Serverless Experiences -- On Your Own Servers?
summary
Argues serverless means a developer model focused on a single unit of work with the platform auto-scaling and hiding capacity concerns; gives concrete examples — Google BigQuery runs SQL across terabytes to petabytes and returns results in seconds, object storage behaves as an early serverless service, Lambda/Cloud Functions/OpenWhisk handle short stateless invocations, Cloud Run and Apache Beam target other shapes of work. Notes practical limits: Kafka’s partition/broker and coordinator failover issues, nontrivial ops costs (small GCE instance ≈ $15/month), and skepticism that true scale-to-zero eliminates all backend overhead.
AI-written summary · about
chapters (14)
- 0:00Intro
- 1:02Servers
- 4:01Function as a Service
- 8:45Unit of Work
- 13:48What does it take
- 17:47Multitenancy
- 18:54User Journeys
- 20:13Kafka
- 21:12Whats not serverless
- 22:37Not everything is serverless
- 24:30Call to action
- 25:48Questions
- 30:01Its okay to fake it
- 30:38How important is it
related talks
53mServerless is the Abstraction We Deserve
42mNo Server Is Better Than No Server: Building Blocks for Serverless Architectures
51mConfusion In The Land Of The Serverless
28mThe Brave New World of serverless computing
01h 00mServerless - reality or BS - notes from the trenches
28mServerless - the way microservices should be
01h 00mMicroservices and Serverless
31mWhat is "Serverless"?