Data Access vs. Data Management

What is the value of separating data access from data management? To answer that question, one must first understand the origin of data warehousing. Fifteen years ago, organizations were asking, "Why build a redundant database dedicated to reporting?" The answer is priority and performance. Transaction systems have historically held higher priority than reporting applications. Heavy query access impacts performance of transactional systems to unacceptable levels. In order to maintain high performance, a second repository is necessary.

The reason to separate data management from data access is also priority and performance. To meet current demands, access to analytical data must be near immediate. When users drill, data is immediately returned. Specifically, fast performing ETL does not justify slow running queries. In the practical world, both ETL performance and query performance must be fast.

To achieve adequate performance on the movement and query of data, the enterprise solution must optimize and separate each function. While this separation increases cost, this investment is essential to establishing large-scale solutions.