Published in TDAN.com July 2005
“Enterprise Metadata Management ~ Fast Forward to 2005”
The term “Enterprise Metadata Management” has been in use since the late 1980’s when many of us worked in classic “mainframe environments” using “corporate data
dictionaries” to manage our COBOL programs, Job Control Language (JCL), screens, reports, copybooks, and database schema definitions. Fast-forward to 2005 ~ has anything changed? What does
“enterprise metadata management” mean today? I recently surveyed a number of leading metadata management vendors, Global 5000 companies, and industry thought-leaders seeking their
input.
It is Summer 2005. What do the words “Enterprise Metadata Management” mean to you?
“Enterprise Metadata Management means collecting and cataloging the ‘contextual information’ about every aspect of the enterprise: data, information, systems,
applications, processes, stakeholders, business rules, architectures and infrastructure” says Alan Perkins, Chief Solutions Architect for ASG-RochadeTM and ASG-becubicTM
at Allen Systems Group Inc. (Naples, FL). “Enterprise metadata provides the answers to: ‘What is it?’ ‘Where is it?’ ‘How did it get there?’ ‘Why do
we have it?’ ‘Who owns it?’ and ‘Who is responsible for it?’ It also answers: ‘Who can change it and what is the process for doing so?’” This
statement, which is still quite relevant today, could have also been made in the 1980’s.
“Enterprise Metadata Management is much broader in scope and definition than ever before” says Peter Aiken (Ph.D.), Managing Director at Data Blueprint, a consulting firm in Richmond,
VA. “Today it encompasses a broad range of disciplines, involving topics such as XML, business metadata, technical metadata, data warehousing, vocabularies, taxonomies, & ontologies,
master/reference metadata, GIS metadata, and even reusable source-code components. In my consulting programs and training workshops, we offer a broad range of metadata topics that extend to include
the skills required to articulate metadata to business knowledge workers and management”. This suggests that our corporate environments are more complex than before. In 2005, enterprises,
large and small, have more data, more kinds of data, and more compelling reasons for organizing that data. Simply stated, they have “more metadata”.
“Instead we turn the question back to our customers and ask: what does it mean to them?” says Guy Hoffman, President & CEO of Metallect (Plano, TX). “Our objective is to
understand how the customer wants to derive value from enterprise metadata management, whether it be to reduce the escalating costs of application maintenance, compliance, or the adoption of a
service oriented architecture. Our goal is to leverage enterprise metadata to provide increased visibility and automation to reduce costs, cycle times, and risks associated with aligning underlying
applications and their databases with strategic business initiatives.”
Charlie Wertz, former professor (sometimes author and consultant) says “I got involved with metadata but dropped the ball after publishing The Data Dictionary: Concepts and Uses in the
1980’s. Today, I feel like an odd sort of Rip van Winkle; when I awoke, I found little had changed. There is more data than ever and there are many more reasons to create an enterprise metadata
store, but it seems that the folks who write the checks are still having trouble in seeing the value proposition.”
What are some of the biggest Metadata Management Issues at Global 5000 Companies today?
“The biggest PAIN is that my customers (corporate employees) cannot find THE single, definitive answer to any of our questions about data” says Bruce McTavish, metadata manager at
Starbucks Coffee (Seattle, WA). “The more people I ask, the more answers I get. And this problem is NOT unique to Starbucks. It occurs across all major corporations worldwide.”
Perhaps this means that we still don’t understand the way data proliferates and the different ways various individuals view and use it.
“We have multiple applications and some have similar (or identical) data.” says Diane Colvin, a data architect at Radian Group Inc. (Philadelphia, PA). “There is nothing that ties
this data together so that we can more easily understand and maintain our key applications and systems.”
“We help companies to combine and transform their data into a single master database” says Anurag Wadehra, VP of Marketing at Siperian Inc. (San Mateo, CA). “A metadata-driven
framework is MANDATORY to enable companies to understand the different forms, types, and definitions that common data elements share with each other. It is important to maintain the distinction
between managing metadata through a generalized metadata tool versus having a metadata-driven framework designed for a specific purpose, such as supporting customer data integration and
master/reference data management. In my experience, the most successful companies combine ‘best practices’ from both approaches.”
