Metraipolis › archive › dc_86fe1b49337b

Halley: A Predictive Control Plane for Optical AI Data Center Networks — arXiv:2611.04408

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.04408
License (as archived)
CC0-1.0
Archived
2026-11-04T18:57:39.846Z
SHA-256 of the archived text
415fbef2c94cb1e80b177ee2ed6f25521deb600aab2e74d85a8f0338b4482d7b
Stable URL of this page
https://metraipolis.com/documento/dc_86fe1b49337b

How to cite

Halley: A Predictive Control Plane for Optical AI Data Center Networks — arXiv:2611.04408 — arXiv (en), https://arxiv.org/abs/2611.04408 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_86fe1b49337b (Metraipolis archive dc_86fe1b49337b, sha256 415fbef2c94cb1e80b177ee2ed6f25521deb600aab2e74d85a8f0338b4482d7b)
BibTeX
@misc{dc86fe1b49337b,
  title = {Halley: A Predictive Control Plane for Optical AI Data Center Networks — arXiv:2611.04408},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04408},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_86fe1b49337b; sha256 415fbef2c94cb1e80b177ee2ed6f25521deb600aab2e74d85a8f0338b4482d7b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04408. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A frontier AI model is trained in an AI data center, on tens of thousands of accelerators joined by a dedicated network, and every training step ends with all of them exchanging data at once. Between racks, that network is optical, and each of its four layers, from lasers to switches, runs a fixed or after-the-fact policy. Lasers sharing an enclosure, cooling, and power can fail together. At a fixed spare ratio, correlated failure provably caps a pool's availability however large it grows. In this paper, we present a comprehensive predictive control plane framework, called Halley, that reads each resource's next state and acts before the failure occurs. It has four components, each carrying a formal guarantee: LightReserve that sizes the laser pool to provide a coverage bound, FoldFit that schedules the switch within $11/9$ of the optimal, ThermaLead that holds a microring with one-step-ahead control, and TriageGuard that protects the highest-risk modules first. Halley gives each layer a predictive controller that defers to the deployed policy when unsure. Composed from simulations at that point, effective goodput, the share of ideal throughput kept, rises from 65.2\% to 76.5\% if a window that exhausts its spares is lost. If only the dark channels' bandwidth is lost, it reaches 67.4\%. Per megawatt, that is up to 1.03--1.17 times the training; for the same training, up to 3.3--14.7\% less data center capacity. \textsf{Halley}'s fallback circuits are formally verified and are shown to achieve close timing on a commercial FPGA, and pass a 130~nm signoff.

Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.NI, nuovi lavori del giorno).
Categorie arXiv: Networking and Internet Architecture (cs.NI).
Pagina del lavoro: https://arxiv.org/abs/2611.04408 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-04T18:36:07.106Z dalla pagina https://arxiv.org/list/cs.NI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).