This preview shows page 1. Sign up to view the full content.
Unformatted text preview: LAP systems are:
1. MOLAP (Multidimensional OLAP) and
2. ROLAP (Relational OLAP)
The analytical ability and the interface of these two architectures are the same,
what differs is the way data is physically stored. In MOLAP, data is stored in a proprietary multidimensional database (MDDB) to be viewed in a
multidimensional manner. On the other hand, in ROLAP, data is stored in a
conventional relational database and query languages (like SQL) are used to
perform analysis on the data in the data warehouse.
The relative advantages and limitations of the two architectures are as follows:
1. Pre-identified queries are easier and faster to process with MOLAP than with
ROLAP because the multiple dimensions of the multidimensional database can be
properly structured to handle such queries efficiently.
Since MOLAP architecture stores the same data in the form of multiple
dimensions, the storage space requirement is more in case of MOLAP than
ROLAP. Hence MOLAP architecture has a limit on the physical database size it
can handle. ROLAP architecture is scalable up to several terabytes.
MOLAP architecture can handle a limited number of dimensions. ROLAP
architecture does not have any limit on the number of dimensions it can handle.
A proprietary multidimensional database is essential for MOLAP. ROLAP
can be implemented using conventional relational database technology.
Star and Snowflake Schemas
The two commonly used schema models for designing data warehouses are:
1. The star schema, and
2. The snowflake schema
In the star schema model, the data warehouse consists of one fact (major) table
and many dimension (minor) tables. The fact table stores the primary keys for all
dimension tables. A simple example of a star schema is shown in Figure 16.17.
Here sales is a fact table, whereas customer, product, time, and store are
dimension tables. The design appears like a star, with the fact table in the center,
and dimension tables connected to it by radial lines. The main advantages of star
schema are that it is easy to define, reduces the numbe...
View Full Document
This document was uploaded on 04/07/2014.
- Spring '14