Given that the average client restaurant can only hold one day of on hand inventory, all restaurants within the chain are placing orders every day. Typically, a restaurant counts its inventory just one time per month. The client restaurant was relying on a SaaS solution that leveraged sales data to estimate the depletion of on-hand inventory levels, but that system was old and outdated and struggled to handle the client restaurant’s current volume of sales data. Artisan was engaged to build a new perpetual inventory engine service capable of processing over 1 billion transactions annually to power the restaurant’s new modernized supply chain technology stack.
In order to migrate the client off of its legacy platform and scale a system to process a billion+ sales transactions annually, Artisan turned to AWS. A key component of the solution architecture would be the ability to scale rapidly with no downtime during peak demand hours around lunch and dinner time. With that in mind, the team leveraged Elastic Kubernetes Service + DynamoDB for sales transaction processing and pushed aggregated inventory data in 5-minute increments down into Amazon RDS where relational integrity with other required data points for estimating inventory could be easily maintained.
The perpetual inventory engine runs continuously and is highly available. The transaction processing power of the architecture easily handles multiple orders of magnitude; more data than is required to support the client restaurant’s current chain-wide needs. The client intends to double its current revenue over the next 10 years, so this solution will be capable of supporting its needs for the foreseeable future.
