计算机科学 ›› 2022, Vol. 49 ›› Issue (9): 249-259.doi: 10.11896/jsjkx.220500222

• 计算机网络 • 上一篇    下一篇

融智算力网络及其功能架构

胡玉姣1, 贾庆民1, 孙庆爽2, 谢人超1,3, 黄韬1,3   

  1. 1 网络通信与安全紫金山实验室未来网络研究中心 南京 211111
    2 西北工业大学计算机学院 西安 710072
    3 北京邮电大学网络与交换国家重点实验室 北京 100876
  • 收稿日期:2022-05-23 修回日期:2022-06-24 出版日期:2022-09-15 发布日期:2022-09-09
  • 通讯作者: 胡玉姣(huyujiao@pmlabs.com.cn)
  • 基金资助:
    国家自然科学基金面上项目(62171046)

Functional Architecture to Intelligent Computing Power Network

HU Yu-jiao1, JIA Qing-min1, SUN Qing-shuang2, XIE Ren-chao1,3, HUANG Tao1,3   

  1. 1 Future Network Research Center,Purple Mountain Laboratories,Nanjing 211111,China
    2 School of Computer Science,Northwestern Polytechnical University,Xi'an 710072,China
    3 State Key Laboratory of Networking and Switching Technology,Beijing University of Posts and Telecommunications,Beijing 100876,China
  • Received:2022-05-23 Revised:2022-06-24 Online:2022-09-15 Published:2022-09-09
  • About author:HU Yu-jiao,born in 1993,Ph.D,is a member of China Computer Federation.Her main research interests include cyber-physical system,computing power network,edge computing,and deep learning.
  • Supported by:
    National Natural Science Foundation of China(62171046).

摘要: 面向算力网络新兴研究领域,针对其迫切需要提升智能性与精准服务能力的问题,提出了云边端算力资源、网络资源、智能模型及算法协同共生的融智算力网络建设理念,引入了内生智能和业务智能两个层面的智能性。内生智能指算力网络为保障系统准确运行所具备的自感知、自适应、自决策、自学习能力,业务智能指算力网络为增强对行业/应用的适应性所具备的智能资源封装及自主部署能力、业务编排与认知能力。进一步地,设计了层次化功能架构,从基础资源、资源管理、业务编排、运营服务以及系统优化5个层面明确了AI赋能融智算力网络内生智能与业务智能的具体表现。最后,将智能制造及智慧安防两类应用中的3个场景(车间物流、基于机器视觉的质检品控、社区及道路监测)作为仿真实验原型场景,并设计了对比实验组。实验结果表明,应用融智算力网络至车间物流场景中,性能提升幅度与场景规模有关,规划用时提升了约2~50倍,规划结果提升了约2~5倍;基于机器视觉的质检品控场景中,算力设备部署成本下降为原来的1/5、检测准确率提升约4.5%;社区及道路监测场景中,系统的算力部署成本可降低为原来的1/10。

关键词: 算力网络, 内生智能, 业务智能, 功能架构, 层次化模型

Abstract: The computing power network as a new research fieldis in urgent need to improve intelligence and provide on-demand services.To solve the problems,an intelligent computing power network that combines cloud-edge-terminal computing resource,communication resource,and AI approaches together is proposed.At the same time,the system-level intelligence would be built from two aspects of endogenous intelligence and application intelligence.The endogenous intelligence refers to the abilities of self-perception,self-adaptation,self-decision and self-learning to make the computing power network ensure the accurate operation of the system.The application intelligence refers to the abilities in resource deployment,business arrangement and cognition,so that the computing power network could enhance the adaptability of the businesses.Further,a functional architecture is constructed,which would gradually build the endogenous intelligence and application intelligence in the layers of basic resource,resource mana-gement,business choreography,operation service and optimization.Finally,three important scenarios are selected from two domains,i.e.workshop logistics and quality control based on machine vision in the smart manufacturing,road and community monitoring in the intelligent security.Three groups of simulations are designed based on the scenarios,respectively.Experimental results show that while adopting the intelligent computing power network with endogenous intelligent and application intelligence,in the scenario of workshop logistics,the performance improvement is related with the scale of the scenarios,the planning time could be improved by approximately 2~50 times and the planning result could be improved by about 2~5 times.In the scenario of quality control based on machine vision,the deploymentcost of computing equipment would reduce to 1/5 of the original and the detection accuracy could improve by about 4.5%.In the scenario of road and community monitoring,the deployment cost of computing equipment could reduceto 1/10 of the original.

