Evolving Serverless Architectures
summary
Shows how to replace bespoke Lambda glue with managed services: use EventBridge Pipes and DynamoDB streams to remove event-publish code, integrate Step Functions directly with Bedrock via SDK, and use wait-for-callback task token to pause workflows for human review. Concrete results: DVLA photo pipeline handles 1.5M apps/year, moves Lambda logic into infra, cuts function size, adds ~100ms latency, and cites rough costs like ~$0.40 per million for pipes and ~$0.20 per million Lambda savings, simplifying ops and improving resilience.
AI-written summary · about
chapters (10)
- 0:00Intro
- 3:00Agenda
- 3:48Why evolve serverless architectures
- 5:01Pattern catalog
- 12:30Operational considerations
- 16:32Integration into your development flow
- 20:01Real world examples from DVLA
- 23:05Examples of evolving architecture patterns
- 28:30Summary
- 29:39Outro
related talks
23mSimplifying design complexities using Serverless Architecture on AWS
40mRefactoring to serverless
38mRefactoring to serverless
34mDesigning Data Intensive Applications In Serverless Architecture
42mBest practices for serverless developers
43mBest Practices for Serverless Developers
49mMeasuring and increasing developer productivity with the help of Serverless
23mFrom 1 to 101 Lambda Functions in Production: Evolving a Serverless Architecture–Slobodan Stojanovic
