Halley paper proposes a predictive control plane for optical AI data center networks
Manav Mishra · hf · 2026-11-06
A paper from Prompteer.ai and IIT Madras presents Halley, a predictive control plane for optical AI data center networks. At hundred-thousand-accelerator scale, the inter-rack optical network runs a fixed fail-then-fix policy across four layers (lasers to switches), and at a fixed spare ratio correlated failures provably cap pool availability.
Halley reads each resource's next state and acts before failure occurs, with four components carrying formal guarantees: LightReserve (laser pool sizing to a coverage bound), FoldFit (switch scheduling within 11/9 of optimal), ThermalLead (one-step-ahead microring heater control), and TriageGuard (protecting highest-risk modules first). Deployed policies replace reactive short-fixes—either fixing in advance or catching events it should have anticipated—improving effective throughput and reducing capacity needs.
More from Infra
- Veteran investor: memory stocks aren't cheap even at P/E of 3 as LTAs are meaningless — bookwormengr · 2026-11-06
- Enterprise AI's future is local: tactical workloads favor Dell, HPE, Lenovo, Supermicro over frontier labs — DavidLinthicum · 2026-11-06
- Texas Governor-Elect: Approved Data Centers Can Be Completed, New Projects to Pay Taxes — GavinSBaker · 2026-11-06
- Marin's open 535B-A23B MoE finishes main pretraining on 18T tokens, checkpoints soon — dlwh · 2026-11-06
- Buoyant Open-Sources Kinference, an Inference Control Plane for vLLM on Kubernetes — vllm_project · 2026-11-06
- vLLM ranks #2 most-engaged open-source repo in GitHub Octoverse, #8 by contributor growth — vllm_project · 2026-11-06