Energy Proportional Edge Computing Infrastructure

on YouTube ↗47m

" summary

Prototype shows ~30% energy reduction by shifting video-analytics workloads from host CPU to a network IPU and turning off CPU slices; a model for 42,000 edge sites projects about $70M/year saved and ~140M kg CO2 avoided. Uses Energy Proportional Edge Computing Infrastructure with intent-driven orchestration that applies IPU offload and SST power management, plus time- and space-shifting across off-grid solar sites and federated observability to predict when to move or scale workloads.

AI-written summary · about

" chapters (17)

1. 0:00Energy management at scale
2. 1:05Edge computing deployment models
3. 3:37Sustainability challenges
4. 5:02Energy management pillars
5. 8:15Hardware and software efficiency
6. 10:03Energy proportional edge workloads
7. 14:00Adaptive orchestration
8. 16:27Proportional systems modeling
9. 22:31Prototype savings
10. 25:40Observability and federated learning
11. 28:30Energy time and space shifting
12. 32:41Intent-driven power savings
13. 37:36Federated Nearby Computing
14. 39:01Standardization needs
15. 40:49AI policies and emissions
16. 41:52Attestation and transparency
17. 42:35Q&A