : Finding patterns

Use Case : What are the worst cascaded Views?

  • These View are mostly the root of performance problems
  • These Views are simply not maintainable and therefore candidates for redesign