Serverless Compute at the Heart of Your EDA • Julian Wood • GOTO 2024

Learn how to leverage AWS Lambda functions and containers in event-driven architectures, with insights on choosing the right compute option for your workloads.

Key takeaways
  • Serverless compute options include AWS Lambda for functions and ECS/Fargate for containers, each optimized for different use cases

  • Lambda functions can scale instantly up to 10,000 concurrent executions within 90 seconds, with no infrastructure management needed

  • Lambda is best for event-driven workloads with unpredictable traffic patterns, while Fargate containers work better for stable, long-running workloads

  • Key Lambda constraints: 15 minute maximum runtime, up to 10GB memory, pay-per-millisecond pricing model

  • Fargate provides more control and flexibility for container workloads but requires more configuration and doesn’t scale as rapidly as Lambda

  • Event-driven architectures can be built using both Lambda functions and containers - they complement each other rather than compete

  • Lambda functions can be packaged as container images for consistency with container workflows while maintaining serverless benefits

  • AWS EventBridge enables serverless event routing between AWS services and external SaaS providers

  • Step Functions allow visual orchestration of workflows combining Lambda functions and container tasks

  • Moving to serverless reduces operational overhead but requires learning new architectural patterns and best practices