Databases Reference
In-Depth Information
Data
Volume
Frequency
of Update
Criteria
Cost
Consumption
Availability
Tier 4 Platform - Enterprise Decision
Support Platforms
Te radata / Oracle / DB2
Large amount
of users
Tier 4
High
Large
High
Intra-Day
Tier 3 Platform - Reporting and
Analytics Platforms
Columnar DB / Appliance
Lower
compared
to tier 1
Mixed
volume of
users
Lower
compared
to tier 1
Tier 3
High
Daily / Batch
Tier 2 Platform - Data Exploration
Platforms
Hadoop / MySQL / Columnar
Lower
compared
to tier 1
Mixed
volume of
users
Lower
compared
to tier 2
Tier 2
Extra Large
Batch
Tier 1 Platform - Enterprise Data
Repository
Hadoop
Lower
compared
to tier 1
Lower
compared
to tier 2
Low volume
of users
Near Real
Time
Tier 1
Very Large
Tier 0 Platform - Data Storage
Platform
Hadoop
Lower
compared
to tier 1
Lower
compared
to tier 2
Low volume
of users
Monthly
Batch
Tier 0
Very Large
FIGURE 14.2
Tiered technology architecture.
solutions. A data map was developed to match each tier of data, which enabled the integration of other
tiers of the architecture on new or incumbent technologies as available in the enterprise.
Once the data architecture was deployed and laid out across the new data warehouse, the next step
was to address the reporting and analytics platforms. The multiple reporting platforms were retired
and three new platforms in total were selected for enterprise analytics and reporting, which resulted in
a new program for managing this migration.
The migration and implementation of this new architecture was deployed as a multiple-phased
migration plan and it lasted several phases and programs that were executed in parallel.
Outcomes
At the end of about ten months into the migration to the new data architecture platform for the data
warehouse, the enterprise began the implementation of the key business drivers that were the underly-
ing reason for the exercise of the data platform. The benefits of a customer-centric business transfor-
mation provide the enterprise with immense business benefits that were measurable results in terms
of improvement in profitability, reacquisition of customer confidence, ability to understand customer
sentiment beyond the call center, ability to execute campaigns with predictable outcomes, manage
store performance with deeper insights on customers and competition, and perform profitability ana-
lytics by integrating market behaviors to store performance. All of these activities are performed on
both Hadoop and RDBMS platforms in different stages and phases of analytics and reporting. The
