Infrastructure From Code: Next up for DevOps?
summary
Argues that a new layer—called infrastructure from code or provisioning runtime—infers app needs from business logic and provisions AWS resources (API Gateway, Lambda, DynamoDB, S3) via SDKs, annotations or new languages (Model, Encore, Shuttle, Wing) rather than YAML/CloudFormation. Shows a provider that issues customer AWS accounts and pre-provisions sandboxes, one user already handling 60 million requests/day, and demos LLM-assisted generation of deployable apps. Claims cost-driven actions like automatic switching to Fargate, hopes for self-provisioning and self-optimization but says full portability and no-ops aren’t solved yet.
AI-written summary · about
chapters (10)
- 0:00Introduction and motivation
- 3:08Speaker introduction
- 4:12Evolution of software shipping
- 8:04Serverless confusion and challenges
- 10:34Developer productivity issues
- 14:28What is infrastructure from code
- 17:18State of infrastructure from code
- 20:31Vision and self-managing systems
- 23:18Less Ops and future automation
- 25:46Q&A session
related talks
06mInfrastructure From Code: Next up for DevOps?
59mInfrastructure as Actual Code
42mInfrastructure as Code - Lessons learned from Dev to Ops
44mInfrastructure from Code: The Next Generation of Cloud Management
44mInfrastructure as Software
12mThe State of Infrastructure-From-Code 2023
28mContinuous everything: Moving to DevOps on AWS
46mImplementing Infrastructure as Code