《全球6G技术大会:2024年意图驱动自智网络白皮书(英文版)(57页).pdf》由会员分享,可在线阅读,更多相关《全球6G技术大会:2024年意图驱动自智网络白皮书(英文版)(57页).pdf(57页珍藏版)》请在三个皮匠报告上搜索。
1、1/56ContentsI.Introduction.3II.Motivation and Development History of IDAN.42.1 Research Background and Motivation.52.1.1 Challenge 1:Diverse and Changing Service Requirements of Users.52.1.2 Challenge 2:Heterogeneity and Complexity of Intelligent Networks.62.1.3 Challenge 3:Personalized and Customiz
2、ed On-Demand Services 62.2 Development History of IDAN.6III.Conception Meaning andAdvantages of IDAN.83.1 Concept of IDAN.93.2 Techniques of IDAN.93.3 Advantages of IDAN.12IV.Architecture and Experimental Validation of IDAN.134.1 Hierarchical Operation Architecture of IDAN.134.2 Technical Implementa
3、tion Architecture of IDAN.144.3 End-to-End Implementation of IDAN.164.4ExperimentalValidationoftheIntent-DrivenOpticalNetworkArchitecture.174.4.1 Design and Optimization of the Deployment Method.174.4.2 Operation and Maintenance Methods and Policies.204.4.3 Integration Methods Compatible with the Ex
4、isting Network.21V.Key Technologies Involved in the Full-lifecycle of IDAN.245.1 Top-Down Intent Fulfillment.255.1.1 Intent Modeling.255.1.2 Intent Understanding.265.1.3 Policy Verification.285.2 Intent Realization Knowledge Management.30VI.Application Cases of IDAN.326.1 Application Cases-Intent-Dr
5、iven CLL.326.2 Application Cases-China Unicom.366.2.1 Intent-driven Autonomous Orchestration Case.366.2.2 Intent-Driven Autonomous Energy Saving.376.2.3 Automatic Drive Test Intelligent Evaluation Solution.386.3 Application Cases-China Mobile.396.3.1 Intent-driven Hierarchical Service Assurance Case
6、.392/566.3.2 Intent-driven Sensing Certainty Assurance Case.416.3.3 Autonomous Intent O&M Assurance High-quality Industry-specificLeased Lines.436.3.4 Implementation of General Intent Solutions via ONAP.456.4 Application Cases-Emergency Communication Network.476.4.1 Case Analysis of Call on Session
7、Initiation Protocol(SIP)476.4.2 Integrated Space-Air-Ground Communication Network for EmergencyRescue.48VII.Technical andApplication Challenges of IDAN.497.1 Technical Bottleneck Analysis.497.1.1 Uncertainty of Global Information.507.1.2 Variability of Dynamic Intent.507.1.3 Full-Lifecycle Closed Lo
8、op.507.2 Aspects to Explore in the Future.517.2.1 Bottom-Up Intent Assurance.517.2.2 Large Model of Intent-Driven Network.527.2.3 Intent-Driven Cross-Domain Autonomy.53Abbreviations.54References.55Participants.563/56I.IntroductionIn response to the on-demand services in all scenarios,the future netw
9、orks are expected toallow users to enjoy services,change networks and share resources as desired.In the future,allscenarios will coexist in the network,and there will be many global network parameters,diverseresource conflicts,highly complicated network configuration,and mismatched delivery.Therefor
10、e,it is necessary to enable intelligent management and control of the network based on service intent.Intelligent networks are the product of deep integration of networks and artificial intelligence(AI).Related new network concepts include Intent-Based Network(IBN),Intent-Driven Network(IDN),Autonom
11、ous Driving Network(ADN)and Autonomous Network(AN),etc.The mainpromoters include standards organizations such as IETF,ONF,TMForum,and CCSA,ICTequipment manufacturers such as Cisco and Huawei,and telecommunication operators such asChina Mobile and China Unicom.This white paper believes that the IDN i
12、s the key technology torealize fully advanced autonomous networks in the future.The intent-driven autonomous network(IDAN)is therefore defined herein,aiming to explore the concepts,architectures,technologies,and use cases,and summarize future work.IDAN has become a consensus in the industry.Various
13、standards organizations,industryorganizations,open source communities,and other players in the communication industry areactively promoting the industrial application layout of autonomous networks,covering all fields ofthe industry,such as communication operators and communication equipment manufact
14、urers.Asthe technical research and development and commercial implementation of IDAN are acceleratingpace,the industry is booming.Through top-down forward intent fulfillment and bottom-up reverse intent assurance,IDANestablishes closed-loop life cycle management of a space-divided(from user space to
15、 digital spaceto physical space)and multi-layer IDN.Based on the zero trust principle,a step-by-step feedbackmethod is adopted to continuously monitor the entire cycle for dynamic optimization.By building automated,intelligent O&M capabilities throughout the network life cycle,IDAN provides customer
16、s with the ultimate service experience of zero wait,zero failure,andzerocontact,andcreateefficientO&Mfeaturingself-configuration,self-fix,andself-optimization for the production front line of the network.By deeply integrating AI with thehardware,software,and systems of the communication network,IDAN
17、 facilitates agile innovationof enablement services,intelligent network operation,and building of intelligence-endogenousnetworks.Building an intent-driven autonomous network is a complex system project that requirescomplete top-level design,and unified definition of standards and technical processe
18、s.This whitepaper analyzes the development trends of IDAN in a comprehensive manner,sorts out theevolution trends of key technologies on IDAN,discusses the industrial application of autonomousnetworks of operators and equipment vendors,and analyzes the technical challenges andengineering problems of
19、 IDAN applications to guide subsequent research and application.4/56II.Motivation and Development History of IDANDriven by the rapid development of the digital economy,the global telecommunicationsindustry is accelerating the transformation and upgrading towards automated,intelligent networks.With t
20、he large-scale deployment and rapid commercialization of communication networks,therapidly developing digital economy has opened the door to the digital era.With the deepintegration of communication networks and AI,autonomous networks can provide consumers andvertical industry users with innovative
21、network services and digital experience 1.Operators have gradually realized the importance of IDAN.China Mobile views theautonomous network as a key trend in improving the quality and efficiency of communicationnetworks and facilitating the digital upgrade of the industry.With technologies such as I
22、DN,ithelps enable automated deployment of sensing and control policies in the network and evolutionto advanced autonomous networks.China Unicom emphasizes that in order to meet the diverseservice needs of users,networks must have the characteristics of load response and servicecustomization in all s
23、cenarios.The introduction of the concept of intent can simplify networkoperations,improve automation level,and meet service demands more effectively.China Telecomproposed the concept of intent-based network at the 2019 SDN/NFV/AI Conference.Driven byintents,IBN works to realize personalized network
24、services and flexible allocation and sharing ofnetwork resources.IBN aims to cover the entire society,industry,and ecosystem,and define anintelligent management and control system for all scenarios and domains.Chinese operators areleading the innovative practice of autonomous networks and promoting
25、a consensus in theindustry across the world.Since 2021 when China Mobile first proposed the goal of reaching L4in 2025,several companies including China Unicom and China Telecom have set relevant goals,aiming to reach L4 level in high-value scenarios or across all networks.Equipment vendors are laun
26、ching intent-driven products and solutions.Huawei has proposedin the industry the concept and grading standard of the autonomous driving network.It has definedgrading standard for autonomous driving in communication networks in terms of serviceexperience,labor-saving level,and complexity of the netw
27、ork environment.According to thisgrading standard,L5,i.e.the fully autonomous network,which is the ultimate development goalof telecommunications networks,is also the development goal of IDAN.ZTE has proposed theautonomous evolving network(AEN)solution.By using ubiquitous AI to promote theintelligen
28、tialization of the entire network,and adopting the hierarchical closed-loop principle tobuild an intelligent network system at the network element level,single-domain level,andcross-domain level,AEN aims to push the network system to gradually realize autonomousoperation,and,as driven by data,enable
29、 self-learning and self-evolution,and the intelligentautonomous network system,so as to optimize the investment efficiency and O&M efficiency ofthe network.ZTE has planned a comprehensive service process covering planning,building,maintenance,optimization,and operation and more than 60 intelligent a
30、pplication scenarios,including single-domain intelligent scenarios in various dedicated domains such as wireless,bearer,and core networks,as well as scenarios of end-to-end digital operation across dedicateddomains.Many universities are exploring breakthroughs in single technical points on autonomou
