How serverless impacts design -
summary
Serverless shifts design from monolithic apps to unbundling from apps to tasks, forcing deployment to shape the model: design event payloads as purpose-built data transfers, make protocols the primary boundary, and plan for version-tolerant upgrades. Real-world numbers: migration from Heroku to AWS Lambda cuts hosting ~50% (IDC cites ~60% average ops savings) and providers bill by invocation and time/memory — pay per request (~$0.20 per million). Examples include Cognito hooks, Atlassian Forge, Twilio functions, and Comic Relief cutting a $83k peak to $5k (≈94% savings).
AI-written summary · about
chapters (9)
- 0:00Introduction and buzzword rant
- 6:49Terminology clarification
- 13:39Serverless as integration pattern
- 21:44Deployment constraints and design
- 28:03Task-based design over applications
- 34:03Distributed system design challenges
- 41:49Protocol design in serverless
- 48:06Security boundaries in tasks
- 51:35Conclusion and future implications
related talks
20mWhy Serverless Architecture is Changing Software Development
58mServerless: Five Key Things you need to Know
23mSimplifying design complexities using Serverless Architecture on AWS
36mTransforming SaaS with Serverless Architectures
05m#Serverless + #Devops = BFF
01h 04mMigrating to Serverless - an experience report
01h 00mServerless - reality or BS - notes from the trenches
35mServerless by Design