Distributed Rate Limiter Design | Dataford Interview Questions

Design a distributed rate-limiter for an API gateway that processes incoming data feeds from various external exchanges for Jane Street.

Implement process_requests(requests, limits, node_count). Each request is a dictionary with timestamp, exchange, and cost; timestamps are nondecreasing. limits maps each exchange to [capacity, refill_rate], where refill rate is tokens per second. Route each exchange to one of node_count logical nodes and apply a token bucket independently per exchange. Return a Boolean list indicating whether each request is accepted. State updates must be atomic and request order must be preserved.

Constraints

1 <= len(requests) <= 1000
1 <= node_count <= 100
Requests are ordered by nondecreasing timestamp
Each timestamp is a nonnegative number
Each request cost is positive
Each exchange in requests appears in limits
Each limit is [capacity, refill_rate] with positive values

Function Signature

def process_requests(requests, limits, node_count):

Your question is Distributed Rate Limiter Design. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Design a distributed rate-limiter for an API gateway that processes incoming data feeds from various external exchanges for Jane Street.

Constraints

1 <= len(requests) <= 1000
1 <= node_count <= 100
Requests are ordered by nondecreasing timestamp
Each timestamp is a nonnegative number
Each request cost is positive
Each exchange in requests appears in limits
Each limit is [capacity, refill_rate] with positive values

Function Signature

def process_requests(requests, limits, node_count):

Your question is Distributed Rate Limiter Design. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.