Databases Reference
In-Depth Information
Multiple
Organizations
Multiple Data
Warehouses
Multiple Bl
Solutions
Multiple Analytics &
Reporting Solutions
Business Units
BU DWs
P
r
o
c
e
s
s
e
s
P
r
o
c
e
s
s
e
s
P
r
o
c
e
s
s
e
s
Store
Management
Services
Retail
DWs
Financial
DWs
Financial Services
Enabling
Functions
DW
FIGURE 14.1
Current state architecture.
●
Provide access to data in a self-service platform from executives to store managers.
●
Create a scalable analytics platform for use by data scientists.
●
Reduce processing complexity.
●
Create an enterprise data repository.
●
Increase data quality.
To create a robust future-state architecture that can satisfy all the data requirements, goals, and
business requirements, the technology platforms that were considered included incumbent technolo-
gies like Teradata and Oracle, Big Data platforms like Hadoop and NoSQL, and applications software
like Datameer and Tableau.
The overall architecture was designed on a combination of all of these technologies and the data
and applications were classified into different tiers that were defined based on performance, data
volumes, cost, and data availability.
The overall benefit of this approach resulted in creating a scalable architecture that utilized tech-
nologies from incumbent platforms and new technology platforms that were implemented at a low
cost of entry and ownership, and retiring several incumbent technologies and data processing layers,
which provided an instant ROI.
The tiered technology approach enabled the enterprise to create the architecture layout as shown
in
Figure 14.2
.
The enterprise architecture team used the tiering approach to present the criteria for the differ-
ent systems to be allocated to a particular tier in the architecture layers. At the end of the exercise
the enterprise data repository was designed and deployed on Hadoop as the storage repository. This
included all the current historical data in the multiple data warehouses and other systems. To complete
the integration, new data architecture models were developed and data processing rules were assigned
to each data layer in the data architecture, along with data quality, auditing, metadata, and master data
