Containers or Serverless Functions - When to choose what?
summary
Compares concrete tradeoffs: containers (Kubernetes/EKS, ECS, Fargate, App Runner) give hardware control, GPUs and long-running processes and often lower cost for steady load, while functions (AWS Lambda, API Gateway, SQS, EventBridge) scale to zero, use Firecracker microVMs, and charge only per execution but suffer cold starts and a 15 minute max execution limit. Shows real results: a renewable energy team runs 20 million predictions/year with 90% less ops, Kora runs MLops on EKS, and Lexoffice mixes containers plus Lambda for spiky invoice lookups.
AI-written summary · about
chapters (8)
- 0:00Introduction and overview
- 1:07Container fundamentals explained
- 7:08Serverless functions basics
- 18:03Cost comparison analysis
- 23:20Flexibility and portability
- 33:29Serverless containers options
- 35:13Real-world use cases
- 40:04Key takeaways summary
related talks
33mContainers and AWS Lambda: Choose the optimal compute for applications
41mIs Serverless the End of Kubernetes?
48mServerless beyond functions
24mContainers vs Serverless from a DevOps standpoint
57mServerless or Containers? Who Cares Pick Both!
35mContainers and AWS Lambda: Choosing optimal compute for modern apps
51mServerless Compute at the Heart of Your EDA
44mServerless and containers are coming together: an AWS perspective