www.qbx-tool.com

QBX: A CASE Tool for Data Mart Design

Antonino Battaglia Matteo Golfarelli Stefano Rizzi

University of Bologna - Italy

Rome - Italy

2

Motivation

! Data warehouses have become an essential element for strategic analyses

! Data warehouse systems are characterized by a long and expensive development process that hardly meets the requirements of today's market ! This is one of the causes behind the low penetration of data

warehouse systems in small-medium firms, and behind the failure of whole projects

! One of the directions to increase the efficiency of the data warehouse development process is to automate the design of data marts ! Some research prototypes of CASE tools have been

developed ! Commercial tools (such as Oracle Warehouse Builder) are

oriented to a single platform and should be considered as design wizards rather than CASE tools

3

Overview

! QBX is a CASE tool resulting from a close collaboration between academy and industry

! It includes two separate components: ! QB-Xpose

•  used by designers for conceptual design, logical design, and deployment of ROLAP data marts in the form of star/snowflake schemata

! QB-Xplore •  accessed via browser by business users and technical

experts to interactively explore project-related knowledge at different levels of abstraction

4

Features

! A basic feature of QBX is that of using conceptual schemata for multidimensional design ! Conceptual modeling provides a high level of abstraction in

describing the multidimensional repository, aimed at achieving independence of implementation issues

! It is recognized to be the necessary foundation for building a data warehouse that is well-documented and fully satisfies user requirements

! It relies on a graphical notation that facilitates writing, understanding, and managing conceptual schemata by both designers and business users

Dimensional Fact Model

5

Functionalities

! Conceptual Design ! Automate conceptual design by choosing relevant facts on a

source operational database schema and letting QBX draw your hierarchies (supply-driven approach)

! Or you can create your data mart conceptual schema from scratch by composing existing hierarchies and reusing conformed dimensions (demand-driven approach)

! Logical Design ! QBX translates your data mart conceptual schema into a

logical (star/snowflake) schema following your design preferences and taking your data volume into account

! QBX will then support you in deploying your data mart by generating SQL code for your relational DBMS and writing meta-data onto the Mondrian and Microstrategy OLAP engines

6

Functionalities

! Reverse Engineering ! Draw the conceptual schema of an existing data mart

starting from its relational schema ! Data Mart Interoperability

! Import and export your conceptual schemata to OLAP engines

! Project Documentation ! QBX creates the documentation for your data mart, including

fact schemata, conformed dimensions, glossaries, data volumes

! Publish expressive and easy-to-read documentation on the web to be explored and annotated by business users and technical experts

7

Forward engineering

CONCEPTUAL DESIGN

LOGICAL DESIGN

Requirements Design Options

Logical Schema Data Volume Operational DB

DEPLOYMENT

OLAP Meta-data

Data Mart

QBXpose

Project Documentation

QBXplore Designer

Technical Expert

Business User

WEB BROWSER

Conceptual Schema Logical Schema

8

Reverse engineering

IMPORT

OLAP Meta-data

Data Mart

QBXpose

Project Documentation

QBXplore

Technical Expert

Business User

WEB BROWSER

Conceptual Schema Logical Schema

9

Conceptual design

10

Logical design