Wideway's DW solution
is subject oriented; add the data over the time; and non changeable
data collection. Wideway's DW solution design is according
following methods:
Project
Plan
Project plan is very important for defining the project objective
and scope, including project evaluation and adjustment of process,
which is a key step for early stage of data warehouse project.
During this stage, focuses are preparation of resources, technical
team, task distribution of the project and planning. This process
is repeated one.
Business Requirements Analysis
Business requirements analysis is very important stage during
the development of data warehouse. Good business requirements
analysis will make the success rate very high. The designer of
data warehouse must understand the key elements which drive the
business and turn it into the effective model of data warehouse
design.
Dimensional Model Design
By requirements analysis,we can determine the data used by end
user for business analysis. There is basic difference between
data model design of data warehouse system and that of general
operational database. This stage is the core stage of the whole
project, during which we will decide the blueprint of data warehouse
system.
During this stage,
granularity of fact table, related dimensions, properties, level
of path of drill down and logic design etc. Logic database includes
right table structure design, relationship between main key and
out key etc.
By these kinds of design,
we will still export source-to-target data(mapping).
Physical Design
Physical design of database focuses on turning logic design into
dimensional physical structure. During the stage, main job is
to define naming criterion and configure database environments
as well as initial index and partition policy.
ETL
Data ETL part is the least noticed one during data warehouse project.
This process mainly includes three parts: extraction, transformation
and load. Extract will easily expose the hidden data errors in
the operational database system. The quality of data decides the
reliability. We need to solve the data quality problems during
ETL stage.
System Structure Design
Various technologies need to be integrated in data warehouse environment.
Technology system structure design is for setting up total technical
framework and vision. During the establishment of technical system
structure of data warehouse, three factors need to be considered:
business requirements, current technical environment and planned
technical future.
Select the Product and Testing
After finish the DW structure design, we will need to estimate
and plan the needs (such as hardware, data management system,
ETL tool and date display tool) for the whole process of the DW
project. Once the tools have been decided, we will also need to
install and test them.
The End-User Application Specification
Normally, it is no needs for every end-user to visit DW, so we
will design certain applications for the end-user to use. Those
applications' design will mostly describe tables, all variables
and other computing needs.
The End-User Application Development
Through the application design documentation, the end-user application
development will have all the variables that user desired and
ability to create different tables.
Deployment
Deployment is put technology, data and application together to
give the end-user a complete package of a system.
At the end of the design
phase, we will provide following documentations:
A documented BDM.
A documented physical schema that is customized for your
chosen RDBMS and data characteristics.
A data mart loaded with an actual data load from your own transactional
systems.