JohnAbeJames.com

AWS Summit June 2020

  • 18/06/2020
  • Yesterday (17th June) I attended the AWS Summit Online for Europe, Middle East & Africa. The event was bookended by talks from CTO Dr. Werner Vogels and CEO Andy Jassy. I was a little disappointed that there were no major announcements, but the talks focussed on general positive outcomes for AWS clients.
    I chose the ‘I build applications’ track and attended the 5 talks:
    Purpose-built databases for modern applications (Level 300) This was good revision for the Solutions Architect Certification I’m working on; nothing new though. The talk briefly mentions the different database options:

  • Amazon Aurora – SQL
  • Amazon DynamoDB – Key/Value
  • Amazon DocumentDB – Mongo Compatible
  • Amazon ElastiCache – Redis or Memchached (in-memory)
  • Amazon Neptune – Graph Database
  • Amazon Timestream – Time Series, good for sensor feeds etc.
  • Amazon Quantum Ledger Database – a ledger
  • Amazon Managed Apache Cassandra Service – wide column
  • Full-stack mobile and web development with AWS Amplify (Level 300) An introduction to Amplify. I’ve not used Amplify but the talk inspired me to follow up with a basic deployment. Lots of stuff taken care of for you out of the box. I’m not sure how I feel about this, I like to know a little about how everything works in case it goes wrong. Still, looks amazing for prototyping.
    Application integration patterns for microservices (Level 300) The merits of SNS, SQS, and Step Functions.
    Event-driven architecture (Level 300) The merits of API Gateway and Event Bridge.
    A path to event sourcing with Amazon MSK (Level 200) MSK: Amazon Managed Streaming for Apache Kafka Event sourcing - Persist entities by storing a sequence of state-changing events. I need to read more on this because it has some relevance for some client work.
    Some points on AWS -
    Netflix, Disney+, and Hulu all use AWS. AWS profit for Q2 this year was $4bn. Some sage advice -
    Work with lots of experiments with '2-way doors' (decisions you can revert). Think carefully and minimise '1-way doors' (decisions which are difficult to revert). Speed disproportionately matters.