Metadata Management Key to Complex Information Architectures

Loraine Lawson

Information management systems are becoming more complex, in large part due to more integration between a wider variety of sources, both within and without the organization's walls. Ventana Research, in a report on how cloud computing is changing integration, notes that organizations are already integrating an average of five or more different data sources.


That's why metadata is becoming a critical management challenge for IT shops. In much the same way we've had to consolidate and centralize master data, IT will need to centralize and consolidate metadata.


Metadata is commonly explained as the data about the data, but that definition belies how complicated metadata management can be. I once explained metadata as similar to table headings. If the table held information about customers, where you label one column "Date" and another column "Last name" and another column "First Name" and so on, "Those labels are metadata - they describe the data in the columns," I wrote.


But that's oversimplified, as Francis Carden, the founder of OpenSpan, pointed out in the reader comments:

The way MetaData is described here is as a field name. However, break your meta data further and ONLY THEN do you find the REAL problem. The spreadsheet data example needs a 4th column for me to explain. Say, it's "Balance". The data is $379,332.22, the MetaData is "Balance" and the MetaDescription is "((History_Bal-Account_Year1_Bal)/12)+(Credit_Score/fudge_Factor)". Now in order for any integration to post this data is dangerous without the business logic. Imagine if 75 percent of your metadata is derived from business logic.

Metadata is critical to integration because when you integrate data what you're actually doing is matching up the metadata, according to Evan Levy of Baseline Consulting (which is now owned by DataFlux). Experts also predict data integration vendors will focus more on supporting metadata management this year.


Metadata is also increasingly important to other initiatives including business intelligence. For example, you can get more out of a prepackaged BI solution by ensuring that it supports metadata from your back-end systems, according to a recent Virtual Circle article by Lou Agosta. Agosta writes:

If an enterprise is operating an ERP system from SAP, Oracle, or Microsoft, then look for technology with metadata that enables the back-end ERP system to be easily and rapidly configured by the BI presentation layer. In other words, do not expect a system designed to work with Microsoft Dynamic AX to provide visibility to proprietary SAP business application process interfaces (BAPIs).

If you'd like to learn more about why metadata management will be a key component to information management, you might want to check out this week's Briefing Room, which will feature IT analyst Robin Bloor. He'll also discuss the value of an information-oriented architecture and discuss low-hanging fruit you can target in your own data environment.


"Metadata Management as the Key to Agile BI" is scheduled to start at 4 p.m. ET tomorrow, June 14, and as always you should preregister. The event will also feature Ian Fyfe of Pentaho who will talk about centralized metadata management and share real-world customer examples of how the company's solutions help organizations avoid data silos.

Add Comment      Leave a comment on this blog post
Jul 23, 2011 4:09 AM Jan Jan  says:

Based on what grounds does one choose some data to be data and other data to be metadata?

How does that choice change when one chooses other grounds?

How many grounds would be there to choose from?

What, uhh,  is metadata management?

Jul 24, 2011 9:37 AM Loraine Lawson Loraine Lawson  says: in response to Jan


You don't choose metadata from data. Rather, it's the information about the data. For instance, if you have a spreadsheet or database table that stores customer names, phone numbers and addresses - it's the information that tells the computer which field is the customer name, which field is the phone number. That's very simplistic - it can also be information about who is allowed to view information, which application it's stored in. You might read this article:

Metadata management is a tool for managing that data. And the reason you would want to do that is it makes integration easier, because you can compare apples to apples, say if  you're merging two databases and they use different languages. The metadata helps you figure out what they share in common and what's different.

Check out this article for a simple explanation of both:

Aug 11, 2011 5:47 AM Jan Jan  says: in response to Loraine Lawson

"Information about the data"... isn't that data too? Labeling data as data or as metadata very much depends on ones interest/perspective... isn't it?

Isn't all data related to other data? Independent data or data 'an sich' does not exist - does it? Isn't all data in some wat meta to other data? Why do we still make such a strict/separatist distinction between the two? The strictness of the distinction is of no use in dynamic contemporary societies.

A good tool for metadata management will also be a good tool for data management and vice versa. Agree?


Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.