Data Warehouse Metadata

Data Warehouse metadata

While considering almost every question anyhow connected with data management, name "metadata" appears at least a few times. What does it exactly mean? In brief, metadata is - strictly - "data describing another data". An example should resolve the doubts.

Think about a library. What do you need to know to find a proper book in a catalogue? Firstly, its author. Secondly, book's title. Or maybe - differently - book's subject. As it was stated, metadata (author, title, subjects) describes data (concrete books).
Another example: telephone book. What is needed to find a proper phone number? Its owner's name, location. Maybe a profession. Then, metadata (owner's name, location, profession) describes data (demanded phone numbers). Back to the formal definition of metadata. While metadata describes another data, this lastly mentioned data might refer to any element of enterprise computing system - regardless whether it is data source architecture, business definitions, data flow among diverse systems, or objects relations. Let's move back to the example for only one more while - telephone book. It doesn't matter where each book comes from - it might be USA, France, Samoa Island or Cambodia. The thing is that all the books mentioned have their numbers sorted by names, locations, etc. Even though the concrete numbers differ, sorting criteria are the same. Metadata is the same, data - by no means.
As it got clearly proven, metadata leads to the unification and ease of integration of diverse databases applying to generally related data. That is why the importance of metadata grows so quickly.

Why else is metadata warehousing so important in recent management?

Firstly, along with data dependence, should also grow data trustworthiness. While many enterprises depend on multiple volumes of data, they hardly understand its sources. They see only result side of ready-to-read reports, slightly ignoring their input information. That doesn't favor neither data comprehension, nor report reliability. For example, financial statements are based on diverse information and factors and being usually accurately analyzed. Even though the general feelings might arise only because of report's final summarization, comprehending the reasons of each situation demand diving into more and more accurate data across diverse data sources. Efficient metadata warehousing, through specialized lineage tools, might facilitate decoding each part of report and lead to related data (its sources, transformations, and destinations).

Metadata warehousing helps ensuring enough IT flexibility. While business evolves every new day, along it, change computing systems to follow newer and more complex demands. Unfortunately, most of changes usually cause a lot of lateral consequences, destroying former data integrality and unity. For example, after computing system's upgrade, a need to change the method of determining ROR (rate of return) might appear. Thereupon, metadata warehousing system's role is to highlight all the objects that have to be anyhow changed. Having these objects highlighted, making change's time and money consuming forecast is significantly simplified.

Efficient metadata warehousing system is indispensable for ensuring single, coherent metadata architecture. While computing systems grows larger and larger, metadata is spread across the whole enterprise, diverse departments and units, therefore keeping their integrality becomes impossible. Specialized metadata warehousing systems support metadata managing across the whole enterprise.