算网融合背景下的异构算力调度系统和应用

龚豪

doi:10.11959/j.issn.1000-0801.DXKX260013

您当前的位置：

首页 >

文章列表页 >

算网融合背景下的异构算力调度系统和应用

工程与应用 | 更新时间：2026-07-21

- 算网融合背景下的异构算力调度系统和应用
- Heterogeneous computing power scheduling system and applications under the background of computing- network convergence
- 电信科学 2026年42卷第6期页码：243-256
- 作者机构：
  
  中国电信股份有限公司上海分公司，上海 200120
- 作者简介：
  
  龚豪（1982- ），男，中国电信股份有限公司上海分公司副总经理，主要研究方向为人工智能、云计算、算力网络技术及产品创新。
- 基金信息：
- DOI：10.11959/j.issn.1000-0801.DXKX260013
  中图分类号： TP319
- 收稿：2025-11-19，
  
  修回：2026-02-25，
  
  录用：2026-03-06，
  
  网络首发：2026-07-21，
  
  纸质出版：2026-06-20
- 稿件说明：
移动端阅览
龚豪.算网融合背景下的异构算力调度系统和应用[J].电信科学,2026,42(06):243-256.

Gong Hao.Heterogeneous computing power scheduling system and applications under the background of computing- network convergence[J].Telecommunications Science,2026,42(06):243-256.
龚豪.算网融合背景下的异构算力调度系统和应用[J].电信科学,2026,42(06):243-256. DOI： 10.11959/j.issn.1000-0801.DXKX260013.

Gong Hao.Heterogeneous computing power scheduling system and applications under the background of computing- network convergence[J].Telecommunications Science,2026,42(06):243-256. DOI： 10.11959/j.issn.1000-0801.DXKX260013.

摘要

针对我国近年算网融合及算网基础设施发展的背景，阐述了异构算力调度的系统定位和功能架构，提炼了系统实现的4项关键组件，包括屏蔽异构算力差异的算力调度适配器、基于调度算法执行两级调度的作业调度控制器、提供算力感知和算网路由功能的算网资源编排模块、实现算力交易一体化服务的算力服务3中心模块，针对调度效果多维约束给出了一套多目标调度算法模型。结合“东数西算”、人工智能（artificial intelligence，AI）训推协同、超算互联、视频直播等场景探讨了异构算力调度的重要作用，结合工程实践给出了实际部署案例，并对异构算力调度在算网融合演进中的创新和应用前景进行了总结。

Abstract

In view of the development of computing-network convergence and computing-network infrastructure in China in recent years

the system positioning and functional architecture of heterogeneous computing power scheduling were expounded

and four key components for system implementation were extracted

including computing power scheduling adapters that shielded the differences of heterogeneous computing power

a job scheduling controller that performed two-level scheduling based on scheduling algorithms

a computing-network resource orchestration module that provided computing awareness and computing-network routing functions

and three center modules for computing power services that realized integrated computing power transaction services. A multi-objective scheduling algorithm model was proposed for the multi-dimensional constraints of scheduling effects. Combined with scenarios such as the East Data West Computing project

artificial intelligence training-inference collaboration

supercomputing interconnection

and live video streaming

the important role of heterogeneous computing power scheduling was discussed. A practical deployment case was given combined with engineering practice

and the innovation and application prospects of heterogeneous computing power scheduling in the evolution of computing-network convergence were summarized.

关键词

Keywords

references

YD/T 4255-2023 . 算力网络总体技术要求 [S ] .

YD/T 4255-2023 . General technical requirements for computing power network [S ] .

工业和信息化部 . 算力互联互通行动计划: 工信部信管〔2025〕119 号[R ] . 2025 .

Ministry of Industry and Information Technology . Action plan for interoperability of computing power: MIIT Letter [2025 ] No. 119[R ] . 2025 .

中国电信股份有限公司研究院 . 《分布式智算中心无损网络技术白皮书》 [R ] . 2024 .

Research Institute of China Telecom Co., Ltd . White paper on lossless network technology for distributed intelligent computing centers [R ] . 2024 .

唐卓 , 蒋冰婷 , 张嘉鹏 , 等 . 算力网络调度基础理论与关键技术现状及展望 [J ] . 中国科学基金 , 2025 , 39 ( 2 ): 240 - 249 .

Tang Z , Jiang B T , Zhang J P , et al . The foundation theory and key technologies of computing first network scheduling: present situation and prospects [J ] . Bulletin of National Natural Science Foundation of China , 2025 , 39 （ 2 ): 240 - 249 .

