Swift Lambdas at Scale: How a Tiny iOS Team Built a Big Backend
summary
Small iOS team replaces Node/Docker servers with serverside Swift running on AWS Lambda, S3 and DynamoDB, orchestrated by Step Functions and fronted by API Gateway; infrastructure is defined in Terraform and uses the Sodo Swift AWS wrapper. Shared Swift packages eliminate duplicated models, CI/CD deploys duplicated prod and sandbox regions, peak load hits ~500 concurrent Lambda executions, Step Functions bypass 15‑minute limits, and monthly bills consistently come in lower than prelaunch estimates.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:46Origins of Song Shift
- 2:16Solving music transfers
- 3:47Early backend architecture
- 5:13Moving to AWS and Swift
- 7:47Getting started with Lambda
- 10:47Optimizing for the team
- 12:07Swift language benefits
- 13:53Overcoming statelessness
- 15:24Advanced cloud integration
- 16:53Scaling with Step Functions
- 19:02Backend architectural design
- 21:16Monitoring and observability
- 22:55Infrastructure as code
- 25:26Testing and production
- 26:05Conclusion and takeaways
related talks
33mAWS Lambda - Logic in the Cloud without Back-end
35mSwift, Server-Side, Serverless
31mAccelerate business innovation with serverless applications
29mGoing Serverless
30mEvolving Serverless Architectures
42mServerless Tales from the Trenches
54mServerless Architectures: no servers allowed!
01h 04mMigrating to Serverless - an experience report