Oracle long running – job offload to Hadoop

Client
Sporting goods retailer
Our client is a full-line sporting goods retailer offering a broad assortment of brand name sporting goods equipment, apparel, and footwear in a specialty store environment.
Challenges
Project scope
The client wanted to offload long running Oracle jobs to Hadoop. In Oracle, jobs were running for over 6 hours, the target was to run in less than 1 hour.
Approach
Sourcing data & eliminating long-running tasks
Datametica helped the client in the following ways:
- Sourced data from multiple data stores to Hadoop (AS400, Oracle and other files)
- Eliminate the additional long-running ETL jobs
Solution
Efficient time management
- Now the job runs in Hadoop in 30 minutes on a small Hadoop cluster
- Provided data feed to downstream systems like SAS
Benefits

Time management
Business user has the data 8 hours earlier than the legacy ETL/Oracle solution

Business cycle planning
It helps them to plan the business cycle an entire day earlier and be ready to deploy their plans at the start of the next business day

Cost efficiency
Cost effective solution compared to Oracle

Analytics platform
Offload to Hadoop made available the data on an analytics platform, enabling them to do analytics on the same data