电信科学 ›› 2023, Vol. 39 ›› Issue (8): 127-135.doi: 10.11959/j.issn.1000-0801.2023153

• 专栏:算力网络 • 上一篇    

通算一体驱动的算力内生网络技术与应用

孙杰, 马雷明, 杨爱东, 叶晓舟, 吕鹏, 王志刚, 欧阳晔   

  1. 亚信科技(中国)有限公司,北京 100193
  • 修回日期:2023-08-07 出版日期:2023-08-01 发布日期:2023-08-01
  • 作者简介:孙杰(1983- ),男,现就职于亚信科技(中国)有限公司研发中心云网规划部,主要研究方向为通信与5G网络智能化
    马雷明(1985- ),男,现就职于亚信科技(中国)有限公司网络与网管产品研发中心,主要研究方向为 5G 专网、算力网络、算力内生网络
    杨爱东(1984- ),男,博士,亚信科技(中国)有限公司通信人工智能实验室首席数据科学家,主要研究方向为 5G 无线通信、大数据挖掘、机器学习及其应用
    叶晓舟(1980- ),男,博士,亚信科技(中国)有限公司通信人工智能实验室主任,主要研究方向为通信人工智能和未来网络技术
    吕鹏(1980- ),男,亚信科技(中国)有限公司 5G 无线网络产品研发总监,主要研究方向为无线网络和网络智能化技术
    王志刚(1979- ),男,亚信科技(中国)有限公司 5G 网络产品研发总监,主要研究方向为核心网和网络智能化技术
    欧阳晔(1981- ),男,博士,IEEE Fellow,亚信科技(中国)有限公司首席技术官、高级副总裁,主要研究方向为移动通信、人工智能、数据科学、科技研发创新与管理

Computing native network (CNN) technology and application driven by joint communication & computing (JC&C)

Jie SUN, Leiming MA, Aidong YANG, Xiaozhou YE, Peng LYU, Zhigang WANG, Ye OUYANG   

  1. AsiaInfo Technologies (China) Co., Ltd., Beijing 100193, China
  • Revised:2023-08-07 Online:2023-08-01 Published:2023-08-01

摘要:

全球通信业均开展 6G 潜在技术架构与关键技术的研究与探索,其中通过通算一体更好地赋能各项6G网络新技术和新业务,从而进一步实现6G智能普惠愿景已成为业界共识。因此,如何实现通信与计算的深度融合与实时协同,以满足动态复杂网络环境下的新业务超低时延与算力实时响应等需求,成为通算一体的挑战与关键。提出一种通算一体驱动的算力内生网络,其通过算力时分复用与智能通算调度技术实现了基站的通信与计算算力解耦、算力池化与算力智能调度,从而实现基站同时支持通信与计算业务。该技术已经过验证并试点应用,实验与试点结果表明,算力内生网络技术可有效提升无线网络的算力效率,并实现基站同时支持通信与计算业务,助力6G通算一体演进。

关键词: 通算一体, 算力内生网络, 感知单元, 智能通算调度

Abstract:

The global telecom industry is carrying out research and exploration of potential 6G technical architectures and key techs, of which it has become a consensus in the industry to further realize the vision of 6G smart inclusion by better enabling all new 6G network technologies and services through joint communication & computing (JC&C).Therefore, how to realize the deep integration and real-time collaboration of communication and computing to meet the needs of ultra-low latency and real-time response of new services in dynamic and complex network environments has become the challenge and key of JC&C.A computing native network (CNN) driven by JC&C was proposed, which realized the decoupling of communication and computing of base station or multiple base stations, computing pooling and intelligent JC&C scheduling through computing time division multiplexing and intelligent JC&C scheduling, thus realizing the base station to support communication and computation services simultaneously.The technology has been technically verified and piloted, and the experimental and pilot results demonstrate that CNN can effectively improve the computing utilization rate of wireless networks and realize the base station to support communication and computing services simultaneously, facilitating the evolution of 6G JC&C.

Key words: joint communication &computing, computing native network, awareness unit, intelligent JC&C scheduling

中图分类号: 

No Suggested Reading articles found!