Key words: Computing power network, Endogenous intelligence, Application intelligence, Functional architecture, Hierarchical model

中图分类号: 

  • TP399
[1]新型数据中心发展三年行动计划(2021-2023年)解读[OL].2021-07-16.http://www.gov.cn/zhengce/2021-07/16/content_5625389.htm.
[2]关于印发《全国一体化大数据中心协同创新体系算力枢纽实施方案》的通知[OL].2021-05-24.http://www.gov.cn/zhengce/zhengceku/2021-05/26/content_5612405.htm.
[3]中华人民共和国国家发展和改革委员会.东数西算[OL].https://www.ndrc.gov.cn/xwdt/ztzl/dsxs/?code=&state=123.
[4]“东数西算”让数字化“脚步”更快更稳[OL].2022-02-28.http://www.gov.cn/zhengce/2022-02/28/content_5675995.htm.
[5]工业和信息化部关于印发《新型数据中心发展三年行动计划(2021-2023年)》的通知[OL].2021-07-04.http://www.gov.cn/zhengce/zhengceku/2021-07/14/content_5624964.htm.
[6]China Mobile.Computing Force Network WhitePaper[M/OL].https://www.vzkoo.com/document/568a222bc5892fb727aa400ea6c8e828.html?keyword=%E7%AE%97%E5%8A%9B%E7%BD%91%E7%BB%9C.
[7]China Academy of Information and Communications Technology.White Paper on China's Computing Power Development Index[M/OL].http://www.caict.ac.cn/kxyj/qwfb/bps/202109/t20210918_390058.htm.
[8]China Mobile.Diversity Computing Technology Vision WhitePaper[M/OL].https://www.vzkoo.com/document/b2d9baf73ecc3e9e5448dd10df64b321.html?keyword=%E5%A4%9A%E6%A0%B7%E6%80%A7%E7%AE%97%E5%8A%9B.
[9]China Unicom.Computer Power Network WhitePaper[M/OL].https://wenku.baidu.com/view/5817c423b868a98271fe910ef12d2af90342a844.html.
[10]China Mobile,HuaWei.Computing-awareNetworking WhitePaper[M/OL].https://www.vzkoo.com/document/297a189187b473cd9132bf706651c333.html?keyword=%E7%AE%97%E5%8A%9B%E7%BD%91%E7%BB%9C%E7%99%BD%E7%9A%AE%E4%B9%A6.
[11]China Institute of Communications.Computing Network Frontier Report[R/OL].https://www.vzkoo.com/document/a9255c0ed9e4db735d93e42b1f3ab839.html?keyword=%E7%AE%97%E5%8A%9B%E5%89%8D%E6%B2%BF.
[12]LEI B,LIU Z Y,WANG X L,et al.Computing network:a new multi-access edge computing[J].Telecommunications Science,2019,35(9):44-51.
[13]DUAN X D,YAO H J,FU Y X,et al.Computing force network technologies for computing and network integration evolution[J].Telecommunications Science,2021,37(10):76-85.
[14]ZHANG S,CAO C,TANG X Y.Computing Power NetworkTechnology Architecture Based on SRv6[J/OL].ZTE Communications,2022.https://www.cnki.com.cn/Article/CJFDTotal-ZXTX202201005.htm.
[15]DI Z,CAO Y F,QIU C,et al.New computing power network architecture and its application case analysis [J/OL].Computer Application.https://d.wanfangdata.com.cn/periodical/jsjyy202206002.
[16]HUANG G P,SHI W Q,TAN B.Computing Power NetworkResources Based on SRv6 and Its Service Arrangement and Scheduling[J/OL].ZTE Communications.https://www.zte.com.cn/china/about/magazine/zte-communications/2021/cn202103/specialtopic/cn202103006.html.
[17]LI J F,CAO C,LI A,et al.Computing Power Modeling for Business Experience in Computing Power Network[J/OL].ZTE Communications.https://s.wanfangdata.com.cn/periodical?q=Mixed%20supervision%20for%20surface-defect%20detec-tion%3A%20From%20weakly%20to%20fully%20super-vised%20learning.