31、snetworks.Beijing University of Posts and Telecommunications has conducted research onknowledge-defined intent-driven networks.Xidian University is trying to achieve a closed-loop5/56intent life cycle following the two technical paths:top-down intent fulfillment and bottom-upintent assurance.Univers
32、ity of Electronic Science and Technology of China has deeply exploredintent representation and the classification and modeling technologies,and applied theintent-driven concept to emergency communications.In order to seize the opportunity first,enterprises have been constantly making adjustmentsand
33、changes.In order to benefit from the market opportunities brought by key technologies infuture networks,cloud computing,and edge computing,operators must also make active changesand devote themselves in.In the future service scenarios,not only networks with low latency,highreliability,and support fo
34、r massive device connections are required,but operators are alsoexpected to provide additional services,including massive data analysis,image recognition,andprecise positioning.Research activities on intent-driven autonomous networks are popular inuniversities,showing the huge and unstoppable develo
35、pment potential of this field.2.1 Research Background and MotivationThe development of the global digital economy is moving towards a new stage of deepapplication,standardized development,and inclusive sharing.The digital transformation is beingadvanced in various industries,expanding from office an
36、d marketing service scenarios to coreproduction and manufacturing scenarios,shifting from efficiency change to value change,andextending from internal digital transformation within the enterprise to collaborative transformationalong the industrial chain and value chain.Against this backdrop,the need
37、 for 5G applications andcloud-based big data-driven intelligence has been more common and urgent among governmentsand enterprises.In the past ten years,China has achieved remarkable achievements in the development of thedigital economy.In 2022,Chinas digital economy registered a GDP of CNY 50.2 tril
38、lion,with ayear-on-year nominal growth of 10.3%,significantly higher than the nominal GDP growth rate inthe same period for 11 consecutive years.The digital economy accounts for 41.5%of GDP,equivalent to the proportion of the secondary industry in the national economy.With its overallscale ranking s
39、econd in the world,the digital economy in China is increasingly playing a leadingand supporting role in economic and social development.2023 witness the emergence of manynew digital services,such as computing-network integration,Internet of Vehicles,and HDICT,provided by communication operators,as w
40、ell as new business models,such as network as aservice,service customization,and project delivery.All this places higher requirements onnetwork infrastructure and automated intelligent operation.Driven by the rapid development of the digital economy,the global telecommunicationsindustry is accelerat
41、ing the transformation and upgrading towards automated,intelligent networks.Against this backdrop,autonomous networks emerge.In order to meet the diverse digital needs ofvertical industries and consumer lives,the autonomous network framework and its key featuresare proposed.However,the evolution of
42、network technology and the improvement in O&Mmanagement have brought many challenges.2.1.1 Challenge 1:Diverse and Changing Service Requirements of UsersWith the development of science and technology,various industries have increasingly urgentneeds for technological innovation.For example,the develo
43、pment of new energy,AI,big data andother fields has created opportunities for transformation and upgrading across traditional6/56industries.User demands for network services are also changing faster and faster.In pursuit ofpersonalized and customized consumer experiences,users have more diverse dema
44、nds forproducts and services.Operator network services must have the agility to capture user intent,intelligently and flexibly adjust resource allocation and operation policies based on user intent,and improve product and service quality to meet the constantly changing needs of users andenhance user
45、 experience.With the saturation of the individual user market,the digitization ofvertical industries and that of consumer lives have become the highly-expected potential growthpoints,and more diverse needs have been proposed for connectivity,bandwidth,latency,reliability,and other aspects of network
46、 services.2.1.2 Challenge 2:Heterogeneity and Complexity of Intelligent NetworksWith the dramatic increase in the number of devices connected to the network,the networkare constantly expanding.This will bring huge challenges to network management,e.g.how toensure reasonable allocation of network res
47、ources,how to improve network performance.Inaddition,intelligent heterogeneous networks in the future will cover multiple access technologies,multiple application scenarios,and deeply integrate with various industries and fields.This willmake the network architecture more complex and require better
48、network management.How tosimplify network management and improve O&M efficiency will be one of the major challenges.2.1.3 Challenge 3:Personalized and Customized On-Demand ServicesOn-demand services in all scenarios aim to provide users with personalized and customizednetwork services to meet the ne
49、eds of different scenarios.With AI technology,intent-drivenintelligent policies can better sense user needs,enable intelligent scheduling and optimization ofnetwork resources,and thus improve network performance and user experience.The integration ofon-demand services in all scenarios and intelligen
50、t policies can better meet users network needs indifferent scenarios and improve network resource utilization.The autonomous network has anetwork system with autonomous learning,adaptive adjustment and self-optimization.It canautomatically adjust network parameters based on user behavior and network
51、 conditions toreasonably allocate and optimize network resources.The IDN drives the adjustment andoptimization of network behavior with user needs and intent at the core.The integration ofautonomousnetworksandIDNbringsmanyadvantages,includingimprovednetworkperformance,better user experience,greater
52、network security protection,lower management cost,and adaptability to the development requirements in future networks.To sum up,in order to meet diverse service needs of users,the introduction of intent becomesthe key to solving problems in the future development of network load features and service
53、customization in all scenarios,so as to simplify network operations,improve automation level,and better meet service needs.2.2 Development History of IDANAs many domestic and foreign standards organizations,including TM Forum,3GPP,ITU-T,ETSI,and CCSA,have been carrying out the work related to autono
54、mous networks,enterprisesare actively promoting the standardized layout of autonomous networks to compete for a voice inthe industry 5-7.In 2015,the Northbound Interface Working Group of the Open Networking Foundation7/56published the Intent-Based Networking White Paper.Gartner,Cisco,and Huawei have
55、successively proposed the definition of IDN,clarifying the basic components of IDN,such asintent translation,intent verification,automatic implementation,and holographic sensing 710.As a brand-new network model that analyzes user intent,IDN translates the intent intocorresponding network policies,an
56、d ultimately realizes automated deployment of network sensingand control policies,and thus the network is evolved from a static resource system into a dynamicsystem that can meet business goals continuously.In 2015,ONF released a standard proposal titled Intent:Dont Tell Me What to Do!Tell MeWhat Yo
57、u Want:In intent-based networks,intelligent software decides how to translate intent intoconfigurations for specific infrastructure so that the network can operate in the desired manner.In 2017,Cisco proposed in the white paper titled Moving towards Intent-Based Networks:The network team can use con
58、cise words to describe the work they want to accomplish,and thenthe network can translate the intent into numerous policies that will make appropriateconfiguration and setting changes in complex,heterogeneous environments with the automatedfunctions.In 2017,the Network Management Research Group(NMRG
59、)of the International InternetResearch Task Force IRTF launched the intent research,focusing on the analysis of conceptdefinitions,main principles,intent classification,and life cycle management,and released RFC9315 and RFC 9316,which define intent as the abstract,high-level policy for operating net
60、works5.In 2018,the IDMS_MN project initiated by 3GPP SA5 described the concept and scenariosof intent-driven management in terms of the mobile communication network management layer,and recommended standardizing intent expression in the specification stage.3GPP released the TS28.312 standard in R18.
