A high-demand API between internal systems and Oracle NetSuite

Context

La Casa de las Carcasas is a mobile accessories retailer with more than 5,000 employees, more than 1,200 stores worldwide, and warehouses in Europe and Latin America. The company runs an ERP project built around Oracle NetSuite.

Problem

The internal systems and NetSuite have to share every kind of data, at high volume:

• The item master: more than 70,000 references, synced continuously and on demand.
• Sales: 1,200 stores selling every day, which means at least 50,000 records a day that have to arrive with little delay.
• Stock movements: they have to reach NetSuite in the right chronological order, or the stock stops adding up.

The order in which events are processed matters for some kinds of events but not for others.

What I did

• Designed the high-demand API that syncs the internal systems with NetSuite. I am its main developer.
• Built it on Laravel Queues, Redis and Horizon.
• Made FIFO ordering configurable per event type, so each kind of event is processed with the guarantees it needs.
• Coordinated stock movements chronologically, to avoid stock mismatches in NetSuite.
• Contributed to the Snowflake integrations within the same ERP project.

Result

• More than 70,000 item references synced continuously and on demand.
• At least 50,000 sales records a day from 1,200 stores, with little delay.
• Stock movements coordinated chronologically to avoid mismatches.
• One API for every kind of data, with ordering guarantees chosen per event type.

Stack

Laravel, Laravel Queues, Redis, Horizon, Oracle NetSuite, Snowflake.

What I learned

How to work with large volumes of data while keeping both performance and visibility high: a master of more than 70,000 references synced continuously and on demand, at least 50,000 sales records a day with little delay, and stock movements that must land in NetSuite in the right chronological order to avoid mismatches.