A developer is designing a modem, distributed microservice enterprise application. The application will be integrating with other systems and focus on a large deployment, so control of API calls is necessary. What is the best practice to reduce application response latency and protect the application from excessive use?