Is Serverless the End of Kubernetes?
summary
Compares serverless and Kubernetes as complementary choices driven by use case, and examines concrete tradeoffs and tools including AWS Lambda, ECS Fargate, EKS/ROSA and managed services like DynamoDB, SageMaker and IoT Core. Presents a Formula One pipeline that ingests ~1.1M datapoints/sec from 300 sensors, routes data via API Gateway into Lambda for inference and DynamoDB for storage (claiming millions of messages/sec), and recommends using serverless for short event-driven APIs while keeping containers for long-running or GPU workloads, with cautions about vendor lock-in, community support and staffing.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 3:54Container basics and Kubernetes
- 7:29What is Serverless
- 12:07Cloud Native 2023
- 16:20Formula 1 case: Serverless
- 20:48Serverless vs Kubernetes
- 33:02Final decision and pitfalls
- 36:14Q&A
related talks
41mContainers or Serverless Functions - When to choose what?
42mNo Server Is Better Than No Server: Building Blocks for Serverless Architectures
31mThe Evolution of Serverless: Past, Present and Future
31mAccelerate business innovation with serverless applications
40mWho Needs Serverless?
32mThe Serverless Tidal Wave
48mThe future will be Serverless
36mTransforming SaaS with Serverless Architectures