61、In 2024,the FS_IDMS_MN_Ph3 project was launched in R19 to conductfurther research on enhanced intent-driven management services for mobile networks.In 2019,ETSI launched the comprehensive research on intent standards.The first releasedETSI GS ENI 005 defines the ENI system policy management model,wh
62、ich supports declarativepolicies,imperative policies and intent policies.ETSI GR ENI 013,released in 2022,focuses onintent information model standardization,and carries out gap analysis based on the current statusof the industry.ETSI IFA 050,launched at the same year,mainly studies intent-driven int
63、erfacesand model standards.ETSI SOL 021,launched in 2023,mainly studies the RESTful protocol anddata model specifications regarding the requirements on intent management service interface.The autonomous network,jointly proposed by TM Forum and industry partners in 2019,aimsto lead the automated and
64、intelligent transformation of network infrastructure and operationsystems.Aftermorethanfouryearsofdevelopment,systematicconcepts,standards,implementation methods,and application cases of the autonomous network have formed,withremarkable achievements in industry consensus,standard formulation,and pra
65、ctice anddeployment 7.(1)Industryconsensus:Zero-X/Self-Xvision,L0-L5grading,three-layerandfour-closed-looparchitecture,single-domainautonomyandcross-domaincollaboration,intent-driven and full stack AI and many other concepts have become widely-accepted industryconsensus;(2)Standard formulation:9 maj
66、or standards organizations,including TM Forum,3GPP,8/56CCSA,and ETSI,have been focusing on 5 major standard directions,and have initiated/publishedmore than 80 standards/research subjects in total.In addition,they rely on the M-SDO Initiative,involving multiple standards organizations,to ensure homo
67、geneous architectures and unifiedstandards;(3)Practice and deployment:14 leading operators,including China Mobile,China Telecom,China Unicom,Deutsche Telekom,and Vodafone,have included autonomous networks into theirgroup strategies.Driven by commercial value and operational effectiveness,they aim to
68、 conducthierarchical assessment and autonomous capability planning and building in an iterative way withthe goal of reaching L4 autonomy between 2025 and 2027.In 2021,TMF officially launched the work on standards related to cross-layer andcross-domain general intent.In the same year,it delivered IG1
69、253 and a series of intent-relatedspecifications,and conducted in-depth research on intent modeling,intent information models,intent life cycle management,interfaces and intent use cases 4.In 2022,it launched research onthe TR290 general intent model,TR291 extended intent model,and TR292 intent mana
70、gementontology.And it published two research reports on the development of intent in autonomousnetworks:Autonomous Networks:Empowering Digital Transformation-From Strategy toImplementation in 2021 and Intent in Autonomous Networks in 2022 8.The NFV special project group SP1 of CCSA released a series
71、 of standards for technicalrequirements for network functions virtualization(NFV)management and orchestration in 2021,stipulating the overall architecture and functional requirements of intent management,intentmanagement information models and intent management interfaces in the technical requiremen
72、tsfor NFV management and orchestration.The Network Management Technical Committee TC7 ofCCSA launched technical research on intent management in autonomous networks in June 2022,focusing on research and discussion on intent management architectures and related referencepoints,intent expression model
73、s,intent life cycle management,etc.And it officially launched theformulation of standards for technical requirements for intent management in autonomousnetworks in June 2023 9.Duringtheabove-mentionedstandardizationresearchprocess,variousstandardsorganizations and manufacturers have been actively de
74、voted in the research on IDN and proposedtheir solutions.IDNs are gradually developing from theory to practice,and are constantly beingimproved and optimized in practice,showing great potential in network management,automateddeployment,and optimized network resource allocation.However,the current re
75、search andpractice of IDNs are still facing some challenges and problems.For example,how to accuratelyunderstand and translate user intent,how to ensure the accuracy of intent verification andimplementation,and how to enable comprehensive sensing and dynamic adjustment of thenetwork.All these proble
76、ms have to be studied and solved in future research and practice.Inaddition,autonomous networks,as an important application scenario of IDNs,are getting moreand more attention.With the introduction of the concept of intent,autonomous networks enablesautomated,intelligent,and dynamic networks,thereby
77、 better meeting the needs of users andbusiness development.Therefore,how to achieve effective management and application of intentin autonomous networks will be an important subject in the future research.III.Conception Meaning and Advantages of IDAN9/56Based on the above-mentioned existing standard
78、ization research results,there is an urgentneed for a consensus on the definition of intent in autonomous networks,and the definition,architecture and key technologies of IDNs.And the theoretical and technical system of IDNsneeds to be improved and optimized to better promote the development and pro
79、gress of thenetwork.This chapter introduces the basic concepts of IDAN,including the overall framework,core ideas,and key capabilities.3.1 Concept of IDANAs defined in TMF IG1253 4,intent is the formal specification of all expectations,including the requirements,goals and constraints for technical s
80、ystems.The intent owner is thecreator of the intent and is responsible for the management of the intent life cycle.The intentprocessor is the receiver of an intent,and is responsible for intent realization and satisfaction,andthe management of the life cycle of the intent instance.IDAN is an advance
81、d version of IDN.It emphasizes that the network no longer passivelyreceives service instructions,but tries to actively understand the intent of human administrators toanticipate opportunities and act according to the intent as far as possible,and takes into accountboth the accuracy of the intent and
82、 network optimization goals.In addition,it can dynamicallyadjust network parameters to improve network performance according to the network environmentand user needs.According to the autonomous network grading standards,at the L4(advancedautonomous network)stage,IDAN are implemented by introducing i
83、ntent.IDANs meaningcovers the following aspects:With user intent at the core:User intent is the main consideration in the design and operationof network systems.Network resources are automatically adjusted by analyzing and understandinguser needs.It also takes time to provide optimized services.Not
84、all intent is of equal priority,so itshould configure a priority for each intent and reserve that of high priority.Automated and intelligent:With advanced automation and intelligence technologies,such asartificial intelligence and machine learning,network resources can be optimally allocated,andauto
85、matic configuration of intent priority,optimization,and failure recovery can be enabled.Thesetechnologies can learn,sense and make decisions autonomously,thereby improving theadaptability of the network.Zero wait,zero contact,and zero failure:With precise resource control and automated O&Mcapabiliti
86、es,real-time service provisioning,ready-to-use and zero-wait experience can be realized;with end-to-end network monitoring and intelligent configuration failure recovery,zero-failureexperience can be realized;with open network data and capabilities and self-services,zero-contactexperience can be rea
87、lized.Network autonomy in all scenarios:Building systematic capabilities based on the fourmanagement layers of network elements,resources,service,and business,enables networkautonomy in all scenarios.By building systematic capabilities at different layers,collaboration andclosed-loop control between
88、 different management levels can be achieved,which boosts externalbusiness growth and improves internal efficiency.3.2 Techniques of IDANAn autonomous network is a network infrastructure,operation management system andservice system that integrate intelligence and automation 2.At present,the automat
89、ed O&M10/56capabilities of autonomous networks rely largely on preset empirical rules carefully devised byexperts.However,these rule-based automated policies may face many challenges in real-worldapplications.First,the design of rules cannot cover all scenarios,resulting in poor performances inspeci
90、fic or complex environments.Secondly,parameter configuration is usually static,and cannotadapt to instant changes in the network environment,reducing system adaptability and responsespeed.In addition,parameter configurations are often divided into several fixed ranges,andcannot be refined according
91、to actual needs,affecting resource utilization efficiency.Finally,intervention of experts is required to identify and resolve potential conflicts between rules,increasing operational complexity and cost.In response to these problems,autonomous networksneed to use more intelligent and flexible techno
92、logies to adapt to the constantly changing networkenvironment and needs.This is a key technical research direction for implementing IDAN.(1)External input intent and endogenous intentAccording to the way of generation,intent in IDAN can be divided into human-generatedintent and automatically generat
93、ed intent.Human-generated intent is called extrinsic intent,andautomatically generated intent is called intrinsic intent.The key technology mainly involved inextrinsic intent is intent understanding.Intrinsic intent is endogenous intent that is usually evolvedto better accomplish human intent,and th
94、e key technology involved is intent prediction.Autonomous is reflected in the process of intent understanding and prediction.In intentunderstanding,the system must have the ability to analyze and understand user input,which mayinvolve natural language processing,image recognition,speech recognition
95、and other technologies.In intent prediction,the system needs to use data analysis,model training,and other methods topredict the possible next intent of the user or evolve secondary intent based on the users currentintent to complete the current service.These operations enable the system to intellig
96、ently senseand predict user intent,making the network serve users in a better and smarter way.(2)Intent-based prediction technologyIntent prediction technology plays a unique role especially in emergency scenarios.In suddenemergency scenarios such as natural disasters,man-made accidents,or infrastru
97、cture failures,IDAN can automatically and quickly schedule network resources to deal with the incidents.Takethe prevention of and response to sudden forest fires as an example.Emergency disasters usuallyrequire millisecond-level response.However,due to the complicated network integrating air,space,a
98、nd ground,a large number of network configuration operations are required,which placesa huge burden on human operators.Through IDAN,a large number of sensors and satelliteimagery can be heterogeneously accessed with the help of the intent-driven proxy to automaticallycollect warning messages from th
99、e physical network.With efficient intent translation and rapiddecision-making,intent can be automatically and accurately issued to direct relevant personnel totake emergency measures,and related resources across the entire network can be automaticallyscheduled to provide immediate communication and
100、coordination support.In addition,with thehelp of big data analysis and machine learning,autonomous networks can predict the spreaddirection of the fire,so as to predict the possible development trend of the disaster in advance,generate various plans for human administrators,and help take more timely
