Design a Quota / Rate-Limiting Service

Problem

Design a distributed quota/rate-limiting service for a data platform (Snowflake-style design, not a generic web app).

Requirements

Functional:

  • Enforce per-account compute/query quotas
  • Distributed sliding-window rate limiting
  • Fair sharing across clients
  • Real-time quota reads

Non-functional:

  • Millions of requests/sec
  • Accurate under bursts
  • Low added latency

Discussion points

  1. Sliding-window vs token-bucket at scale
  2. Distributed counters (Redis) and consistency
  3. Fairness and burst handling
  4. Fault tolerance and clock skew
  5. Hot-account mitigation
added …
LeaderboardSalaryAccount