Databases Reference
In-Depth Information
●
Created a best-practices approach for deploying Cassandra and HBase across the different regions
of the world, where the data interfaces were deployed based on corporate standards. The design
also included cross-data center replications to ensure availability.
●
Integrated a semantic data discovery platform with interfaces to internal and external products and
service hierarchies.
●
Created and deployed a tagging process for data management and integration, especially with
unstructured data.
●
Implemented a metadata integration and reference architecture for centralized management of
metadata across the enterprise.
●
Designed and developed data parsing algorithms for integrating semi-structured and unstructured
data components.
●
Integrated algorithms for parsing machine logs and clickstream logs to understand the user and
device behaviors.
●
Integrated algorithms for predictive analytics integration.
●
Integrated algorithms for statistical modeling and integration.
●
Integrated reporting and visualization platforms into the different data layers.
●
Provided business users capabilities to discover, analyze, and visualize data across the enterprise
in an integrated platform.
The implementation process was a planned migration to the heterogeneous platform over 24
months on a global basis. The main benefits of the architecture were:
●
Design once and deploy in parallel worldwide
●
Modular scalability
●
Standardization
●
Lower cost of maintenance
●
Higher scalability and flexibility
●
Security standards compliant
●
Always available
●
Self-service capabilities
●
Fault tolerant
●
Easy recovery from failure
Benefits
Based on the revised architecture approach to implementing a new data warehouse, the enterprise has
started realizing a quick ROI on their analytics and data discovery processes. The business users have
been able to measure the performance of the enterprise and its competition with clearer insights and
have been able to start recovering some of the brand issues that were not previously salvageable.
Extending the data warehouse and reengineering the architecture was another way of building the
next-generation data warehouse. Another technique that was implemented by a leading manufactur-
ing organization was integration of processed outputs from Big Data, which provided them insights
into the market, channel performance, and customer sentiment across geographies, and helped them
improve efficiencies in procurement and logistics processes.
