Introduction to Data Warehousing and Analysis Services 2005 - Professional SQL Server Analysis Services 2005 with MDX - page 22

expect better query performance than the case where the data is being re-
trieved from the relational database. This proprietary format helps Analysis
Services 2005 to retrieve the data efficiently and thereby improves the query
performance. Based on where the data and/or aggregated fact data is stored
you can classify the storage types as MOLAP (Multi-dimensional OLAP),
ROLAP (Relational OLAP), or HOLAP (Hybrid OLAP).
MOLAP is the storage mode in which the data and aggregated data are both
stored in proprietary format on the Analysis Services instance. This is the de-
fault and recommended storage mode for Analysis Services databases since
you get better query performance as compared to the other storage types.
The key advantages of this storage mode is fast data retrieval while analyzing
sections of data and therefore provides good query performance and the abil-
ity to handle complex calculations. Two potential disadvantages of MOLAP
mode are storage needed for large databases and the inability to see new
data entering your data warehouse.
ROLAP is the storage mode in which the data is left in the relational data-
base. Aggregated or summary data is also stored in the relational database.
Queries against the Analysis Services are appropriately changed to queries
to the relational database to retrieve the right section of data requested. The
key advantage of this mode is that the ability to handle large cubes is limited
by the relational backend only. The most important disadvantage of the
ROLAP storage mode are slow query performance. You will encounter slower
query performance in ROLAP mode due to the fact that each query to the
Analysis Services is translated into one or more queries to the relational
backend.
The HOLAP storage mode combines the best of MOLAP and ROLAP modes.
The data in the relational database is not touched while the aggregated or
summary data is stored on the Analysis Services instance in a proprietary
format. If the queries to Analysis Services request aggregated data, they are
retrieved from the summary data stored on the Analysis Services instance
and they would be faster than data being retrieved from the relational
backend. If the queries request detailed data, appropriate queries are sent to
the relational backend and these queries can take a long time based on the
relational backend.