Architecture
A possible agnostic architecture, applying what you have learned in previous chapters, looks like Figure 5.1. Different applications connect to a frontend component, which handles connection termination, SSL certificates, and so on, and orchestrates and load balances backend connections where the submitted photos are parsed and analyzed.

Figure 5.1 – Photo classification architecture
A diagram using AWS services looks like the one in Figure 5.2: a three-component diagram, with API Gateway, Lambda, and Rekognition. Different types of applications will connect to API Gateway using HTTPS, invoking a Lambda function, which queries Rekognition for image analysis and parses the response accordingly.

Figure 5.2 – Photo classification architecture on AWS
In the following section, you will find a detailed explanation of why Rekognition is the better tool for this use case. For now, trust us.
In this architecture...