“One of the biggest issues I see is to tie ‘technical metadata’ to ‘business metadata’” says Zvi Schreiber, Founder & CEO of Unicorn Systems (New York, NY).
“Unicorn has been pioneering semantic technologies to achieve the link between ‘technical and ‘business metadata’ more effectively, delivering the vision of ONE Enterprise
even when I.T. has thousands of systems and applications components.” This statement suggests that “cataloging the existence of tables, columns or files (and their content)”
is NOT the same thing as” cataloging the business meanings and USES of the information”.
“Our customers keep asking for a ‘Google or Yahoo-like’ application that can leverage a web-browser to help them find key information assets and business data components more
easily” says Geoff Rayner, CEO at Data Advantage Group (San Francisco, CA). “For our product (called ‘MetaCenter’) we hired a professional graphics designer to significantly
improve the user interface and overall user experience over other metadata products available today. We want our product to be easy to use and easy to implement so we also provide
‘real-time’ metadata interfaces to popular development tools. This helps our customers’ implementation process to be quicker and more successful.”
“One of the biggest challenges facing organizations is managing unstructured data” said Jeff Dirks, President and CEO of SchemaLogic (Seattle, WA). “Most of the information stored
throughout an enterprise is unstructured, from emails to word documents to marketing materials. As unstructured data continues to grow, organizations are having an increasingly difficult time
organizing the content for successful search and reuse. By defining and managing the structure and context of information, companies will be able to improve information ‘findability’
and capitalize on existing corporate information assets, which will ultimately help them react more quickly to changing business environments and new market opportunities.”
Within the next 2-3 years, how do you see the “Enterprise Metadata Market” evolving or changing?
“Increased integration and information sharing across the ‘enterprise’ will become more and more important” says Greg Coticchia, CEO at LogicLibrary Inc. (Pittsburgh, PA).
“We are focusing on supporting enterprise architectures, reusable components, and SOA (service oriented architecture) plus ‘proactive, cooperative integration’ with other leading
metadata solution vendors.” This will become a bigger issue as SOA proliferates. Monitoring and managing flows between the monolithic applications and programs of the 80’s and 90’s was
simple in comparison.
“To help increase the customer’s chance of true implementation success and real ROI, my company has introduced the industry’s first ‘metadata integration
appliance’” says Brian Brewer, CTO at InfoLibrarian Corp. (Rochester, NY). “We ship our metadata management solution pre-installed on a network interface appliance. This helps our
customers to catalog and access their key information assets with HOURS after installation instead of the weeks & months that traditional metadata solutions have taken. It truly increases the
customer’s value proposition.”
“Enterprise Metadata Management will become more ubiquitous” says Perkins. “More and more companies will recognize the real value of metadata.” We’d like to think so.
“Enterprise Metadata Management will become more automatic, ‘transparent’ and easier to implement for the customer” says Rayner.
Summary and Conclusion
Perhaps the biggest changes from the 80’s and 90’s are that businesses are more concerned about “more kinds of data”, such as unstructured documents, email, scanned images, web
pages, and web services. There are more reasons for being concerned, such as high-visibility projects involving data mining, business intelligence, regulatory compliance, and data security. Most
Global 5000 companies seem to be looking for better “search and discovery” products at the enterprise level. Such products enable people to find information assets more reliably,
quickly, and easily, in a self-service manner that is less dependant upon specific “corporate subject-matter experts”.
The “good news for metadata” is that there are more vendor solutions today than ever, including many new start-up companies that have been launched in the new millennium. And we are
also seeing established software companies (such as SAP and Informatica) incorporating metadata management utilities within their standard product offerings. I have personally counted over 29
different vendors who participate either directly or indirectly in the metadata management market. Several of these vendors do not officially classify themselves as “metadata management
tools” but still catalog and organize corporate metadata definitions.
Traditionally, the metadata management market has not been a large (multi-billion $$$) market, but Global 5000 companies are still spending hundreds of thousands (or even millions of dollars) per
year in their quest for easier access to corporate information. And venture capital companies are funding the development of new metadata management technologies! If your company would like easier
access and better management of your key information assets, I encourage you to take a look at the “modern metadata market”.