101、 and effectiveemergency measures.(3)Intent-driven management serviceThe intent-driven management service is a key component to implement the intent-driven11/56autonomous function.In SDN-based implementation,it can be implemented on the SDN gatewayand interact with the intent translation function thr
102、ough the southbound interface to translate theintent in natural language into that using the intent model.The intent-driven controller isresponsible for parsing and delivering the intent to the intent-driven management service,and theproxy is responsible for the specific network configuration.The in
103、tent-driven proxy providesaccess and interconnection support for end links,such as satellite links,self-organizing networks,and cluster networks,through physical network interfaces.(4)Intent priorityNetwork resources are always limited,and it takes time to schedule them.Variouscommunication services
104、 are not of equal priority.Therefore,important intent can be configuredwith higher priority to prioritize key communication services by reserving network resources andother means.The configuration of intent priority plays an important role in resolving intentconflicts in emergency communications.The
105、 intent priority may be affected by the clarity of userintent,context,urgency,system policy,user expectations,task complexity and other factors.Eventhe same intent,if issued by different users at different times,its priority may not be the same.(5)Intent pyramid modelThe first stage is configuration
106、 detail management,represented by SNMP.This stage focuseson the basic monitoring and configuration of underlying network devices.The second stage issystem policy management,represented by PBNM.This stage introduces more complex policiesto make network management more targeted and flexible.The last s
107、tage is service intentmanagement,represented by IDNM.This stage emphasizes the implementation of a higher levelof adaptive network management based on the users service intent.During the evolution of thethree stages,as the abstraction level of the underlying facilities is gradually raised,thecomplic
108、ated technical details are continuously screened out,finally implementing networkmanagement based on the service intent of the highest abstraction level,forming mapping fromintent to policy to final configuration and providing more flexible and intelligent means fornetwork management.The pyramid mod
109、el of“intent-policy-configuration”is shown in Figure 1.12/56Figure 1 Schematic Diagram of the Intent-Policy-Configuration Pyramid Model3.3 Advantages of IDANIDAN shows four highlights.First,it brings greater flexibility,intelligence and efficiency tothe business model,provides users with more innova
110、tive and valuable products and services,offering enterprises new business opportunities and growth space.Second,by capturing data fromthe entire network in real time and leveraging massive data analysis,in-depth insights,predictionsand knowledge extraction,the potential of data flow is released to f
111、acilitate data-based productinnovation.Third,it improves resource scheduling,automation and intelligence,effectivelyreduces labor cost,and enhances the utilization efficiency of terminals,base stations andcomputing centers,thus cutting operating cost for enterprises.Finally,IDAN has excellent riskid
112、entification,prediction and emergency response capabilities,significantly reducing potentialrisks and losses.By using automation and intelligent means to dynamically balance the usageneeds of the entire network,producing remarkable effect on energy saving and emission reduction3.With its unique adva
113、ntages,IDAN is reshaping the traditional network operation model 11.IDAN has remarkable advantages in improving network performance,optimizing resourceutilization,and providing personalized services,which help promote the innovation anddevelopment of network technology,mainly reflected in the follow
114、ing aspects:(1)Improving network resource utilization:By monitoring network state and user needs inreal time,IDAN can intelligently adjust network resource allocation for efficient resourceutilization and waste reduction.(2)Optimizing network performance:IDAN continuously optimizes network performan
115、cebased on advanced algorithms and models.This helps improve network stability and reducelatency and packet loss,thus improving user experience.(3)Automated and intelligent:With autonomous learning capabilities,IDAN can extract13/56valuable information from massive data,and automatically adjust netw
116、ork parameters and policies.This offers more convenient network management and lower manual intervention cost.(4)Personalized services:IDAN can provide users with personalized and customizednetwork services based on user behavior and preference.This helps improve user satisfaction andloyalty,which i
117、s an important direction for business development.(5)Open and scalable:With good openness and scalability,IDAN can be seamlesslyintegrated with other network technologies and applications to adapt to constantly changingnetwork requirements.This provides strong support for network innovation and deve
118、lopment.(6)Coping with complex scenarios:IDAN can cope with diverse network environments andcomplex scenarios,optimizing support for various applications.This helps increase networkadaptability and flexibility.IV.Architecture and Experimental Validation of IDAN4.1 Hierarchical Operation Architecture
119、 of IDANThe deployment practice of multilateral access gateways and autonomous networks centersaround the core idea of single-domain autonomy and cross-domain collaboration to buildsystematic capabilities in layers to achieve network autonomy in all scenarios.Its targetarchitecture includes four lay
120、ers,work element management(dynamic sensing andautomatic optimization of equipment components and operation state,and opening up automatedoperation capabilities),network management(automatically converting the upper-layer serviceand application intent-driven controller into network behavior based on
121、 network management,control and analysis capabilities,and,with the intra-domain closed-loop control capability ofsouthbound sensing/analysis/decision-making/execution,in combination with local intelligence,continuously guaranteeing the SLA commitment of network connections or functions and thusachie
122、ving closed-loop management of single-domain network autonomy),service management(based on the end-to-end process of planning,building,maintenance,and optimization,incombination with the three major capabilities of service collaboration,guarantee,and analysis,enabling cross-vendor and cross-domain s
123、ervice layer autonomy and closed-loop management)and business management(mainly targeting at autonomous network businesses,providingcustomers,ecosystem and partners with business enablement and operation capabilities)frombottom to top,as well as three closed loops,i.e.the resource closed loop(managi
124、ng resources of asingle field,achieving single domain autonomy),business closed loop(business-orientedend-to-end management across fields to achieve cross-domain collaboration)and user closed loop(user and business management,including user information,operation,billing,and customerservice)2.The hie
125、rarchical architecture model reduces the complexity of the overall system.Each layer can evolve independently and operate autonomously,and hide the domainimplementation technology,intra-domain operations and intra-domain functional details from theinterface and its consumer layer,as shown in Figure
126、2.Among them,the A-type reference point islocated between the hierarchical decoupled management architecture layers,which can serve asthe autonomous system northbound to provide the upper layer with the call and management entryfor intent processing capabilities.The K-type reference point is located
127、 between the centralizedautonomous node and the hierarchical decoupled management layer,providing the latter with thecall and management entry of centralized intelligent support capabilities required to achieve intent14/56management.An intent overlay layer can be introduced at each layer of the auto
128、nomous network andinteract between layers to provide access and interconnection support for end links such assatellite links,self-organizing networks,and cluster networks.The multilateral access gatewayprovides access interfaces for heterogeneous network interfaces,and the intent-driven controller i
129、sresponsible for simplification and abstraction.In the intelligence stage of autonomous networkevolution,based on user-specified intent goals and combined with AI technology,the systemprovides closed loop automation withinlayers.Heterogeneous network intent interfaceinteractions between each layer a
130、re controlled through the gateway,so that upper-layer servicecalls can be implemented independent of the lower layer.This enables intent issuance and intentcoverage across layers and domains of the autonomous network,further reducing humanoperations,improving management efficiency,and accelerating t
131、echnical iteration.Figure 2 IDAN Architecture4.2 Technical Implementation Architecture of IDANFigure 3 shows a reference technical implementation architecture of IDAN,which adopts alayered design.According to the functional abstraction level and service logic of each layer,thenetwork is divided into
132、 the infrastructure layer,service management layer and businessmanagement layer from the bottom up.The distributed artificial intelligence(DAI)module spansall layers and provides powerful support for each process.The infrastructure layer,consisting of various hardware facilities,software systems,and
133、 data,can be viewed as the cornerstone of IDAN.This layer senses and manages computing resources,network resources,and storage resources across the network in a unified manner,and canexcellently manage and schedule computing and storage resources based on service needs.WithDAI,the infrastructure lay
134、er expands the depth and dimension of its own information sensing,including resource sensing,performance sensing and failure sensing,providing reliable andcomprehensive decision-making input to the network management layer.In addition,throughanalysis and decision-making at the data source,functions
135、such as real-time uninterrupted service15/56response,intelligent adjustment of equipment energy consumption,and computing network failuresensing and repairingcan beenabled toimprovethe self-response,self-healing,andself-optimization capabilities of the system.Regarded as the brain of IDAN,the networ
136、k management layer is responsible forimplementing system functions.This layer receives the state information and service intent of theinfrastructure layer through northbound and southbound interfaces respectively,and performsanalysis,decision-making and network control,including state awareness,reso
137、urce scheduling,computing power management,service orchestration,and failure analysis and self-healing.Thenetwork management layer adopts the hierarchical progression of single-domain autonomy andcross-domain collaboration,emphasizing automated,intelligent execution of sub-functionalmodules of the s
138、ystem,while supporting adaptive evolution of resource state and user intent.Cross-domain collaboration connects various autonomous domains,solves complicated problemsthrough multi-domain collaboration,and enables the automation of functional processes andhigh-level intelligent closed-loop processing
139、 of services.The network management layer deeplyembeds AI into all levels of the computing network to improve the intelligent learning ability andscenario adaptability of each functional module to ensure the quality of service of the computingnetwork in providing current and new services.Figure 3 De
