English

Modeling the Impact of Fiber Latency on Compute-Communication Overlap in Geo-Distributed Multi-Datacenter AI Training

Performance 2026-05-20 v1 Distributed, Parallel, and Cluster Computing
Authors: Ioannis Papavasileiou , Sairam Prabhakar , Indu Kant Deo , Sergejs Makovejs
View on arXiv ↗ PDF ↗

Abstract

We use discrete-event simulation to quantify the impact of fiber latency on the efficacy of geo-distributed AI model training with data parallelism. We conclude that the optimum distances between two AI clusters is 10-100km, over which hollow-core fiber enables 25% higher compute-communication overlap.

Keywords

fiber latency compute-communication overlap distributed AI training data parallelism hollow-core fiber

Cite

@article{arxiv.2605.19169,
  title  = {Modeling the Impact of Fiber Latency on Compute-Communication Overlap in Geo-Distributed Multi-Datacenter AI Training},
  author = {Ioannis Papavasileiou and Sairam Prabhakar and Indu Kant Deo and Sergejs Makovejs},
  journal= {arXiv preprint arXiv:2605.19169},
  year   = {2026}
}

Comments

4 pages, 4 figures. Submitted to ECOC 2026

Related papers

View all related →
Distributed, Parallel, and Cluster Computing · Computer Science
Characterizing Compute-Communication Overlap in GPU-Accelerated Distributed Deep Learning: Performance and Power Implications
Seonho Lee, Jihwan Oh, Junkyum Kim, Seokjin Go +2
2025-07-08
Information Theory · Computer Science
Deploying Federated Learning in Large-Scale Cellular Networks: Spatial Convergence Analysis
Zhenyi Lin, Xiaoyang Li, Vincent K. N. Lau, Yi Gong +1
2021-03-11
Networking and Internet Architecture · Computer Science
Interplay between Distributed AI Workflow and URLLC
Milad Ganjalizadeh, Hossein S. Ghadikolaei, Johan Haraldson, Marina Petrova
2022-08-03
Computation and Language · Computer Science
CO2: Efficient Distributed Training with Full Communication-Computation Overlap
Weigao Sun, Zhen Qin, Weixuan Sun, Shidi Li +4
2024-01-30
Networking and Internet Architecture · Computer Science
Optimizing Energy and Latency in 6G Smart Cities with Edge CyberTwins
Amine Abouaomar, Badr Ben Elallid, Nabil Benamar
2025-11-07
Networking and Internet Architecture · Computer Science
Latency Versus Survivability in Geo-Distributed Data Center Design
Rodrigo de Souza Couto, Stefano Secci, Miguel Elias Mitre Campista, Luís Henrique Maciel Kosmalski Costa
2015-10-19
Machine Learning · Computer Science
CAFE: Carbon-Aware Federated Learning in Geographically Distributed Data Centers
Jieming Bian, Lei Wang, Shaolei Ren, Jie Xu
2024-02-07
Distributed, Parallel, and Cluster Computing · Computer Science
Hiding Communication Cost in Distributed LLM Training via Micro-batch Co-execution
Haiquan Wang, Chaoyi Ruan, Jia He, Jiaqi Ruan +3
2024-11-26
Machine Learning · Computer Science
Echo: Simulating Distributed Training At Scale
Yicheng Feng, Yuetao Chen, Kaiwen Chen, Jingzong Li +6
2024-12-18
Machine Learning · Computer Science
Consensus Control for Decentralized Deep Learning
Lingjing Kong, Tao Lin, Anastasia Koloskova, Martin Jaggi +1
2021-06-21
Optics · Physics
Beyond the fundamental noise limit in coherent optical fiber links
C. E. Calosso, E. Bertacco, D. Calonico, C. Clivati +6
2015-06-19
Machine Learning · Computer Science
Analyzing the benefits of communication channels between deep learning models
Philippe Lacaille
2019-04-22
Machine Learning · Computer Science
Optimizing Pipelined Computation and Communication for Latency-Constrained Edge Learning
Nicolas Skatchkovsky, Osvaldo Simeone
2019-06-13
Distributed, Parallel, and Cluster Computing · Computer Science
Data movement limits to frontier model training
Ege Erdil, David Schneider-Joseph
2024-11-14
Machine Learning · Computer Science
Personalizing Federated Learning with Over-the-Air Computations
Zihan Chen, Zeshen Li, Howard H. Yang, Tony Q. S. Quek
2023-02-27
Systems and Control · Electrical Eng. & Systems
Delay-Tolerant Augmented-Consensus-based Distributed Directed Optimization
Mohammadreza Doostmohammadian, Narahari Kasagatta Ramesh, Alireza Aghasi
2025-10-06
Computer Vision and Pattern Recognition · Computer Science
Decentralized Diffusion Models
David McAllister, Matthew Tancik, Jiaming Song, Angjoo Kanazawa
2025-01-13
Distributed, Parallel, and Cluster Computing · Computer Science
MAD Max Beyond Single-Node: Enabling Large Machine Learning Model Acceleration on Distributed Systems
Samuel Hsia, Alicia Golden, Bilge Acun, Newsha Ardalani +4
2024-06-12
Information Theory · Computer Science
Dynamic Clustering and Power Control for Two-Tier Wireless Federated Learning
Wei Guo, Chuan Huang, Xiaoqi Qin, Lian Yang +1
2022-05-20
Machine Learning · Computer Science
Over-the-Air Federated Learning in Cell-Free MIMO with Long-term Power Constraint
Yifan Wang, Cheng Zhang, Yuanndon Zhuang, Mingzeng Dai +2
2024-10-24
Machine Learning · Computer Science
CALF: Communication-Aware Learning Framework for Distributed Reinforcement Learning
Carlos Purves, Pietro Lio'
2026-03-16
Machine Learning · Computer Science
Overlap Local-SGD: An Algorithmic Approach to Hide Communication Delays in Distributed SGD
Jianyu Wang, Hao Liang, Gauri Joshi
2020-02-25
Distributed, Parallel, and Cluster Computing · Computer Science
Distributed SLIDE: Enabling Training Large Neural Networks on Low Bandwidth and Simple CPU-Clusters via Model Parallelism and Sparsity
Minghao Yan, Nicholas Meisburger, Tharun Medini, Anshumali Shrivastava
2022-02-01
Networking and Internet Architecture · Computer Science
Stochastic Modeling and Resource Dimensioning of Multi-Cellular Edge Intelligent Systems
Jaume Anguera Peris, Joakim Jaldén
2026-01-26
Machine Learning · Computer Science
Time Minimization in Hierarchical Federated Learning
Chang Liu, Terence Jie Chua, Jun Zhao
2022-10-11