Traditional Metadata Definition
Traditional definition of Metadata is ‘data’ about ‘data’. This is now outdated as the scope and usage of meta-data has gone far beyond this definition.
Revised Definition of Metadata
Metadata is ‘data, information and knowledge’ on the ‘data, information and knowledge’ of an organization, both from physical and logical perspective.
Revised and more simplistic definition
Meta-data captures all characteristics and details about an organization.
The definition explained
METADATA IS
- Data – Where it is lying, what elements it contain?, when was it created?, when was it changed?, what is the data-model? who created it?
- Information- Description of what is contains? How is it linked to the other meta-data? Any quality issues?
- Knowledge- How important is it? Is it used extensively? Where all is it used? What will happen, if this resource is not there?
ON THE
- Data- All data tables, worksheets, flat files, data warehouse databases, OLTP databases...
- Information- dash-boards, scorecards, MIS, Reports...
- Knowledge- Heuristics, documents, policies, business rules
OF AN ORGANIZATION BOTH FROM
- Physical and- Physical location in terms of servers, lockers & safes, people drawers, network drives, local drives? Physical models...
- Logical- Which roles owns it? Logical models...
PERSPECTIVE
Another way to describe Meta-Data
Though it’s difficult to find an exact analogy for Metadata, here is one effort. Meta-Data can also be considered as an equivalent of Amazon book store. If we consider each data element as a book, the meta-data will contain name of the book, summary of the book, assessments about the book, the date of publication, high level description of what it contains, who are the publishers, how you can find the book, author of the book, whether the book is available OR not. This information helps you to:
- Search for the book
- Access the book
- Understand about the book before you access OR buy it.
Some other de-mystifying statements
Metadata is not only a BI subject
Metadata Management (like Master Data Management) is not a pure BI subject, though it has many applications around BI components like Data Warehouse and OLAP. Metadata Management (as obvious from the definition) serves every possible stakeholder within an organization. Technically speaking, it can be used even by people who have never seen a computer. ERP systems are as much a stakeholder for Meta-data project as a data Warehouse.
Meta-Data is a business initiative
Metadata (like other data management and BI initiatives) is a business owned initiative. Business provides sponsorship and also most of the components of metadata. Metadata in it's extremely restricted form (whereby, only the data lying in the IT systems), may be thought of as an IT initiative. However, even for the data lying in IT systems have a major Business stake-holding. (For example, if business which is going to define on if an agent can report to a single OR multiple sales people...)
It’s true that a well-functioning metadata repository will require IT platforms, but that part comes in the ‘support’ domain and not in ‘ownership domain’
Metadata Management comes under the ownership of a Data-Steward
Metadata management (like other data management initiatives like data quality), comes under the ownership and accountability of a data steward (which is a business role). Refer Maximizing Effectiveness of Data Steward and Business Intelligence Organization- Roles & Skills for a better perspective on this role.
How Metadata is placed as an overall component of Data Management
Metadata Repository
Metadata repository contains all details on an organization management environment. These details are placed in a central repository OR in well-connected synchronized repositories (for example Data Warehouse could be having its own repository and an ERP could be having its own). An ideal situation (for which we are not aware of a case-study) is to have a single repository. When there are multiple repositories, one does run into a challenge of making sure that they are well-synchronized and integrated. There have been issues around creating a single metadata standard. Typically organizations rely on XML to integrated and exchange metadata information.
Meta-Data Model
A metadata repository has its own data model. For example there will be a data-model for storing database tables. A typical data model for database table metadata will contain fields like name of the table, location of the table, data of table creation, systems accessing the table etc...
Usage (covered more in detail on ‘why Metadata’)
Whenever one wants to have any usage of a given set of metadata, one can log into a data repository and get information on that meta-data. For example, if you are looking for Standards of Business Conduct, you may log into the metadata repository and find out on its intranet address OR the person who has the custody from where you can ask for a copy. The systems also use meta-data to conduct their operations. For example, before every extraction done by a data warehouse, it will check the meta-data on the location of the database tables of the source systems (this is called the 'operational use' of metadata).
Metadata Management
Metadata Repository is one of the outcomes of metadata management process. Metadata management is an end-to-end process for creating, enhancing and maintaining meta-data repository and associated processes. Metadata management includes establishing processes, mind-sets, organization and capabilities to build a metadata environment. Like BI and Master Data Management, bigger challenge on meta-data management is related business process discipline and culture.
|
