Albert Heijn's smart replenishment team is responsible for distributing forecasts to different systems and providing a BI solution where the business can analyze the forecasts. Managing real-time aggregation of approx. 7 billion forecasts within a store forecasting system using Azure SQL Database presents significant data challenges. Performance bottlenecks often arise due to the sheer data volume, impacting the system's ability to process information swiftly. Additionally, the costs associated with transferring and processing such vast amounts of data can be prohibitively high. Addressing these issues requires strategic solutions to optimize performance and reduce expenses.
In this talk we present our approach to try resolving these challenges using Serverless DBSQL.