Modernizing and Optimizing Applications for the Cloud
summary
Workshop shows a hands‑on path: port a dummy LLM chatbot (with a C++ shared library) to ARM64, recompile binaries, build multi‑arch container images and manifest lists, and deploy via EKS or EC2 autoscaling while measuring requests/sec and latency. Tools used: EC2, RDS, Lambda, EKS + Carpenter, web VS Code, Amazon Q and CI/CD pipelines. Key findings: Graviton (ARM64) often delivers ~40% better price‑performance, cross‑compilation on emulation is very slow, and multi‑arch images + phased rollouts work and reveal real production gains.
AI-written summary · about
chapters (11)
- 0:00Workshop introduction
- 1:09Cloud application optimization focus
- 6:05EC2 instance types and CPU choices
- 17:41Graviton architecture deep dive
- 28:12vCPU and core mapping explained
- 30:23Workshop lab setup overview
- 47:47Second workshop segment begins
- 48:53Performance benchmarking strategies
- 57:07Application deployment options
- 1:07:57Deploying multi-platform artifacts
- 1:11:41Workshop recap and closing
related talks
36mWhy and How to Choose Arm64 for your Cloud Applications
31mContinuous Delivery on Arm: A Practical Approach
31mOptimizing Cost and Performance with arm64
33mOvercoming migration hurdles
44mWhat to know before adopting Arm: Lessons learned at Datadog
27mWhat's Going ARM: Adopting ARM64 At Airbnb
34mARM Migration Made Practical
06mChange Your Architecture and Save the World