Enhanced forecasting reduces food waste

BUSINESS OBJECTIVE

A high-end supermarket chain wished to deploy a business intelligence environment, to enable enhanced stock forecasting and other crucial, data-powered business processes. Enhanced forecasting reduces food waste Unlocking compute capacity for complex business analytics

RETAIL

SCENARIO

• Faced with increased pressure to cut food waste, the retailer was looking to drive compute-intensive analytics processes such as stock forecasting, in near real-time, that would increase replenishment accuracy.
• The retailer planned to introduce Hadoop to its existing relational database infrastructure, and enable next-generation SAS analytics tools, using multiple datasets, including customer-facing and supply chain databases.
• The company operated two data centers, 20km distant, for failover and backup, and could not afford operational disruption.

CHALLENGES

• Hadoop replication solutions were overly complex and expensive, requiring too much admin effort.
• Mission-critical replenishment forecasting is a continuous process, and many solutions offered two-hour RPO and eight-hour RTO – too slow for the retail trade.
• With a mixed storage, server, and application environment, the retailer was faced with the cost and complexity of deploying multiple solutions.

SOLUTION

Deploy WANdisco LiveData Platform to replicate data automatically to its second site for disaster recovery, offering high availability with near-zero RPO and RTO. Run specific compute-intensive operations on the idle capacity at the second data center, and use the WANdisco LiveData Platform to replicate the results to the primary data center.

ENVIRONMENT, TECHNOLOGY, AND TOOLS

• Teradata Hadoop Appliances
• Relational databases
• Ab Initio ETL
• WANdisco LiveData Platform

CAPABILITIES INTRODUCED

WANdisco LiveData Platform provides real-time data replication to enable separate data centers to function as one, with near-zero RPO and RTO disaster recovery.

BENEFITS & OUTCOMES

• Automates replication of 100% of Hadoop and relational database data, without need for separate solutions
• Creates a LiveData environment with consistent data at both data centers
• Large savings compared with running multiple solutions
• Protects against data and value loss in case of disaster