aws elastic container service intro

Containers are similar to a virtual machine, more like a virtual operating environment.

• Standardized - These are standardized units with everything the software needs to run (libraries,

system tools, code, etc.).

• Microservices - applications are created using independent, stateless components

or microservices running in containers * These are typically stateless and can be started/stopped/scaled independently * AWS has a whitepaper called "Running Microservices on AWS"

• Docker or windows containers - use Docker for any non-windows loads, Windows containers

for any Windows loads

Container architecture:

Advantages of containers & microservices architectures

• Highly scalable - if the application becomes over loaded, scale only the services you need to
• Fault tolerant - a single error in 1 of your containers won't bring down the whole app
• Easy to maintain - easier to maintain, update, and change than large mono-apps

What about ECS?

Container orchestration service that supports Docker and Windows containers.

This is similar to Kubernetes at a high level, but iwth deep integration with AWS services like IAM, VPC, Route53.

Quickly deploy and scale containerized workloads without having to install, configure, and manage/scale your own orchestration platform.

• Clusters of Virtual machines - ECS will run your containers on clusters of virtual machines
  • EC2 instances have a lot more control
• Fargate for serverless - use Fargate for serverless containers to not worry about

the underlying EC2 instances

Elastic Container Registry

• Registry of container images for ECS to connect to and deploy Docker containers

Services using ECS

• Amazon Sagemaker - quickly deploy and scale ML models for training
• Amazon Lex - build conversational interfaces
• Amazon's own recommendation engine is runnong on ECS