Skip to main content

AWS Re:invent Day One

ยท 3 min read
Jeremy Care

Quick feedback from my first day at Re:invent 2021. My first sessions were about Multi-tenant-aware SaaS Microservices, and Automated content moderation using Media2Cloud.



AWS Certification Loungeโ€‹

I started the day with a visit on the AWS Certification Lounge in the Venetian Hotel, you need to have a certification to be able to enter. In this place you have streaming of all the keynotes, coffee, snacks etc. It's a pretty chill atmosphere. You can fill a survey to win an AWS CERTIFIED Badge, and a water-flask!

Water Flask

(Yes, I'm flexing with my AWS Amplify Video shiny sticker.)

GPS311 Building Multi-tenant-aware SaaS microservices on AWS (Chalk Talk)โ€‹

I could not reserve my seat on the application, it was already full. I arrived early to be the first one in the Walk-up queue. One hour before, it was too early, no one was there, but around 20 mins before the session the line was huge. When all the registered attendees entered the room, they filled the empty chairs with people from the walk-up queue. I could go in.

This session was hosted by Micheal Beardsley, member of the AWS SaaS Factory team. the subject of the session was multi-tenant-aware microservices. He showed us an evolution of architecture, from a multi users to a multi-tenant pattern.

The main idea was to inject the tenant context in the microservice:

  • Reduce compute permissions to only tenant-specific resources using policy, leveraging Cognito and IAM. (For example, only allows access to specific DynamoDB Partition based on the tenantID)
  • Enriched logging, add more context than just a log, as a JSON format to be able to ingest it in another processing pipeline. It's really important since SaaS are often metrics driven software. Where you add feature based on your user usage.



AIM307-R2 Automating content moderation to improve digital safety (Chalk Talk)โ€‹

This time I was able to reserve my seat, it was more comfortable to attend.

This session was about AWS Media2Cloud open source project focused on the content moderation use cases.

It's a solution we used for multiple projects. It creates a pipeline of to analyze media, providing integrations for all the AI/ML services. You can use this solution to build the base of your Curation Video content for example.

They gave examples of video analyzes and explicit content flagging. I recommend watching the replay if you want to build something similar.

It was a good showcase of the new version of Media2Cloud the v3. With the addition of OpenSearch, Rekognition custom labels etc.



We went for a diner with Amplify-Video team, to meet with them after working on the project for a year now, it was really fun!