The following example describes an ecommerce platform built by an auto parts manufacturer using Amazon SNS FIFO topics and Amazon SQS queues. The platform is composed of four serverless applications:
-
Inventory managers use a price management application to set the price for each item in stock. At this company, product prices can change based on currency exchange fluctuation, market demand, and shifts in sales strategy. The price management application uses an AWS Lambda function that publishes price updates to an Amazon SNS FIFO topic whenever prices change.
-
A wholesale application provides the backend for a website where auto body shops and car manufacturers can buy the company's auto parts in bulk. To get price change notifications, the wholesale application subscribes its Amazon SQS FIFO queue to the price management application's Amazon SNS FIFO topic.
-
A retail application provides the backend for another website where car owners and car tuning enthusiasts can purchase individual auto parts for their vehicles. To get price change notifications, the retail application also subscribes its Amazon SQS FIFO queue to the price management application's Amazon SNS FIFO topic.
-
An analytics application that aggregates price updates and stores them into an Amazon S3 bucket, enabling Amazon Athena to query the bucket for business intelligence (BI) purposes. To get price change notifications, the analytics application subscribes its Amazon SQS standard queue to the price management application's Amazon SNS FIFO topic. Unlike the other applications, the analytics one doesn't require the price updates to be strictly ordered.
For the wholesale and retail applications to receive price updates in the correct order, the price management application must use a strictly ordered message distribution system. Using Amazon SNS FIFO topics and Amazon SQS FIFO queues enables the processing of messages in order and with no duplication. For more information, see Amazon SNS message ordering details for FIFO topics. For code snippets that implement this use case, see Amazon SNS code examples for FIFO topics.