[18]GONG C Y,SHU H F,ZHANG X.Multi-level computing po-wer network centralized indivisible task scheduling algorithm[J/OL].ZTE Communications.https://www.zte.com.cn/china/about/magazine/zte-communications/2021/cn202103/specialtopic/cn202103008.
[19]LIU Z N,LI K,WU L T,et al.CATS:Cost Aware Task Sche-duling in Multi-Tier Computing Networks[J].Journal of Computer Research and Development,2020,57(9):1810-1822.
[20]SUN Y K,ZHANG X,LEI B.Research on intelligent computing power-aware routing allocation strategy in edge computing po-wer network[J].Radio Communications Technology,2022,48(1):60-67.
[21]JIA Q M,DING R,LIU H,et al.Survey on research progress for compute first networking[J].Chinese Journal of Network and Information Security,2021,7(5):1-12.
[22]HU Y J,YAO Y,LEE W S.A reinforcement learning approachfor optimizing multiple traveling salesman problems over graphs[J].Knowledge-Based Systems,2020,204:106244.1-106244.14.
[23]KAEMPFER Y,WOLF L.Learning the multiple travelingsalesmen problem with permutation invariant pooling networks[J].arXiv:1803.09621,2018.
[24]BOŽIČ J,TABERNIK D,SKOČAJ D.Mixed supervision for surface-defect detection:From weakly to fully supervised lear-ning[J].Computers in Industry,2021,129:103459.
[25]BOŽIČ J,TABERNIK D,SKOČAJ D.End-to-end training of a two-stage neural network for defect detection[C]//2020 25th International Conference on Pattern Recognition (ICPR).IEEE,2021:5619-5626.
[26]TABERNIK D,ŠELA S,SKVARČ J,et al.Segmentation-based deep-learning approach for surface-defect detection[J].Journal of Intelligent Manufacturing,2020,31(3):759-776.
[27]LOPES V,ALEXANDRE L A.Auto-classifier:A robust defect detector based on an automl head[C]//International Conference on Neural Information Processing.Cham:Springer,2020:137-149.
[28]RUDOLPH M,WANDT B,ROSENHAHN B.Same same butdiffernet:Semi-supervised defect detection with normalizing flows[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision.2021:1907-1916.
[29]ZAMIR S W,ARORA A,KHAN S,et al.Restormer:Efficient transformer for high-resolution image restoration[C]//Procee-dings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2022:5728-5739.
[30]DAI Z,LIU H,LE Q V,et al.Coatnet:Marrying convolution and attention for all data sizes[J].Advances in Neural Information Processing Systems,2021,34:3965-3977.
[31]XU Q,ZHOU Y,WANG W,et al.Spg:Unsupervised domain adaptation for 3d object detection via semantic point generation[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.2021:15446-15456.
[32]REDDY N D,GUIGUES L,PISHCHULIN L,et al.Tessetrack:End-to-end learnable multi-person articulated 3d pose tracking[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2021:15190-15200.
[1] 李梦菲, 毛莺池, 屠子健, 王瑄, 徐淑芳.
基于深度确定性策略梯度的服务器可靠性任务卸载策略
