AWS - Non EC2 Compute

by Rahul Gupta

ECS : Elastic Container Service

  • Launch Docker containers on AWS
  • You must provision and maintain the infrastructure (the EC2 instances)
  • AWS takes care of starting / stopping containers
  • Has integrations with the Application Load Balancer

Fargate

  • Launch Docker containers on AWS
  • No need to provision the infrastructure (no EC2 instances to manage)
  • Serverless offering
  • AWS runs containers for you based on the CPU/RAM you need

ECR -  Elastic Container Registry

  • Private Docker Registry on AWS
  • This is where you store your Docker images so they can be run by ECS or Fargate

Serverless - Lambda

  • AWS EC2 vs Lambda
  • Easy Pricing:

— Pay per request and compute time

— Free tier of 1 million AWS lambda requests then $0.20 per 1 million thereafter and 400K GBs of compute time means 200 seconds if function is 1GB RAM means 3.2 million seconds, if function is 128MB RAM after that $1 for 600K GB-second

  • Integrated with the whole AWS suite of services
  • Event Driver: functions get invoked by AWS when needed
  • Integrated with many programming languages - Node.js, Python, Java, C# (.net core), Golang, C#/Powershell, Ruby, Custome Runtime API (example Rust)
  • Easy monitoring through AWS CloudWatch
  • Easy to get more resources per functions (Up to 3GB of RAM)
  • Increasing the RAM will also improve CPU and network!

Example 1: Serverless Thumbnail creation

Example 2: Serverless CRON Job

AWS Batch

  • Fully managed batch processing at any scale
  • Efficiently run 100 thousands of computing batch jobs on AWS
  • A batch job is a job with start and an end (opposed to continuous)
  • batch will dynamically launch EC2 instances or spot instances
  • AWS batch provisions the right amount of compute/memory
  • You submit or schedule batch jobs and AWS Batch does the rest
  • Batch jobs are defined as Docker images and run on ECS
  • Helpful for cost optimizations and focusing less on infrastructure

Example - Image Processing

Lambda vs Batch

  • Lambda has time limit, limited runtime, limited temporary disk space and its serverless
  • Batch has no time limit, any runtime as log as its packaged as a Docker image. Rely on EBS/instance store for disk space. Batch is not serverless, but its managed service relies on EC2 instances

Lightsail

  • Virtual servers, storage, DB and networking at one place
  • Low and predictable pricing
  • Simpler alternative to using EC2, RDS, ELB, EBS, Route53 etc
  • Great for people with little cloud experience
  • Can setup notifications and monitoring of your Lightsail resources
  • Use cases:

— simple web application (has templates for LAMP, Nginx, MEAN, Node.js etc)

— Website (templates for Wordpress, Magento, Ples, Joomla)

— Dev/Test environment

  • Has HA but no auto scaling, limited AWS integration