MDD / Multidimensional Databases for Not Yet Techies

MDD / Multidimensional Databases are optimized for data warehouse and online analytical processing (OLAP) applications. It allows users to ask more complicated questions. For instance, a standard relational database of a company's customers will cross refer their customer number, name, address, date of last order etc. You can look up a lot of information that way, but it's only good for relatively low level information. Say, a salesperson wants to know the last time a customer ordered, before he goes to a sales meeting with them.

Or in what state do the most customers live? What is their average order and how many of them have bought books online? This is complex information that could have tremendous impact on the evaluation of past marketing campaigns and the designing of future campaigns. A company could decide the best use of their marketing dollars is to be buy ads in Southern newspaper rather than national TV commercials. MDD / Multidimensional Databases can help them.

MDD / Multidimensional Databases can deliver impressive query performance by pre-calculating or pre-consolidating transactional data rather than calculating on-the-fly. They pre-calculate and store every measure at every hierarchy summary level at load time and store them in efficiently indexed cells for immediate retrieval. This means that MDD / Multidimensional Databases require a huge amount of processing storage. They are not for small businesses who just need to keep normal accounting and sales records on their customers. They are also not well-suited for large volumes of data (over 50GB).

To manage large amounts of data, MDD / Multidimensional Databases servers aggregate data along hierarchies. This also provides navigation by zooming in and out of detail.

Next: MySAP