Enterprise Data Management for Transaction and Analytical Processing - Benchmarking Transaction and Analytical Processing Systems - page 16

Used only
in OLAP
In OLTP
and OLAP

OLAP

Dimensional
Modeling

OLTP

Flat File

Data
Models

Hierarchical Network Relational Multidimensional Hybrid

Database
Design

Star
Schema

Snowflake
Schema

1NF 3NF

Fig. 2.1 Data management alternatives in OLTP and OLAP systems

from all database design levels: These are normalization on the logical level,
optimization of access paths on the physical database design level, for example,
indexes and views, and specific physical storage layouts used in database systems
for transactional and analytical processing.

2.1 Data Models for Transaction and Analytical Processing

Codd [ 40 ] gives a definition of the term data model as the combination of

1. A collection of data structure types,
2. A set of operators and rules to retrieve or modify data from any part of the
aforementioned data structure types, and
3. A number of general integrity rules that define consistent database states and
changes of state.

Codd [ 40 ] emphasizes that all three parts of this definition of data models are equally
important and argues that 2 and 3 are essential to understand how a structure based
on a certain data model behaves. Without a defined set of operations any application
using a structure has to be made aware of the structures internal workings regarding,
for example, the conjunction of parts of the structure.
All examples in this section refer to a customer order scenario. Figure 2.2
presents the conceptual overview of the entities and relations for this scenario
where customers order products and the respective products are being shipped to the
customers. Each customer can place as many orders for as many products as needed