Data Warehouse
“Data Warehouse” can be understood as a data storage. It is used to archive data for some
time and to make it available for various evaluations. A data warehouse is clearly queryoriented.
As a rule, instead of individual data records being read, sums over many data
records are calculated. The data model is optimized for reading many data records. It is
denormalized, i.e. intentionally contains redundancy.
OLAP
OLAP is the abbreviation for “Online Analytical Processing.” Contrary to OLTP, OLAP does
not mutate data but reads it for analysis. OLAP is an analysis tool and presents data in various
compression levels. There are two different kinds of OLAP: relational (ROLAP) and multidimensional
(MOLAP).
HOLAP stands for Hybrid OLAP and is the combination of ROLAP and MOLAP.
OLTP
OLTP stands for “Online Transaction Processing.” Data is continuously written, mutated and
read. In contrast to OLAP systems, these systems are transaction-oriented. The underlying
data model is relational and optimized for writing and reading individual data records.
Data Mart
A “Data Mart” is an extract of data from a data warehouse. The data is presented as a cube
which presents the base data as well as the compressions. With regard to data mart and cube,
the term “slice and dice” is also well known.
Slice means that the cube can be regarded from any perspective by exchanging the
dimensions or by simply examining a slice.
Dice means that a selection between the various compression levels is possible with the
option of changing between the levels using Drill Down and Drill Up.
Comments
Post a Comment