Serverless Boom or Bust? An Analysis of Economic Incentives
summary
An economic model compares provider utilization and price ratios against customer benefits and defines an alpha classification of customers: alpha <1 (need better serverless features), 1–c^{-1} (want consumption pricing but unprofitable for providers), and >c^{-1} (providers can profit). Simulation shows breakeven at ~50%/25% of peak for Poisson and bursty arrivals respectively and that oracle one‑minute autoscaling barely shifts those points. Deliverable insight: alpha ≈ 10^2 to hit p95<1s, implying modest QoS guarantees could unlock substantial serverless value and that hybrid provision/consumption often wins.
AI-written summary · about
chapters (5)
- 0:00Intro
- 1:36Questions we want to reason about
- 3:20Developing the model
- 7:29Examining Autoscaling
- 12:30Conclusions
related talks
21mBeyond the Basics: Production Serverless Patterns for Extreme Scale
17mCloud Computing Made Better: An AWS View on Serverless Computing
18mMaking Serverless Pay-For-Use a Reality with Leopard
27mAzure Cosmos DB Serverless Unleashed
40mThe State of Serverless Computing
46mServerless: Making Sense with a Cloud Paradox
43mDissecting serverless runtimes - from functions to databases
04mWhat if serverless was real?