DAMA + Wilshire Meta-Data Conference - Data Strategy Track
Intel recognized the need for an enterprise metadata container and implemented a repository to support data management. This presentation is a case study from Intel Corporation of how the metadata processes and the metadata repository are used to capture and show tangible evidence, in terms of $-value, that the productivity, governance, and alignment objectives of the enterprise architecture program are being realized. It includes a description of the metadata processes that support the creation and packaging of asset building blocks for work products and services. It also includes the philosophy and techniques for valuating and capturing reuse of these assets. It addresses the role of governance and the enterprise architecture framework, as well as the use of taxonomies to aid cataloging, administering, and searching for unstructured information resources, assets, and services. The presentation will discuss how the metadata repository is a key enabling technology for SOA by providing a discovery mechanism that enables consumers to find the reusable services that they need. Finally, it provides Intel’s experience about positive returns for the enterprise architecture program. Intel was recognized for its metadata program in 2004, when it received the Wilshire Award for Outstanding Metadata Implementation.
  •  Relationship between metadata, repository and architecture. How we tie each strategy together.
  •  Our federated repository approach.
  •  Our metadata and asset reuse framework methodology
  •  How we determine appropriate standard work products.
  •  How we are able to achieve buy-in from the Architecture community.
  •  How we align architecture work products to Enterprise Architecture.
  •  How metadata about services is captured in the repository, and utilized at design time by architects.
  •  How metadata and asset reuse is integrated with architecture governance.
  •  How the SOA repository improves programming productivity and increases re-use of software assets
Why do most Metadata Repository implementation projects fail? There are various reasons why this occurs. Lack of "metadata" knowledge. Data managers attempting to implement metadata projects. And of course politics. But the number one reason is that we usually try to deliver something on a grand scale, get bogged down and fail. As a result Metadata Repositories get a bad name.

Bob Schork has implemented both Platinum, Rochade Repositories and a home grown metadata usage repository.

Bob will cover the following:
  •  The overall difference between build verses buy
  •  The problem with repository implementation initiatives.
  •  Analyzing and implementing good Metadata Standards.
  •  Describe what a Metadata Usage Repository is.
  •  How to set up Scanners and populate your Metamodel.
  •  Describe Scanner pitfalls and how to get around them.
As interest in repository technologies continues to grow, many organizations are becoming interested in how best to approach the selection process. We have found that a "try before you buy" approach is very helpful. Organizations often take a somewhat by-the-book approach and designate – say 2007 – as "the year for studying our organizational metadata needs." The theory being that well understood organizational needs will better inform interactions with various repository vendors. This session suggests and describes an additional task – building a prototype repository that you can try out and gain experience with – as a compliment to the metadata study. Our experience has shown that organizations can benefit greatly from a weeklong intensive effort to build a "prototype" metadata repository. By picking an obviously useful area, building a "quick and dirty" repository, and implementing a plan to keep it current, organizations can greatly compliments the theoretical studies and are much better prepared for future repository related decisions.

After 35+ years there have been few sustained successes for enterprise metadata repositories. Why is this?

The need for data dictionaries—currently known as metadata repository—was recognized at the dawn of the database age in the early 1970s.

The purpose then as now was to keep track of all the artifacts—data elements, screen layouts, schemas, subroutines, programs, files, and systems—PLUS the relationships between said artifacts.

The need for this bill-of-materials-impact-analysis database engine is even greater today now that applications are spread across many hardware platforms, operating systems, and hundreds of software languages.

A well populated repository serves as a reliable source of corporate systems knowledge and memory.

Without a populated and accessible metadata repository no one can reliably answer senior management 's question: "How many places in our systems are we exposed to privacy breaches where we use unencrypted social security number?"

This presentation will be an overview of why metadata repositories consistently failed in the past, what are the lessons to be learned from those failures, and what you can do to avoid repeating the same costly mistakes.


Close Window
Wishire Conferences DAMA International