Research on Cloud-Edge Collaborative Distributed Training and Inference Solution

LI Zhiyong; XI Zhuoning; TIAN Wen; ZHANG Ming; ZHENG Kunsong; CAO Yuanming; SUN Ao; WU Zhenyu; DING Guoqiang; WEI Yizhe

doi:10.11959/j.issn.1000-0801.DXKX260285

您当前的位置：

首页 >

文章列表页 >

Research on Cloud-Edge Collaborative Distributed Training and Inference Solution

更新时间：2026-06-24

- Research on Cloud-Edge Collaborative Distributed Training and Inference Solution
- Telecommunications Science (2026)
- 作者机构：
  
  1.中国移动通信集团有限公司，北京，100033
  2.中国移动通信集团设计院有限公司，北京，100080
  3.中国移动通信集团有限公司供应链管理中心，北京，100053
- 作者简介：
- 基金信息：
- DOI：10.11959/j.issn.1000-0801.DXKX260285
  CLC： TP393
- Received：02 May 2026，
  
  Revised：2026-06-12，
  
  Accepted：22 June 2026，
- 稿件说明：
移动端阅览
LI Zhiyong, XI Zhuoning, TIAN Wen, et al. Research on Cloud-Edge Collaborative Distributed Training and Inference Solution[J/OL]. Telecommunications Science, 2026.
DOI：

LI Zhiyong, XI Zhuoning, TIAN Wen, et al. Research on Cloud-Edge Collaborative Distributed Training and Inference Solution[J/OL]. Telecommunications Science, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260285.

摘要

企业智能化转型中，大模型训练与推理需要大量智能算力资源，自建算力往往成本高、建设周期长，而租赁算力服务则面临数据出域安全风险等问题。本文提出云边协同分布式训推方案，通过模型分层部署，边缘企业侧处理敏感数据，云端智算中心承载大规模计算，中间通过支持RDMA的跨域无损传输网络传递敏感性更低的中间变量，在保障企业数据隐私性的基础上，依靠安全网络架构，实现算力弹性扩展，形成适配不同行业场景的高性能组网解决方案。

Abstract

In the context of enterprise intelligent transformation

large model training and inference require massive intelligent computing resources. Self-built computing power often suffers from high costs and lengthy construction cycles

while rented computing power services face issues such as data exfiltration security risks. This paper proposes a Cloud-Edge Collaborative Distributed Training and Inference Solution. Through hierarchical model deployment

the enterprise edge side processes sensitive data

while the cloud-based intelligent computing center undertakes large-scale computing tasks. Between the cloud and edge ends

non-sensitive intermediate variables are transmitted via an RDMA (Remote Direct Memory Access)-supported cross-domain lossless transmission network. On the basis of ensuring enterprise data privacy

the solution achieves elastic scaling of computing power based on a secure network architecture

and finally forms a high-performance networking solution that is adaptable to diverse industry scenarios.

关键词

Keywords

references

Li Z , Wu S , Li L , et al . Energy-Efficient Split Learning for Fine-Tuning Large Language Models in Edge Networks [J ] . IEEE Networking Letters , 2025 , 7 ( 3 ): 176 - 180 . DOI: 10.1109/LNET.2025.3530430 http://dx.doi.org/10.1109/LNET.2025.3530430 .

Yang Q , Dong Z , Liang S , et al . Research on Data Privacy and Information Security Evaluation and Optimization for Large Language Models [C ] . 2025 6th International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering (ICBAIE) , Octobe r17 -19 2025, Shanghai, China : IEEE ,2025. DOI: 10.1109/ICBAIE66852.2025.11326645 http://dx.doi.org/10.1109/ICBAIE66852.2025.11326645 .

Yang Z , Jin Y , Zhang Y , et al . Research on Large Language Model Cross-Cloud Privacy Protection and Collaborative Training Based on Federated Learning [C ] . 2025 IEEE 6th International Seminar on Artificial Intelligence, Networking and Information Technology (AINIT) , April 11 -13 2025, Shenzhen,China : IEEE , 2025. DOI: 10.1109/AINIT65432.2025.11035133 http://dx.doi.org/10.1109/AINIT65432.2025.11035133 .

