Rate Limiting

[SeanMcOwen]

Figure out where this goes:

Rate Limiting

Rate limiting limits the amount of work (i.e. Web3 access) a Servicer can provide to an Application throughout the duration of a Session.

During each Session, the amount of POKT an Application has staked (see Application Protocol for more details) is mapped to "Service Tokens" that represent the amount of work a Servicer can provide using the SessionTokenBucketCoefficient governance parameter. The Token Bucket rate limiting algorithm is used to determine the maximum number of requests a Servicer can relay, and be rewarded for, thereby disincentivizing it to process relays for the Application once the cap is reached.

At the beginning of the session, each Servicer initializes: AppSessionTokens = (AppStakeAmount * SessionTokenBucketCoefficient) / NumServicersPerSession. When one of the Servicers in the session is out of session tokens, the Application can continue to use other Servicers until every they are all exhausted.

The mechanism described above enables future iterations of the protocol where different types of request may vary the required number of AppSessionTokens per request.

[SeanMcOwen]

Make a submit_relay_request_policy for figuring out service tokens and what needs to be computed

[SeanMcOwen]

Done