SAS/MDDB Server Administrator's Guide

Why Should I Use MDDBs?

There are several advantages to creating and using MDDBs, including the following:

Using MDDBs with SAS/EIS software allows users to retrieve the underlying detail data that make up a multidimensional view.

Integration

The OLAP functionality achieved with the SAS/MDDB Server extends the decision support environment by providing multidimensional views of business data within a data warehouse architecture. This means superior integration with warehouse managed data, analytic capabilities, and presentation capabilities.

Optimized Storage and Performance

SAS/MDDB Server software minimizes the amount of space needed to store multidimensional data. This reduction is due largely to the aggregation of relational warehouse data into dimensions in the MDDB, but other factors include

Controlling the storage of statistics. The size of an MDDB can be affected significantly by whether statistics are stored or derived at run time. By default, a minimum number of statistics are stored in MDDBs. Keeping the size of the MDDBs to a minimum by storing only simple statistics means that MDDBs require short build times and can be easily and quickly maintained by data warehouse administrators.
Preventing data expansion through efficient handling of sparse data. The best measurement of the efficiency of a multidimensional data model is in the number of values that it can store within a given amount of space. MDDBs do not store data in the NWAY cube or in subcubes for crossings that are not populated in the input data. Therefore, the actual input cells are smaller than the number of potential cells proportional to the sparsity of the data. By storing only the cells where data actually exists, the SAS/MDDB Server software keeps the size of the MDDBs to a minimum.
Preventing data explosion through user-defined subtables. Data explosion is related to the way in which multidimensional data operations can cause extensive increases in data volume. SAS/MDDB Server software controls data explosion by allowing users to control which aggregations are stored as subtables.
Storing dimensional labels separately from summaries. By storing labels separately, SAS/MDDB Server saves storage space for each record.

Security

Security is a key issue in the implementation of an OLAP solution. SAS/MDDB Server software provides applications-level security by enabling password protection. Also, when using SAS/EIS software to access your multidimensional data, you can use the SAS/EIS Access Control facility to implement cell-level security. For details on the SAS/EIS Access Control Facility, see the SAS/EIS Software: Administrator's Guide.

You should consider each of these factors when deciding how and when to use MDDBs and the SAS/MDDB Server in your application development environment.

Chapter Contents
Previous
Next
Top of Page