Skip to content

M-CSN: Joint Architecture and Flow Scheduling for Metro-Scale AI Fabric Based on Supernodes

2026 · IEEE Transactions on Network Science and Engineering · Vol 13, pp. 11513-11531 · 0 citations · 78 references

Abstract

Deploying trillion-parameter large language models across metropolitan environments is required to sustain real-time inference. Urban power constraints, however, prohibit monolithic GPU clusters, forcing the integration of distributed supernodes into a citywide compute fabric. Over 100-km distances, optical propagation delays invalidate reactive congestion control for 12.8 Tbps cross-domain pipeline parallelism (PP) traffic. The resulting high bandwidth-delay product causes in-flight data to saturate edge switch buffers before the feedback loop triggers source throttling. We address this limitation with M-CSN, an AI compute architecture featuring LOTUS, a model-driven scheduling engine for metropolitan computing power networks (CPNs). LOTUS replaces delayed reactive signaling with a proactive bi-level scheduling strategy. At the macro-layer, an inverse-SLA mechanism partitions capacity among concurrent flows to guarantee performance isolation. At the micro-layer, the scheduler decouples transmission rates from dynamic window probing: it injects calculated pacing rates and static hyper-BDP windows into source RDMA queue pairs (QPs). This source-side enforcement eliminates feedback lag. Packet-level simulations demonstrate that LOTUS sustains a 92% operational load with PFC-free execution. Eliminating pause-induced jitter and buffer saturation reduces critical-flow latency by 2.6× relative to reactive baselines. This enables fragmented urban resources to operate as a unified, deterministic, supernode-based computing power network.

View source

Similar papers

Book Open access Aug 2026

CSIG: Congestion Signaling for Datacenter Transports

This work introduces CSIG, a protocol that delivers precise, multi-bit bottleneck congestion signals via a fixed-length Ethernet header, and proposes Fast Ramp-Up, a congestion control primitive that leverages these bottleneck signals to reduce median RPC latency by 20% and unclaimed bandwidth by 60% in production.

Abhiram Ravi, Nandita Dukkipati, Weiwu Pang et al. · 0 citations
Sep 2026

The network fabric as a storage orchestrator: a switch-centric architecture

A new class of highly scalable, effective, and resilient network-native services is made possible by Synapse, which removes the main storage bottleneck in contemporary data centers by offloading the entire telemetry capture process to the network fabric.

Seyed Hossein Ahmadpanah, A. Sahafi, S. Erfani · 0 citations
Preprint Aug 2026

Scaling 5G-TSN Bridges: Operating Regimes, Scheduling, and Time Synchronisation Under Heterogeneous Industrial Traffic

The nascTime framework on OMNeT++/Simu5G is used to evaluate how many TSN endpoints a single 5G NR cell can bridge before per-flow QoS degrades, showing that sub-3 ms TSN deadlines may require radio-configuration changes such as configured grants or higher numerology.

Mohamed A. M. Seliem, U. Roedig, C. Sreenan et al. · 0 citations
Open access Aug 2026

Hierarchical Scheduler with Adaptive Time-Budget Reallocation for Time-Triggered Edge-Fog-Cloud Architectures

The lack of determinism restricts the integration of safety-critical applications into Edge–Fog–Cloud (EFC) architectures. Existing EFC schedulers are typically designed for dynamic, best-effort operation based on unmanaged resource allocation and elastic virtualization. This paradigm introduces unbounded queueing, res...

Omar Hekal, Josepaul Paulachan, Daniel Onwuchekwa et al. · 0 citations
Book Open access Aug 2026

CCSwitch: A Scalable Data Plane for Non-Blocking In-Network Collective Communication

This work presents CCSwitch, a modular switching fabric built from 4×4 non-blocking Collective Engines, a modular switching fabric built from 4×4 non-blocking Collective Engines (CEs) that combines spatial and temporal parallelism to perform reductions without accumulation buffers.

Sumukh Pinge, Hardik Soni, Bob Lantz et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.