Urban Outfitters Open Source Refresh

A retailer chooses SymmetricDS to synchronize Oracle POS MySQL store database with central office DB2 database.

The Organization


Philadelphia, PAIndustry
RetailNumber of Nodes

Urban Outfitters chose to refresh their Oracle POS (ORPOS) based store infrastructure, moving from a mostly IBM based store stack to an open source based store stack.  SymmetricDS was chosen to replace an MQ Series solution as the data delivery mechanism to anchor the new store infrastructure.

The Challenge

The challenge was to deliver to the retailer a proven data movement solution that could deliver the massive number of master item and pricing data to multiple brands in a timely manner, and also trickle sales back to the central office throughout the day.  The solution also needed to decrypt and re-encrypt sensitive data as it hit the safe harbor of the central office as well as map several simple data elements to new values for back-end systems.

Another requirement also existed to deliver a subset of the central office sale and item data to another centralized MySQL database which was a data source for a custom web application used by the POS.

The Solution

Having implemented synchronization for ORPOS in the past, JumpMind quickly setup a standard synchronization profile between the MySQL store server and the DB2 central office database.  SymmetricDS was used to capture and trickle back sales and electronic journal data.  It was also used to move item, price, tax and hierarchy data to the store database.  The master item table was routed using the lookup table routing feature of SymmetricDS.  The routing leveraged a table that mapped store to brand.  Re-encryption of sensitive data was accomplished using a bean shell script column data filter at the central office SymmetricDS instance.   Simple data mappings were also accomplished using the same mechanism.

Not only was SymmetricDS used move data between the store and the central office, but it was also used to move data to the second centralized datasource.  In order to do this, a third node group was created.  As data trickled into the central office, the required data for the secondary data source was captured and routed to the additional node group.

The Results

SymmetricDS allowed the retailer to have a resilient and consistent mechanism for moving business critical data both to and from their stores.  SymmetricDS provided an affordable alternative to better known, but more expensive messaging alternatives which required more software to execute at the store itself.