Internet Engineering Task Force (IETF) . OSPFv3 as a Provider Edge to Customer Edge (PE-CE) Routing Protocol:IS SN:2070-1721 [S ] . 2012 .

Guo C , Wu H , Deng Z , et al . RDMA over commodity ethernet at scale [C ] . SIGCOMM '16 : ACM SIGCOMM 2016 Conference , August 22 -26 2016,Florianopolis,Brazil:Association for Computing Machinery, New York , United States ,2016: 202 - 215 .ISBN:978-1-4503- 4193 - 6 .

Catherine J. Doughty , Michael H. Long . The Handbook of Second Language Acquisition [M ] . Blackwell Publishing Ltd , 2003 . Print ISBN:9780631217541,Online ISBN: 9780470756492 , DOI: 10.1002/9780470756492 http://dx.doi.org/10.1002/9780470756492 .

Zhao W , Zhou K , Li J , et al . A Survey of Large Language Models [EB/OL ] . ( 2023-03-31 )[ 2026-3-26 ] . arXiv: 2303 . 18223 v19 [cs.CL ] . https://arxiv.org/abs/2303.18223 https://arxiv.org/abs/2303.18223 .

Naveed H , Khan A , Qiu S , et al . A Comprehensive Overview of Large Language Models [J ] . ACM Transactions on Intelligent Systems and Technology , 2025 , 16 ( 5 ): 1 - 72 .

Shanahan M . Talking about Large Language Models [J ] . Communications of the ACM , 2024 , 67 ( 2 ): 68 - 79 .

Subakan C , Ravanelli M , Cornell S , et al . Attention Is All You Need In Speech Separation [C ] . ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , June 06 -11 2021, Toronto, ON, Canada : IEEE ,2021. DOI: 10.1109/ICASSP39728.2021.9413901 http://dx.doi.org/10.1109/ICASSP39728.2021.9413901

Lei Y , Shen T , Cao Y , et al . Enhancing Lexicon-Based Text Embeddings with Large Language Models [C ] . 63rd Annual Meeting of the Association for Computational Linguistics , July 2025 ,Vienna, Austria. ACL : Association for Computational Linguistics , 2025 : 18986 - 19001 . DOI: 10.18653/v1/2025.acl-long.930 http://dx.doi.org/10.18653/v1/2025.acl-long.930

Matsuda K , Frohlich S , Wang M , et al . Embedding by Unembedding [J ] . Proceedings of the ACM on Programming Languages , 2023 , 7 ( 189 ): 1 - 47 .

M S , Kiruba G , J S , et al . Large Language Model AI Text Generation Detection based on Transformer Deep Fast Quantum Convolutional Neural Networks [C ] 2025 6th International Conference on Inventive Research in Computing Applications (ICIRCA) , June 25 -27 2025,Coimbatore, India. IEEE ,2025. Electronic ISBN :979-8- 3315-2142 - 4 .

Vig J , Belinkov . Analyzing the Structure of Attention in a Transformer Language Model [C ] Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP , August 2019 ,Florence, Italy. Association for Computational Linguistics , 2019 : 63 - 76 .

Yao D , Li B . Is Split Learning Privacy-Preserving for Fine-Tuning Large Language Models? [C ] . IEEE Transactions on Big Data（Early Access） , 2024 : 1 - 12 . DOI: 10.1109/TBDATA.2024.3524101 http://dx.doi.org/10.1109/TBDATA.2024.3524101 .

Qu W , Zhou Y , Wu Y , et al . Prompt Inversion Attack against Collaborative Inference of Large Language Models [C ] . 2025 IEEE Symposium on Security and Privacy（SP），May 12 -15 2025,San Francisco CA USA. IEEE , 2025. DOI: 10.1109/SP61157.2025.00160 http://dx.doi.org/10.1109/SP61157.2025.00160 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Resource scheduling optimization in cloud-edge collaboration

Cloud-edge collaboration based computer vision inference mechanism

Research on a DSTM-based trusted mesh security mechanism for 6G distributed autonomous networks

Multimodal sequential recommendation method based on time-frequency transformation and dynamic graph attention

Low-complexity adaptive phase optimization algorithm for IRS-assisted MIMO system

Related Author

Shuling WANG

Jie SUN

Peng WANG

Aidong YANG

Boheng TANG

Xingang CHAI

Related Institution

No data

⁰