140、sign Framework of IDANThe service application layer is used to open up user-oriented service capability and bearabstract service functions.From the users perspective,the service application layer can schedulethe service application to appropriate nodes based on user intent,achieving optimal resource
141、utilization and ensuring excellent user experience.The DAI module forms a complete intelligent closed loop through data management,learning training,intelligent distribution and continuous learning,providing comprehensiveintelligent services for the infrastructure layer,network management layer and
142、businessapplication layer.As the intelligent capability management and knowledge unification center,the16/56DAI module is deeply integrated at all levels,and the entire life cycle of AI,including design andtraining,inference verification,deployment and application,and iteration and optimization are
143、alldeployed within IDAN,so that AI is born and served internally.By deeply mining the datagenerated during the operation of the computing network,the DAI module collaborativelyintegrates the different data,resources and functions between different network layers to generateeffective solutions in con
144、sideration of operation efficiency,automation level and service quality,and other aspects of IDAN.At the same time,this module supportsAI in continuous learning,withthe ability to adaptively evolve knowledge and perform knowledge fusion and inference to createnew knowledge.4.3 End-to-End Implementat
145、ion of IDANA complete intent processing process includes the following stages,i.e.the start stage,evaluation stage,issuance stage,and implementation stage,as shown in Figure 4.In the start stage,the intent owner incubates new intent objects by detecting how well therequirement can be met to determin
146、e whether to define a new intent object or change the currentone.If there is no need to create a new intent object,the processing process ends.In the evaluation stage,the intent owner and potential intent processor determine the feasibleimplementation method for the intent object through investigati
147、on and negotiation,including theselection of the intent processor and negotiation on intent object parameters,and verify andevaluate the impact of the corresponding method.If it is deemed not feasible by the evaluation,theprocessing process ends;If deemed feasible,the intent owner will determine the
148、 intent processorand corresponding parameter information required to implement the intent object.In the issuance stage,the intent owner issues the evaluated intent object to the intentprocessor to request a new intent instance.If the intent processor accepts the intent object,the newintent instance
149、is created,and the intent life cycle enters the implementation stage.If the intentprocessor does not accept the intent object,the creation of the new intent instance is failed,and theprocessing process ends.Figure 4 Schematic Diagram of the End-to-End Intent Processing ProcessIn the implementation s
150、tage,the intent processor operates its responsibility domain accordingto the accepted intent to realize the desired goal of the intent,continuously ensures that the17/56corresponding expectations of the intent are met before the intent instance is deleted,and reportsthe intent processing progress,ex
151、ecution status,the reason why the requirement cannot be met,etc.to the intent owner as needed.When the intent processor receives the request to update the intentinstance,it determines the feasible method for the update operation through investigation andnegotiation,and verifies and evaluates the imp
152、act of the corresponding method.If the update isnot feasible,the processing process ends;if the update is feasible,the corresponding intentinstance will be updated.When the intent processor receives the request to delete the intentinstance,the processing process ends.In an IDN,depending on different
153、 layers of business management,service management,resource management,and network element management,the upper-layer intent managementfunction can serve as the intent owner,and the lower-layer intent management function can serveas the intent processor.The top-down intent calling process is implemen
154、ted between layers bycalling the intent interface,so that the users end-to-end intent can be satisfied.4.4 Experimental Validation of the Intent-Driven Optical NetworkArchitectureBased on the IDAN framework,the intent-driven optical network platform(IDONP)isdeveloped,including the optimized design o
155、f deployment solutions from intent to policy,a faultlocation mechanism for high-precision intent assurance,and a fast slice reconstruction method forintent assurance.4.4.1 Design and Optimization of the Deployment MethodThis section explains the requirement analysis of IDAN,and two implementation me
156、thodsfrom intent to policy,including intelligent policy generation based on strong intent constraints(PG-RL)and intelligent slicing policy generation based on intent constraints,for optimaladaptation of intent to policy.As the number of access users and service types increases,the traffic of optical
157、 networkschanges in stages and regionally.However,the cost of building and demolishing routes is veryhigh,and it is difficult to ensure real-time building and demolition.Therefore,it is necessary forthe network to be able to appropriately and automatically change on the existing networkinfrastructur
158、e,which can be done through network automation or network orchestration,and isrequired to:1)automatically generate configuration policies under strong intent constraints tomeet the service requirements of the optical network;2)establish a highly robust intent assurancemechanism;3)realize real-time a
159、daptation of intelligent policies to the optical networkenvironment.Technical principles:The mapping of intent to slicing policy implementation usually requiresconfiguring the connection topology(connecting to service nodes),exchanging data to updatecontent,and allocating sufficient computing,storag
160、e,and transmission resources to maintain acertain QoS level.The optical network slicing can adapt to the flexible requirements of diverseservices and provide customized service guarantees.Network slicing allows infrastructureproviders(InP)to support heterogeneous services on a public platform(i.e.cr
161、eate a customizedslice for each service).And the slice can be dynamically scaled up/down to match any change inthe requirements of its services.The slice is created based on the requirements for the intenttranslation result(e.g.latency,capacity,and reliability).In the case where there is a change in
162、 time18/56and/or space in the traffic aggregated in the slice,the InP can dynamically scale up/down theprovisioned slices to match the change in service requirements,thereby improving its resourceutilization efficiency.In addition,benefiting from the development of network functionsvirtualization(NF
163、V)and software-defined networking(SDN)technologies,network slicing hasbeen proposed as a key architecture technology in the intent-driven integrated optical-wirelessnetwork.DRL learns from experience by interacting with the network,constantly searches for andadjusts appropriate policies,and dynamica
164、lly adjusts the resources allocated to each slice,therebymaximizing resource utilization while guaranteeing the intent constraints.Deployment method:The intent-driven integrated optical-wireless network architecturemainly includes intent expression,translation,verification and deployment.In terms of
165、 intent formtranslation,the intent flow of natural language,intent primitive,executable policy,and reliableconfiguration is formed.In terms of hierarchical relationship,combined with the latest practicesof Open Daylight(ODL)and Open Network Operating System(ONOS),the intent-drivenintegrated optical-
166、wireless network mainly includes the service application layer,intentnorthbound interface(NBI),intent policy layer,intent assurance layer and infrastructure layer,asshown in Figure 5.Figure 5 Intent-Driven Integrated Optical-Wireless Network ArchitectureThe service application layer generates servic
167、e intent,including different services in differentscenarios.Service intent can be generated directly or indirectly.The service application layerprograms the underlying devices through the programming interface provided by the intent policylayer,thus abstracting the functions of network elements.In a
168、ddition,this layer providesmanagement interfaces to achieve diverse service innovation.The intent northbound interface,a module used to translate intent,connects the serviceapplication layer and the intent policy layer.In addition,the southbound interface(SBI)is basedon the virtualization technology
169、 and connected to various network element devices,virtualizingvarious computing and communication resources.It is mainly used for interaction between the19/56infrastructure layer and the intent policy layer.The intent policy layer is the core of the architecture with management control anddecision-m
170、aking capabilities.This layer parses and checks the service intent translated through theintent northbound interface.User intent is processed into a standardized intent request that can beexecuted by the current network,and the specific resource requirements in the network areobtained through a mapp
171、ing algorithm between the intent and resources.This layer uses anintent-based management and orchestration system for unified scheduling of resources and sliceconfiguration,and introduces the closed-loop configuration into the life cycle management ofnetwork elements.With the intelligent engine,it c
172、ompletes network state data collection,datastorage,data processing,model training,parameter adjustment and other tasks.The infrastructurelayer includes various physical device entities.It also deploys a large number of network datacollection tools to provide feedback information and parameters requi
173、red for policy configuration.Optimized Design:(1)Intelligent policy generation based on strong intent constraints(PG-RL)By studying the extraction of intent keywords and designing intent request messages toencapsulate the intent keywords,the intent can be expressed accurately.Under the constraints o
174、fintent request messages,it studies policy generation based on reinforcement learning(RL),and,bycombining granular policies to generate new configuration methods for optical networks,optimaladaptation of intent to policy can be achieved.Figure 6 Architecture of the Knowledge Graph Matching System fo
175、r Intent Translation(2)Intelligent slicing policy generation based on intent constraints(SPG-RL)By studying the mapping of intent and network resource requirements,it captures the serviceintent and translate it into network policies,achieving accurate translation of intent into networkstate requirem
176、ents.Under the constraints of intent,it studies slicing policy generation based onreinforcement learning(DRL),and,by effectively translating user performance requirements intoslice configuration policies for the integrated optical-wireless network,multidimensional sensingrequirements and optimal ada
177、ptation of intent to policy can be achieved,thus overcoming theshortcomings of traditional optical network resource allocation policiesbeing static and20/56solidified.Figure 7 Intelligent Slicing Workflow of the Intent-Driven Integrated NetworkFigure 8 Intent Parsing Execution and Slicing Policy Res
178、ult4.4.2 Operation and Maintenance Methods and PoliciesDeploy the network,and set the algorithms and computing nodes based on the intent-drivenoptical network platform(IDONP)architecture.With intent assurance in mind,the maintenancepart studies the high-precision fault location algorithm and fast sl
179、ice reconstruction algorithm forintent assurance:(1)To study the intent assurance mechanism based on high-precision fault location methods:by introducing the deep neural evolution network,a high-precision fault location method forlarge-scale alarm sets is proposed,which effectively enables the preci
