Establishing 'Making it Happen' as a 'Formal & Predictable' Discipline

'Change-Readiness' Sub-Section

ENCYCLOPEDIA→ Vendor- IT Tool Domain → - Data Management Tools → - Metadata Management Tools →

Metadata Tool Change Management

A Metadata tool should provide adequate support for metadata projects, change management, metadata models and ease of use.

Project Support

Metadata repository should be able to be shared by multiple projects, with audit trail maintained on usage and access. These projects fall in three categories-

The ones creating or changing the metadata
The ones which are using the metadata.
Both of the above

Most frequent category is the last one, whereby a project is creating and using metadata. The reason is that any new initiative will surely add to the metadata due to all new technical and business information it will create or change. For example a new leads management system would use Metadata repository for customer master table structure, and will add all the new metadata it has created.

A metadata tool should be able to manage the metadata, which is shared across multiple systems/projects vis-a-vis the one, which is specific to a given initiative/system. One should be able to specify the sharing (and non-sharing) rules.

Metadata tool should be able to control various versions of metadata during the project. This include

Development Metadata
Testing Metadata
Implementation Metadata

Sometimes during development phase, one needs to maintain a separate version of the metadata for a project. This version is later merged with the main metadata (this has a crude similarity to the code merge in parallel development tracks).

Metadata Models

This is among most important factors. A metadata tool will be using metadata models (just like a data warehouse uses dimensional models for data) for designing on how to store the metadata. It should be having following capabilities:

Pre-defined metadata models for technical and business metadata to accelerate the implementation. A good tool will be having dozens (if not hundreds) of standard meta-models. These meta-models will be based upon industry best practices. Some examples of standard meta-models are:

Data Table model (model to store information on Relational Database tables)
Program specification model (model to store program specs)
Business Rules model (model to record business rules as part of business metadata)

Extensibility of the meta-models: These Meta-models should be open to change as per the organization specific need. For example- You may add two more attributes to the pre-defined 'Database table model'. While many tools allow this, the key question to ask is on how easy is it to implement? And what are the controls around this model change? In many tools it becomes an arduous task.

Associated ETL and Exchange Interfaces: A tool should be ideally having the predefined exchange mechanisms in line with the predefined meta-model.

Developer Support

Metadata tool should ideally use the open and popular programming languages like visual basic
Metadata tool should be able to work with the standard DBMS platforms and support all DBMS related functions.
One should be able embed SQL statements in the metadata programs.
Metadata tools should be able to generate the code as user works on a wizard (say for Metadata ETL)
A metadata tool should allow multiple developers and designers to work simultaneously.

Collaborative Support

Collaborative support comes with collaboration and communication tools like e-mail, pagers.
Solution may be able to line to a kind of work-flow solution for Metadata management and changes. For example- One can have the workflow to propose metadata change, get it approved, tested and put into production?

Supporting Business Metadata

While most of the Metadata Tools are better in supporting technical metadata, not much progress has been done in the case of business metadata. Here are some of the features one would look forward to:

Ability to create the business meta-models. Most of the metadata tools lack in the pre-defined business metadata models.
A metadata tool by itself or with the help of data integration should be able to pick data from structured and unstructured sources.
A metadata tool should have guides and administrative support for business metadata. For example, it should give me a listing of possible business metadata which I may need to create. Once I select an item (for example management policies), it should provide me with possible sources of the same, and ask me to select. In other words, it should help and guide me to create business metadata.