I attended the AWS re:Invent 2017 conference this year and wanted to pass along some observations. The intent this year was to provide the opportunity for attendees to reserve seats in advance for the sessions and hold back 20% of the seating capacity for walk ups. Unfortunately, there was an issue with the seat reservation system and I was unable to make any reservations for the first hour that it opened up. This resulted in my being unable to make reservations for most of the sessions that I was interested in, primarily those for Alexa and serverless related topics in the Aria hotel.

The first day was by far the most crowded and I ended up having to get in the walk up lines at least 1.5 hours in advance to ensure that I made it in to the session. The second day improved somewhat but that may have been due to my planning on being in each line very early so I only focused on a few sessions that I particularly wanted to attend. The 3rd through 5th days were quite a bit better and the crowds seemed reduced.

There were 4 venues where sessions were occurring with shuttle buses between them. The bus travel time was very unpredictable so I only used them twice to go to the Venetian hotel.

The focus areas for AWS were voice first, serverless, machine learning and containers.

I am a believer in the Alexa (Echo) service and in serverless architectures and re:Invent was a great opportunity to learn more about both.

The sessions that I made it to follow

  • Alexa and AWS IoT: Now With 100% More Voice
  • serverless authentication and authorization
  • deep dive on aws cloudformation
  • aws batch: easy & efficient batch computing on AWS
  • keynote (and some of the announcements):
    • cloud9 - https://aws.amazon.com/cloud9/
    • business alexa
    • voice first
    • serverless - don’t pay for idle (lambda, containers, database)
    • machine learning
    • transcribe - https://aws.amazon.com/transcribe/
    • comprehend - https://aws.amazon.com/comprehend/
    • deeplens - https://aws.amazon.com/deeplens/
    • kubernetes support
    • bare metal workload
    • aurora severless
    • automatic text translation
    • spot instance pause and resume
  • designing multimodal experiences for alexa
  • alexa workshop - petmatch
  • alexa - conversation and memory
  • container networking deep dive with amazon ecs
  • analyzing streaming data in real time with kinesis video streams
  • orchestrating lambda with step functions
  • serverless application repository
  • step function patterns
  • the art and science of conversation applied to alexa skills
  • building alexa analytics using aws

re:Invent is a very motivating experience and I recommend it but in the future I would definitely expect to have reserved seating available or else to have the sessions livestreamed in real time so that they can be watched from any location.

AWS (Amazon) is providing all sorts of amazing services and products that will lead to many fascinating applications.