180、se location of fault nodes inoptical networks,and helps the policy layer isolate faults in time to best assurance intent.Thesimulation results show that if there are more than 10,000 alarms,the location accuracy isincreased to 92%and the location time is reduced by 0.5 seconds thanks to the embedded
181、21/56deployment of the federated deep neural evolution network(FL-DNEN).Figure 9 Improved Fault Propagation ModelAlgorithm(2)The fast slice reconstruction algorithm for intent guarantee(SR-DDQN)studies the intentassurance mechanism based on the fast slice reconstruction method.By introducing the Due
182、lingDQN network,a fast slice reconstruction method for highly dynamic network environments isproposed,which effectively adjusts the policy of allocation of computing and storage resources indifferent domains and the deployment of VNF to match the dynamic changes in slicing requests,and thus to best
183、assurance intent.The proposed model can continuously detect the consistencybetween the execution policy and the original intent based on the real-time network state,andadopt fast slice reconstruction when the intent requirements are not met.The reconstruction timeis reduced by 39.2%,and the resource
184、 utilization is increased by 17.6%,lowering the intentconstraint violation rate while improving network resource utilization.By combining the above two algorithms and the structure of the existing network,thenetwork can realize automatic operation and maintenance.In addition,both algorithms have the
185、continuous learning capability.With the deployment and operation of the network,thisautonomous learning ability plus human supervision and regulation can help further optimize thenetwork.4.4.3 Integration Methods Compatible with the Existing NetworkThis section briefly introduces the existing networ
186、k structure,and how to integrate the IDNand develop IDONP based on the existing network,describes the structure of IDONP and thedevelopment process of each layer,as shown in Figure 10,and clarifies the workflow of IDONP.All this offers integration ideas for IDNs.(1)Introduction to the existing netwo
187、rk:The software-defined optical network(SDON)is anew network paradigm that enables flexible and efficient network management.SDONtechnology allows the underlying infrastructure to be abstracted and used as a virtual entity byapplications and network services.This allows network operators to define a
188、nd operate the logicalmapping of the network,and create multiple coexisting network slices(virtual networks)that areindependent of the underlying transmission technology and network protocols.SDON provides aunified control plane platform to realize the integration of the access network,optical netwo
189、rk,metropolitan area network and core network segments,as well as the integration within andbetween data centers.The global network view can be used to make optimal network controldecisions.SDON offers four main features,such as programmability,agility,flexibility and22/56independence to the network
190、 management domain.The idea of network functions virtualization(NFV)is to virtualize network resources on shared facilities into virtualized network functions(VNF).NFV allocates the capabilities of virtual machines(VM)or containers through a commonshared infrastructure with a VM manager,and directly
191、 controls hardware resources.TheSDON/NFV joint framework is the best way to realize intelligent control and orchestration ofintent-driven integrated optical-wireless networks.The SDN-based open network operatingsystem(ONOS)includes the intent framework component,which is designed to provide theovera
192、ll runtime environment and framework.As a subsystem of ONOS,the Intent Framework isan integral part of intent connection.The Intent Framework views the service intent as apolicy-based directive,allowing applications to broadcast their network requirements externallybased on policy and management.(2)
193、Combined with the integration methods of the existing network:Based on the intent andrequirements,design the network architecture,including network topology,device layout,andconnectivity.Adopt a layered structure and modular design to facilitate future expansion andoptimization.The development of ID
194、ONP has the following practical significance:Providing users with an interactive tool corresponding to the interaction layer in IDONP.Users can issue intent,manage policies in the network,and obtain the network operatingconditions in real time through rendered web interfaces,APIs,or command line too
195、ls.Providing a data model corresponding to the intent northbound interface(Intent NBI)inIDONP.This model describes the detailed semantics behind the natural language and canstandardize the expression of semantics.Moreover,this data model is hierarchical,and ahigh-level intent can be decomposed into
196、several low-level intents.Providing a policy engine corresponding to the JBoss Drools policy engine library in IDONP.The policy engine separates granular network policies from applications,and can manage thesepolicies,including adding,deleting,modifying,and persisting them.This facilitates the polic
197、ygeneration model based on reinforcement learning(PG-RL)to extract granular policies andcombine them to generate new network policies.IDONP can translate user intent into deployableoptical network policies to provide customized services that match user intent.And it integrates anintent assurance mec
198、hanism based on the fault location method to form closed-loop control of theintent life cycle,mainly including the interaction layer,intent layer,policy layer and fault locationlayer.23/56Figure 10 Design of the IDONP StructureThe interaction layer performs rendering mainly based on the HTML5 Canvas
199、.It supportsinfinite scaling of vector graphics,clearly and smoothly displays the operating conditions of theoptical network topology and online visualizes the returned back-end network data.Theinteraction layer and the intent layer can,based on the HTTP protocol,issue the user intent to theintent l
200、ayer through Post requests,and obtain the intent translation result,policy generationsituation,and real-time network operating conditions through Get requests.In addition,theinteraction layer can also manage the policies in the JBoss Drools engine library of the policylayer through APIs.The intent l
201、ayer receives the intent request issued by the interaction layer,parses the intentsemantics through the IA-LDA module,and obtains the intent keyword information.For example,in the intent It is necessary to ensure that sports events can be watched smoothly on the clientsduring the World Cup,QoE,to im
202、prove business priority,and to increase bandwidth allocationand other keyword information can be extracted.The intent NBI then encapsulates the keywordsinto a json data structure containing the service information and issues it to the policy layer.The policy layer includes the JBoss Drools policy en
203、gine library and the policy generationmodel based on reinforcement learning.The policy engine library stores several granular rules,mainly including,in IDONP,service sorting rules,routing rules,and optical network resourceallocation rules.Under the constraints of the data model issued by the intent
204、NBI,PG-RLcombines the granular policies in the policy engine library to generate a new policy that matchesthe intent.In the fault location layer,FL-DNEN includes the fault propagation model and the DNENmodel.By monitoring the network and receiving alarms reported by the controller,it canaccurately l
205、ocate the fault in the optical network and report it to the policy layer in a timelymanner,which helps the policy layer adjust service routing in a timely manner to avoid faulty24/56links or nodes and ensure effective intent assurance.With all the above three aspects in mind,IDAN can be better desig
206、ned,deployed,andoperated to offer more efficient,intelligent,and flexible network services.V.Key Technologies Involved in the Full-lifecycle of IDANThe core goal of IDAN is to translate the structured intent expressed by users into a set ofgranular policies that can be executed by the network.Differ
207、ent from the traditional frameworkthat only views natural language as the source of intent,this technical white paper expands thedefinition of intent to include various expression forms,such as action and voice,to capture theuser intent from more dimensions.Figure 11 Intent Translation ProcessNamed-
208、entity recognition,speech recognition,sequence annotation and other technologiesare used to accurately identify and extract key entities in the intent expression.Then,withtechnologies such as text classification,semantic representation,and text generation,these entitiesare mapped to predefined inten
209、t templates,thereby enriching and improving the intent expressionin the knowledge base.Afterwards,by using the policy library embedded in the knowledge base,the enriched intent templates are translated into a set of specific sub-policies to form the policy setrequired for execution.Ultimately,the co
210、ntrol layer will execute each sub-policy in this policy set in order,so as toaccurately implement the user intent.The overall process and structure of the intent translationtechnology can be understood and implemented with reference to the technical flow chart shown25/56in Figure 11 to ensure effici
211、ent and accurate translation of user intent.To sum up,this chapter willprovide a detailed introduction to the key technologies involved in the above-mentioned spacesand loops of IDNs.5.1 Top-Down Intent Fulfillment5.1.1 Intent ModelingTo achieve accurate mapping from user input to underlying network
212、 configuration,intentneeds to be classified.Human intent should not be classified from a certain perspective,as it ismulti-dimensional with different dimensions overlapping with each other.We can think ofdescribing an intent from multi-dimensions as tagging a certain intent.If there are sufficient t
213、agsto fully cover the dimensions,the intent can be clearly described.In terms of information modeling,the intent can be composed of a set of expectations,andsuch expectations can be based on a common model 3.When specifically defining theexpectation state list,the intent can be described based on th
214、e domain information model.For eachintent,there may be multiple dimensions of intent expectations and corresponding contexts,andfor each expectation,it may be composed of several expectation goals and a set of correspondingexpectation contexts.Intermsofdatamodeling,cross-systemintentexpressionneedst
215、ouseamachine-recognizable language without ambiguous words for expression and exchange.Forexample,the unified modeling language(UML)is a standard modeling language for softwaredevelopment.It allows developers to describe,build,visualize and document the organizationalstructure,behavior and interacti
216、on of software systems.The resource description framework(RDF)is essentially a data model,and formally expressed as the subject-predicate-object(SPO)triple.Itis sometimes also called statements.YANG is a data modeling language with a tree structureconsisting of countless leaves,lists,leaf-lists,and
217、containers.Figure 12 Intent Classification ModelAccording to user goals and intent application scenarios,intent can be classified intocustomer service intent,network service intent,network intent,operational task intent,policy26/56intent,etc.According to whether the name of the target object is expl
218、icitly specified,intent can beclassified into explicit intent and implicit intent.If the target component is clear,it is explicitintent;otherwise,it is implicit intent.According to whether it has life cycle managementcapabilities,intent can be classified into transient(operational)intent and persist
219、ent(service)intent.Transient intent is a simple abstraction of network management operations.Once thespecified operation is executed,the intent is completed,and will no longer affect the target object,so no life cycle management is required.On the contrary,for persistent intent,life cyclemanagement