YD/T 6064-2024 . 算力网络算网编排管理技术要求 [S ] .

YD/T 6064-2024 . Technical requirements of orchestration management of computing and network convergence [S ] .

YD/T 6044-2024 . 算力网络算力度量与算力建模技术要求 [S ] .

YD/T 6044-2024 . Technical requirements of computing measurement and computing modeling in computing and network convergence [S ] .

王蕴 , 林霄 , 楼芝兰 , 等 . 面向边缘光算力网络的上行链路资源协同调度算法 [J ] . 光通信技术 , 2024 , 48 ( 3 ): 45 - 51 .

Wang Y , Lin X , Lou Z L , et al . Uplink resource coordinated scheduling algorithm for edge-oriented optical computing power networks [J ] . Optical Communication Technology , 2024 , 48 ( 3 ): 45 - 51 .

Yi C Y , Cai J , Su Z . A multi-user mobile computation offloading and transmission scheduling mechanism for delay-sensitive applications [J ] . IEEE Transactions on Mobile Computing , 2020 , 19 ( 1 ): 29 - 43 .

归律 , 王悦 , 张钟琴 . 东数西算中多云管理技术应用浅析 [J ] . 信息通信技术与政策 , 2022 , 48 ( 11 ): 84 - 92 .

Gui L , Wang Y , Zhang Z Q . The application of multicloud management technology in east-data-west-computing project [J ] . Information and Communications Technology and Policy , 2022 , 48 ( 11 ): 84 - 92 .

Afifi H , Pochaba S , Boltres A , et al . Machine learning with computer networks: techniques, datasets, and models [J ] . IEEE Access , 2024 , 12 : 54673 - 54720 .

叶宁 , 付康 , 胡少文 , 等 . 基于MPI的异构算力资源融合调度平台 [J ] . 计算机与现代化 , 2025 ( 12 ): 38 - 45 .

Ye N , Fu K , Hu S W , et al . MPI-based heterogeneous computing resource integration and scheduling platform [J ] . Computer and Modernization , 2025 ( 12 ): 38 - 45 .

Long Y X , Wang Z Y , Lan S Z , et al . Energy-latency tradeoff for task offloading and resource allocation in vehicular edge computing [J ] . Computer Networks , 2025 , 258 : 111026 .

Pang S C , Cheng Y Z , He X , et al . Multi-user motion state task offloading strategy for load balancing in mobile edge computing networks [J ] . Ad Hoc Networks , 2025 , 169 : 103759 .

Mahesar A R , Li X P , Sajnani D K . Enhancing task scheduling and QoS optimization in mobile edge computing via microservice-oriented container selection [J ] . Computing , 2025 , 107 ( 2 ): 60 .

Deb K , Pratap A , Agarwal S , et al . A fast and elitist multiobjective genetic algorithm: NSGA-II [J ] . IEEE Transactions on Evolutionary Computation , 2002 , 6 ( 2 ): 182 - 197 .

陆威然 , 鲁腾飞 , 赵恪振 , 等 . 基于博弈论-VIKOR的Kubernetes调度策略研究 [J ] . 计算机时代 , 2023 ( 8 ): 52 - 56 .

Lu W R , Lu T F , Zhao K Z , et al . Kubernetes scheduling strategy based on game theory-VIKOR [J ] . Computer Era , 2023 ( 8 ): 52 - 56 .

中国电信股份有限公司上海分公司 , 华为技术有限公司 . 50G PON全光万兆发展白皮书 [R ] . 2023 .

China Telecom Corp. , Ltd . Shanghai Branch, Huawei Technologies Co., Ltd. White paper on all-optical 10G development based on 50G PON [R ] . 2023 .

马智勇 , 张坚平 , 严森垒 . 基于50G-PON的全光万兆示范与应用 [J ] . 电信科学 , 2025 , 41 ( Z1 ): 274 - 278 .

Ma Z Y , Zhang J P , Yan S L . Demonstration and application of all-optical 10G access based on 50G-PON [J ] . Telecommunications Science , 2025 , 41 ( Z1 ): 274 - 278 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于数据增强技术的光传送网健康度模型测评研究

面向6G智能体网络的行为安全协同架构

智能体互联网服务语义标识和路由技术研究

面向网络运维的智能体体系构建

基于机器学习的SRv6动态性能优化方法研究