You have a container image: Now what?
summary
Claims there are 34 ways to run containers on AWS and groups services into four practical buckets, then focuses on the two that accept arbitrary container images. It frames hosting as a three-layer model — capacity (EC2, Fargate, Anywhere), orchestration (ECS, EKS, ROSA), and provisioning (App Runner, Copilot, Docker Compose, Amplify, Batch, Beanstalk, Greengrass). Fargate is a serverless capacity used via ECS/EKS. The team builds a prototype decision tree using product scorecards to recommend services based on workload, ops model, portability and hybrid needs.
AI-written summary · about
chapters (12)
- 0:00Introduction and context
- 1:03AWS container strategy
- 5:45Categorizing AWS services
- 12:20Layered architecture approach
- 16:37Visualizing service interactions
- 22:04Strategic vs tactical decisions
- 23:44Understanding AWS Fargate
- 27:31Orchestration and runtime
- 34:00Choosing orchestrators
- 37:15Portability and management
- 40:44Review of provisioners
- 56:29Prototype and closing
related talks
33mRunning Containers on AWS | Running Containers on AWS
33mContainers and AWS Lambda: Choose the optimal compute for applications
43mBringing EDA to the Containers World
38mService decisions for containerized applications
44mServerless and containers are coming together: an AWS perspective
35mContainers and AWS Lambda: Choosing optimal compute for modern apps
35mKubernetes on AWS: Zero to hero
18mAbby Fuller, AWS | DockerCon 2018