Traffic Management w. Rate Limiting to Ensure API Scalability & Resiliency
summary
Autodesk Forge enforces multi-level traffic controls: an ingress API Gateway that applies a global spike arrest 20 rps, per-proxy quotas and hierarchies, and per-consumer limits like quota 3 requests per API key. Rate algorithms include token bucket, fixed window, rolling window. A demo uses API Gateway + Lambda showing 429 + Retry-After behavior; SDKs embed exponential backoff and circuit breakers so throttling degrades service gracefully and preserves platform availability.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 0:38Agenda
- 1:33What is Autodesk
- 2:02Autodesk Ecosystem
- 3:14Revit
- 4:29Architecture
- 5:58Bad actors
- 6:54Solving the problem
- 7:07Throttling
- 10:48Algorithms
- 11:12Token Bucket
- 12:37Fixed Window
- 13:47Rolling Window Log
- 15:17Drivers
- 16:53Limited capacity
- 18:39Monetization
- 19:07Demo
- 19:34PG API Management
- 22:32Resiliency SDK
- 24:45Rate Limiting for Services
related talks
25mArt and Science of Rate Limits for APIs
42mStop Rate Limiting! Capacity Management Done Right
28mAPI Gateway: critical component of your digital architecture
54mFive-step plan for securing your APIs
25mSecuring your APIs at Scale
16mLightning Talk: More Intelligent Global Rate Limiting
50mRate Limiting in Node js applications
21mAPI Security & Performance for Gaming