AWS re:Invent 2020 Virtual Conference
- Date: December 1-18, 2020
- Location: Virtual (100% online)
- Attendance: 500,000+ registered attendees (up from 60,000 typical in-person)
- Sessions: 500+ sessions, 800+ videos on YouTube
- Andy Jassy CEO Keynote - AWS CEO sharing insights on customers, products, and services
- Machine Learning Keynote - Swami Sivasubramanian (first-ever ML keynote at re:Invent)
- Infrastructure Keynote - Peter DeSantis, SVP Global Infrastructure
- AWS Partner Keynote - Doug Yeum, Sandy Carter, Dave McCann
- Werner Vogels Keynote - CTO perspective on building the future
- Container Image Support - Package and deploy Lambda functions as container images up to 10 GB
- Billing Granularity - Reduced from 100ms to 1ms, automatic cost savings
- Memory Expansion - Functions with up to 10 GB memory and 6 vCPUs
- SageMaker Pipelines - First ML CI/CD service, reduces months of coding to hours
- SageMaker Data Wrangler - Simplified data preparation from single interface
- SageMaker Feature Store - Consistent features for training and inference
- SageMaker Clarify - Detect bias in ML models, increase transparency
- SageMaker JumpStart - One-click access to popular models and solutions
- SageMaker Edge Manager - Manage ML models across edge devices
- Amazon EKS Distro - Kubernetes distribution for self-managed clusters
- Amazon ECS Anywhere - Run ECS on-premises or other clouds
- Amazon EKS Anywhere - Run EKS on-premises or other clouds
- Amazon ECR Public - Public container registry
- AWS Proton - Automated management for container and serverless deployments
- AWS Trainium - Custom ML training chip designed by AWS
- Habana Gaudi EC2 Instances - ML training instances
- Amazon Monitron - End-to-end equipment monitoring
- AWS Panorama - Computer vision at the edge
- Machine Learning Democratization - Making ML accessible to all developers
- Hybrid Cloud - EKS/ECS Anywhere bringing AWS to any environment
- Container Innovation - Lambda container support bridging serverless and containers
- Custom Silicon - AWS-designed chips for specialized workloads
- Edge Computing - Bringing compute closer to data sources