How Public APIs Manage the Loads?

Question

We have a Web API Service with 10 different web methods which return some data whose size could be like 15MB.

The clients may call 1000 calls per minute to the service which may take down the service!

What's the best practice in managing the client calls load for public APIs?

How does Twitter API manage/limit it?

I could limit the calls per client key and/or limit the calls per POST or GET Calls per Client?

Answer 1

You're on the right track with limiting access per client key e.g., 50 requests per-key per-minute.

Twitter rates per-access token in your control i.e., per-user authenticated with the app in question. You can read all about Twitter's rate limiting approach at the link below:

https://dev.twitter.com/docs/rate-limiting/1.1