220、is required.Once the intent is activated and deployed,the system will keep all relatedintent active until they are deactivated or removed.In addition,IRTF RFC 9316 5 proposed for the first time a standard model for intentclassification,which implements policies based on service scenarios,such as bea
221、rer networks,cloud data center networks,and enterprise private networks,as well as various types of users,suchas network administrators and end customers,and can continuously and iteratively expand thescale of the classification model as technology and services evolve.On the one hand,it improvesthe
222、intent recognition efficiency of the system.On the other hand,it provides differentiatedsolutions based on user needs,expectations and priorities.In order to describe it more clearly,intent can also be classified in terms of time,space,whether it is periodic,etc.Figure 12 shows the model for multi-d
223、imensional description of intent.In terms of the level of intent,intent can be classified into basic intent and advanced intent.Interms of time,intent can be classified into past intent,current intent,and possible future intent(intent prediction).In terms of space,complex large-scope intent can be r
224、efined into small-scopeintent and refined from a global network configuration to a local network configuration.Intent canalso be classified into periodic intent and non-periodic intent.In terms of user service,intent canbe classified into the following types:information query,inter-node communicatio
225、n,networkconfiguration,and device control.In terms of urgency and importance,intent can be classified intothe following types:urgent and important,urgent but not important,important but not urgent,neither urgent nor important,etc.5.1.2 Intent UnderstandingIn IDNs,intent understanding is a key compon
226、ent that aims to translate human naturallanguage intents into structured intents.The basic process of intent understanding typicallyinvolves the following steps:1.User intent input:Users express their service intents through natural language(text orvoice)or graphical interfaces.Intents can range fro
227、m simple commands to complex businesspolicy expressions.2.Intent analysis:The natural language input is processed to extract keywords and phrases tounderstand the specific requirements of user intents.This involves decoding the structure andmeaning of the input through syntax analysis and semantic u
228、nderstanding.3.Semantic understanding and mapping:The extracted user language intents are mapped to aset of predefined network operations and policies.Domain knowledge bases are used to link userintents with specific operations that the network system can perform.4.Intent verification and correction
229、:The system checks whether the extracted intents alignwith network policies and business logic.If necessary,the system will ask users to clarify or27/56correct their intent input.Intent understanding is a critical part of an IDN because it is directly related to whether thenetwork can be correctly c
230、onfigured and managed according to the users business goals andrequirements.Therefore,in order to improve the accuracy of intent understanding and theadaptability of the network,it is usually necessary to combine advanced NLP technology,artificialintelligence,and professional knowledge in the networ
231、k field.(1)Extraction model:GlobalPointer model constructionGlobalPointer is a span-based decoding method that considers the head and tail of an entityas a whole,giving it a more global point.It also ensures that training,prediction,and onlineevaluation are all performed at the entity level,as shown
232、 in Figure 13.Figure 13 Intent Entity Extraction Model(2)Dataset construction:generation of over 1,000 datasets(3)Entity categories:Entities can be transformed into intent models defined by variousstandards organizations.Ultimately,these intents will be mapped to a set of predefined networkoperation
233、s and policies.The information after mapping can be described as the following entities:Objects(constraints on objects to which links apply):nodes,locations,regionsBehavior(constraints on how links are configured):connection,networking,firewall,filtering,balanceEffect(constraints on link configurati
234、on effect):bandwidth,latency,throughput,packet lossrate,QoS,business,network characteristics,protocol typeTime(constraints on link action time):action timeAdditional information(additional information on link configurations):port,priority,application scenarioSerialNumberNetworkOperationEntityCategor
235、yNetworkOperation EntitySub-categoryExampleA1ObjectNodeHost1,A1,host,node,User BA2ObjectLocationPointA,Point BA3ObjectRegionSubnet 1,ad hoc network,UAV network,satellitenetworkB1BehaviorConnectionEstablishconnections,enableconnections,interconnect,communicateB2BehaviorNetworkingBuild subnets,form LA
236、Ns28/56B3BehaviorFirewallEnable firewalls,disable firewallsB4BehaviorFilteringFilter blacklistsB5BehaviorBalanceEnable load balancing,enable controller balancingC1EffectBandwidthBandwidth reaches 500 Mbps,bandwidth not lessthan 500 MbpsC2EffectLatencyLatency of no more than 5 msC3EffectThroughputMax
237、imum network throughputC4EffectPacket loss ratePacket loss rate less than 5%C5EffectQosHighest QoS guarantee,QoS enabledC6EffectBusinessHD video conference,voice connectionC7EffectNetworkcharacteristicsSelf-organizing,self-recovery,highlydestruction-resistantC8EffectProtocol typeIP mode,HTTP,SMTPD1D
238、ateAction time8 a.m.to 8 p.m.every day for one weekE1AdditionalinformationPortPort 439E2AdditionalinformationPriorityHighest priorityE3AdditionalinformationApplicationscenariosSmart cities,wireless image transmission,interimconference,environment monitoring,emergency anddisaster relief5.1.3 Policy V
239、erificationPolicy verification is responsible for ensuring the accuracy and applicability of policytranslation and generation to avoid potential risks such as network failures or service qualitydegradation resulting from its direct implementation.This process requires policy verification tonot only
240、be efficient in evaluating and filtering policies within the required timeframe but alsomaintain a high level of accuracy to ensure that the selected policy can achieve the expectedresults in practical applications.When verifying policy enforceability,we mainly focus on keyfactors such as resource a
241、vailability,potential policy conflicts,and the correctness of the policyitself.At present,mainstream policy verification technologies typically involve two steps:buildinga digital twin and conducting policy simulation and validation.These methods together form atechnical support system to ensure the
242、 reliability and effectiveness of the policy.(1)Building a digital twinBuilding a digital twin involves accurately simulating the physical network,which is aprocess facilitated by advanced digital twin technology.This process is divided into two parts:basic model construction and functional model co
243、nstruction.It aims to achieve a comprehensiveand dynamic description of the physical network.Basic models are created based on key data suchas the basic configuration of network elements,environmental information,operating status,andlink topology.This allows the creation of network element and topol
244、ogy models corresponding tothe physical entity network to ensure a real-time and accurate reflection of the physical networkenvironment.This process is further refined into three key steps:The first step is to build theontology model of the twin network to establish a unified and comprehensive netwo
245、rk twin29/56database.Then,build the network element and topology models on demand to achieve an accuratemapping of the physical network.On the other hand,the construction of functional models focuses on meeting the requirementsfor actual network functions.It supports the dynamic evolution and infere
246、ntial decision-making ofthe network by introducing diverse functional modules throughout the entire lifecycle.Thesefunctional models can achieve comprehensive simulation and prediction of network functionsthrough multi-dimensional construction and expansion according to the specific requirements ofd
247、ifferent network applications.The application of this technology not only enhances the flexibilityand responsiveness of network management,but also provides powerful support for networkplanning and optimization.Figure 14 illustrates a schematic diagram of the digital twin ofFunction C in the example
248、.Figure 14 Schematic Diagram of Digital Twin in Policy Verification(2)Conducting policy simulation and verificationIn order to improve the efficiency of policy verification and ensure the timeliness of policies,a critical step is placed in front of the policy verification process:the pre-classificat
249、ion andorganization of the policies to be validated.This process involves using knowledge inductionmethods to group the network policies based on shared characteristics(such as network protocols,initial port numbers,and priorities),aiming to reduce the workload during the verification processand acc
250、elerate the overall verification progress.As the digital twin is successfully built,the network policies to be validated will be injectedinto the twin for observation and analysis of simulation results.Policy verification typicallyemploys two main methods:control plane verification and data plane ve
251、rification.These methodsrespectively simulate the routing protocols at the network control layer and the data transmissionenvironment of the network to verify the reachability of routing policies and the efficiency ofnetwork packet transmission.Specific verification approaches can be further categor
252、ized into two types:global simulationand local formal verification.Global simulation involves simulating the operations of all protocols30/56within a wide area network(WAN)to generate corresponding data plane processing rules.On theother hand,local formal verification comes into play when uncertaint
253、ies are encountered.Itencodes all potential scenarios with logical expressions and uses solvers to calculate variouspossible situations.This allows for an accurate assessment of the effectiveness and potentialimpacts of policies.This method provides a comprehensive and flexible technical path forval
254、idating network policies,ensuring that the implementation of network policies is both efficientand reliable.5.2 Intent Realization Knowledge ManagementIntent realization knowledge is used to standardize and associate user intent storage andinteraction update,intent translation result parameter stora
255、ge,retrieval,and update,as well asnetwork state feedback information storage.Based on this knowledge,it is possible to achieveclosed-loop management of user intent and network state monitoring,providing feedback on theclosed-loop sensing of user intent.Meanwhile,intent realization knowledge manageme
256、nt providesa function for aggregating and exporting masking user information such as interaction informationon original user intent,intent translation result information,and network state feedbackinformation.It also establishes standard data specifications,which can serve as a foundationaldataset fo
257、r the training and application of intelligent algorithm models related to intent networks inthe future.For the intent realization knowledge management model,an available resource libraryneeds to be designed to store intent realization knowledge and related information configurations,and to handle in
258、tent translation results from the user interaction interface.A data collection andanalysis module needs be designed to query user intent processing results in the available resourcelibrary and manage potential intent updates.Finally,a service orchestration module needs to bedesigned to manage intent