Server-reliability Task Offloading Strategy Based on Deep Deterministic Policy Gradient
计算机科学, 2022, 49(7): 271-279. https://doi.org/10.11896/jsjkx.210600040
[2] 魏鹏, 马玉亮, 袁野, 吴安彪.
用户行为驱动的时序影响力最大化问题研究
Study on Temporal Influence Maximization Driven by User Behavior
计算机科学, 2022, 49(6): 119-126. https://doi.org/10.11896/jsjkx.210700145
[3] 刘丽, 李仁发.
医疗CPS协作网络控制策略优化
Control Strategy Optimization of Medical CPS Cooperative Network
计算机科学, 2022, 49(6A): 39-43. https://doi.org/10.11896/jsjkx.210300230
[4] 沈超, 何希平.
基于纹理特征增强和轻量级网络的人脸防伪算法
Face Anti-spoofing Algorithm Based on Texture Feature Enhancement and Light Neural Network
计算机科学, 2022, 49(6A): 390-396. https://doi.org/10.11896/jsjkx.210600217
[5] 刘云, 董守杰.
基于CUDA核函数的多路视频图像拼接加速算法
Acceleration Algorithm of Multi-channel Video Image Stitching Based on CUDA Kernel Function
计算机科学, 2022, 49(6A): 441-446. https://doi.org/10.11896/jsjkx.210600043
[6] 朱旭辉, 沈国娇, 夏平凡, 倪志伟.
基于螺旋进化萤火虫算法和BP神经网络的模型及其在PPP融资风险预测中的应用
Model Based on Spirally Evolution Glowworm Swarm Optimization and Back Propagation Neural Network and Its Application in PPP Financing Risk Prediction
计算机科学, 2022, 49(6A): 667-674. https://doi.org/10.11896/jsjkx.210800088
[7] 杨晓宇, 殷康宁, 候少麒, 杜文仪, 殷光强.
基于特征定位与融合的行人重识别算法
Person Re-identification Based on Feature Location and Fusion
计算机科学, 2022, 49(3): 170-178. https://doi.org/10.11896/jsjkx.210100132
[8] 陈贵强, 何军.
自然场景下遥感图像超分辨率重建算法研究
Study on Super-resolution Reconstruction Algorithm of Remote Sensing Images in Natural Scene
计算机科学, 2022, 49(2): 116-122. https://doi.org/10.11896/jsjkx.210700095
[9] 李玉, 段宏岳, 殷昱煜, 高洪皓.
基于区块链的去中心化众包技术综述
Survey of Crowdsourcing Applications in Blockchain Systems
计算机科学, 2021, 48(11): 12-27. https://doi.org/10.11896/jsjkx.210600152
[10] 廉文娟, 赵朵朵, 范修斌, 耿玉年, 范新桐.
基于认证及区块链的CFL_BLP_BC模型
CFL_BLP_BC Model Based on Authentication and Blockchain
计算机科学, 2021, 48(11): 36-45. https://doi.org/10.11896/jsjkx.201000002
[11] 杨青, 张亚文, 朱丽, 吴涛.
基于注意力机制和BiGRU融合的文本情感分析
Text Sentiment Analysis Based on Fusion of Attention Mechanism and BiGRU
计算机科学, 2021, 48(11): 307-311. https://doi.org/10.11896/jsjkx.201000075
[12] 高熙, 孙未未.
集装箱码头岸桥最优调度理论研究和高效算法
Theoretical Research and Efficient Algorithm of Container Terminal Quay Crane Optimal Scheduling
计算机科学, 2021, 48(11A): 22-29. https://doi.org/10.11896/jsjkx.201200167
[13] 代闯闯, 栾海晶, 杨雪莹, 过晓冰, 陆忠华, 牛北方.
区块链技术研究综述
Overview of Blockchain Technology
计算机科学, 2021, 48(11A): 500-508. https://doi.org/10.11896/jsjkx.201200163
[14] 刘彤彤, 杨环, 西永明, 郭建伟, 潘振宽, 黄宝香.
机器学习在脊柱疾病智能诊治中的应用综述
Review on Intelligent Diagnosis of Spine Disease Based on Machine Learning
计算机科学, 2021, 48(11A): 597-607. https://doi.org/10.11896/jsjkx.201100006
[15] 穆聪聪, 王一舒, 袁野, 乔百友, 马玉亮.
时序图中Top-k稠密子图查询算法研究
Top-k Densest Subgraphs Search in Temporal Graphs
计算机科学, 2021, 48(10): 152-159. https://doi.org/10.11896/jsjkx.201100005
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!