8 Keys for Successful Serverless Architectures
summary
Argues that serverless keeps value for security and cost-efficiency but fails because functions are stateless and durable data stores don’t autoscale, causing connection storms and billing mismatch. Proposes moving to event sourcing and CRDTs (conflict‑free replicated data types) plus workflow or actor-style runtimes so state is injected and replayable. Names concrete runtimes: Azure Durable Functions (durable task framework) and Lightbend CloudState as working examples of stateful serverless.
AI-written summary · about
chapters (20)
- 0:00Intro
- 0:50What is Serverless Today?
- 1:24Current Value of Serverless
- 3:26Berkeley View on Serverless Computing
- 4:57Challenges in True Serverless
- 7:03Common Limitation of Serverless?
- 9:34Abstract over State
- 11:01Traditional CRUD Database Access Won't Work
- 12:50Better Models for Distributed State
- 14:59CRDT's Provide a Distinct View of Data
- 15:31CRDT Convergent Operations
- 16:16Value of CRDTs in Distributed Systems
- 17:02Event Sourcing in Stateful Serverless
- 18:51Depends on Shape of Data and Level of Consistency
- 19:53Primary Models for Stateful Service Architecture
- 21:05Azure Durable Functions
- 21:57Lightbend Cloud State
- 22:43Serverless Event Sourcing
- 23:41Serverless CRDTs
- 24:29Cloud State - Low-Latency & High Throughput
related talks
01h 00mMicroservices and Serverless
51mConfusion In The Land Of The Serverless
55mConfusion In The Land Of The Serverless
49mYou Don't Want Serverless
52mConfusion In The Land Of The Serverless
14mThinking Twice About Migrating to Serverless
04mWhat if serverless was real?
40mThe State of Serverless Computing