AGL in the Cloud: Progress on AMI Builds
summary
AGL boots as an AWS AMI built from the meta-aws layer and uploaded via the create-e2 script (requires AWS CLI, IAM roles, and an S3 bucket). Images exist for x86 and ARM64 (Graviton t4g); a public x86 AMI is available in us‑west‑2 and runs acceptably on a t3.xlarge. Remote graphics use an RDP backend with TLS certs (VNC planned) and x86 rendering uses llvmpipe; SSH login is to user not root.
AI-written summary · about
chapters (10)
- 0:00AGL in the cloud intro
- 1:27Cloud images simplify dev
- 4:01Cloud machines and builds
- 6:37AGL images for AWS EC2
- 7:43Upload local builds to AWS
- 9:13Launch and access cloud instances
- 11:46Graphical access over RDP
- 13:11Enable RDP and performance
- 15:22Live demo in AWS
- 18:24Next steps and future
related talks
32mImage mode for AI inference across the hybrid cloud
01h 14mModernizing and Optimizing Applications for the Cloud
20mAWS On Air ft. AWS Graviton3 Processors and Amazon EC2 C7g Instances
32mAgile governance on AWS GovCloud (US)
25mAdopt AWS Graviton2 instances for best price-performance in EC2
52mMagma Cloud Deployment on AWS
31mEKS Cost Optimization with EC2 Spot and A1 Instances
30mMagma on Arm – porting and best practices