259、 use cases,scenarios,and resources.Intent realization knowledge management provides CRUD of intent instances.During thecreation of an intent instance,the user interaction interface reads the use case service ID,binds theuser intent to the newly created intent instance ID and the user service ID,and
260、stores them in theavailable resource library.Meanwhile,the available resource library offers a user intentmonitoring interface for the data collection and analysis module to listen for possible updates touser intents.During the intent modification and update process,the user interaction interface re
261、adsthe new use case service ID,stores the new use case service ID for user intents in the availableresource library,and associates the ID with the existing intent instance.Intent realizationknowledge management also provides intent query and intent deletion functions.By using intent realization know
262、ledge management technology and mapping intent instancesto data tables(such as using an intent instance ID as the key ID),it becomes possible to associateintent-related information.This allows for:(1)association of use case service IDs with intentinstance IDs in different scenarios to meet users req
263、uirements across multiple scenarios;(2)association of continuously changing user intents and intent translation information with a unifiedintent instance;(3)association of network change information and intent assurance informationwith a unified intent instance to assess and optimize the networks ab
264、ility to support user intents.Intent realization knowledge management technology also offers a function for aggregatingand exporting masking data,enabling the export of network intent-related information as afoundational dataset for the training and application of intelligent algorithm models relate
265、d to31/56intent networks in the future.32/56VI.Application Cases of IDANThe vision of intent-driven autonomous networks has become a consensus within the industry.Various sectors such as telecommunications operators,equipment manufacturers,and highereducation institutions are actively promoting the
266、practical application of intelligent networks.Theprocess of technology research and development and commercial application is accelerating,andthe entire industry is showing a trend of prosperity and development.6.1 Application Cases-Intent-Driven CLLChina Telecom adopts the service-oriented concept
267、of ONAP,focusing on 6G networkarchitecture design to address on-demand service requirements in all scenarios.China Telecom hasproposed and implemented an intent-driven closed-loop autonomous network architecture solutionand test platform and has validated key scenarios and technologies.In order to p
268、rovide users withdifferentiated and personalized intelligent services,ensure flexible scheduling of global cloudnetwork computing resources,and enable on-demand services in all business scenarios targetingvarious intelligent use cases,China Telecom has implemented end-to-end integrated dynamicorches
269、tration management for 6G access networks,bearer networks,core networks,space-basedand air-based network domains,and cloud computing resource domains.In this example,userintents from the service management layer are mapped to network execution policies at theresource management layer,realizing inten
270、t-driven cloud leased line delivery and assurance.Figure 15 On-demand Services in All Business Scenarios for Intent-driven Closed-loopAutonomous NetworksFrom the perspective of closed-loop autonomous operations,the solution has designed a dualclosed-loop autonomous network architecture,as shown in F
271、igure 15.The outer closed loopincludes two types of data flows,which respectively implement the functions of creating and33/56issuing intents and the functions of modifying and fulfilling intents.The inner closed loopstructure guarantees user intents and consists of four stages:M2A(monitoring),A2D(a
272、nalysis),D2E(decision),and E2M(execution).The tasks of each stage are defined and implemented bymodules in the solution architecture(such as the intent realization module,policy module,andservice orchestration module),and are specifically deployed and implemented in the ONAPplatform.This solution us
273、es a lightweight decoupling method to achieve module decoupling,andadopts standardized network intent verification processes and interfaces,demonstrating highfeasibility for implementation.The intent input and translation module in the intent autonomous network is implementedbased on UUI.This module
274、 is responsible for parsing natural language input from users andextracting corresponding network intent requirement information.The natural language input fromusers can include voice-based or text-based network intent requirement information.Networkintent requirement information includes network qu
275、ality of service requirements and basicnetwork configuration information.Network service quality requirements involve networkconfiguration parameters obtained through user intent awareness.Basic network configurationinformation includes user location information.Network state information covers real
276、-time traffic,latency,and jitter.The voice-based or text-based network intent requirements from users canundergo natural language processing methods(such as the BERT algorithm)for tasks like entityextraction and recognition.Upon natural language processing,user requirements expressed innatural langu
277、age are translated into corresponding network parameters.SDNC and SO areresponsible for controlling SDN,fulfilling the intent,and then providing feedback to DCAE.These policies are verified through the intent verification module to determine whether thenetwork state complies with the service level a
278、greement(SLA).Finally,DCAE sends thedetermination result to UUI for decision-making by users.The intent instance management technology in the intent network enables closed-loopmanagement of user intents and monitoring of network state.When DCAE detects a change in thenetwork state,it sends a notific
279、ation to UUI,requesting user intervention.Users can then modifythe intent instance parameters stored in AAI through UUI.Subsequently,AAI sends amodification notification to DCAE to update the storage of user intents and intent translationresult parameters and read and update the storage of network s
280、tate feedback information.The Natural Language Processing(NLP)functionality in UUI categorizes user intents intomatching use cases or services,extracts parameters from the text,and populates them intoautomatically generated service request forms.These forms are then presented to users forconfirmatio
281、n or modification.A service form is a standard formatted expression of user intents,which specifies the service type,requirements and SLA.The intent orchestration management component in UUI reads the user intent parameters andtranslates them into the corresponding IETF customer service model,which
282、is then sent to SO.The customer service model is a declarative model that describes the services provided by anetwork operator to customers.SO translates the customer service model into the corresponding IETF service delivery modelin the transmission network.The service delivery model defines how se
283、rvices are designed in thenetwork.SDNC converts the service delivery model into the corresponding network configurationmodel and applies it to the physical network.34/56Such model-driven design has its advantages.First,the model used in the translation step canbe standardized through SDO,which helps
284、 standardize IBN solutions.Second,the modelinstances in each step are stored in the AAI database and can be retrieved through a RESTfulinterface.Additionally,model-driven design decouples the data model from the code logic foraccessing(reading/writing)data.This decoupling enables us to improve and e
285、volve our code logicindependent of the data model,thereby minimizing impacts on other components.The inner closed loop of intent guarantee consists of 4 stages:monitoring,analysis,decision-making,and execution.The monitoring stage is implemented on SDNC.It collectsmonitoring and performance data fro
286、m the network controller and forwards the data to DCAE.SDNC determines which data to collect based on SLA parameters obtained from user intentsduring the intent translation process.In the analysis stage,DCAE analyzes the monitoring data received from SDNC and providesfeedback to policy.Its significa
287、nce is to detect network anomalies and notify policy to takecorrective measures.The decision-making stage is implemented on policy.It makes closed-loop decisions basedon the data received from DCAE and issues appropriate recommendations to SO to performservice changes.The execution stage involves a
288、typical SDN orchestration and control workflow.SO andSDNC apply new network configurations to the physical network.All data,including monitoring data,business models,and network resources andconfigurations,are stored in AAI for retrieval and sharing at various stages.The Cloud Leased Line(CLL)servic
289、e connects cloud service users to edge clouds or clouddata centers and connects edge nodes to each other using connection technology orconnection-oriented technology to provide deterministic connectivity performance.Figure 16shows an example of CLL service architecture using Ethernet-over-OTN(EOO)te
290、chnology.Userdata(such as IP network packets or Ethernet frames)is transmitted through bearer Ethernetservices,which run on Ethernet Virtual Connection(EVC).EVC provides the operations,administration,and maintenance(OAM)of the services,while the underlying optical transportnetwork(OTN)service provid
291、es service isolation and traffic protection.Figure 16 EOO-Based CLL Service ArchitectureFigure 17 shows the actual deployment of the IDN architecture.The physical network is adual-domain optical network,with each network domain controlled by a physical networkcontroller(PNC).The standard interface b
292、etween ONAP and PNC complies with the IETF/ACTNstandard(multi-domain service model and protocol interface).35/56Figure 17 Cloud Leased Line Deployment DiagramFigure 18 illustrates the model-driven intent translation process.First,the users intentexpressed in natural language(as shown in Figure 18(a)
293、is translated into an intent model(Figure18(b)through the NLP function.NLP translates the user-expressed endpoints campus A andcloud2intothecorrespondingendpointnamesinthenetworktopology(transportEp_src_campus_A1 and transportEp_dst_cloud2)for SO and SDNC to comprehend.Next,the intent orchestrator i
294、n UUI translates the intent definition into a customer servicemodel,which is a specification of the CLL service the user wants to create(Figure 18(c).Thecustomer service model specifies the connectivity and Service Level Agreement of the CLLservice in a technology-agnostic manner.Finally,the custome
295、r service model is translated into anetwork configuration model,as shown in Figure 18(d).These models represent the actualnetwork configurations of EVC and its underlying OTN tunnels required to implement the CLLservice.Figure 18 Model-Driven Intent Translation36/56Once the intent is fulfilled,UUI p
296、resents the configured network service to the user,as shownin Figure 19.It assists the user in inspecting and validating whether the network aligns with theirintent.In addition,the specific content of intent guarantee can also be displayed in UUI.Figure 20shows the closed-loop process in a CLL use c
297、ase.It can be seen that the actual bandwidth usagecan be retrieved by the system.When the actual bandwidth exceeds the preset bandwidth,thesystem can proactively adjust the maximum bandwidth to avoid traffic congestion.Figure 19 CLL Service in UUIFigure 20 Closed-Loop GuaranteeThe blue curve represe
298、nts the actual bandwidth usage monitored by the intent-drivenautonomous network solution,and the green curve represents the pre-configured maximumbandwidth.When a bandwidth alarm is triggered,the system can send a notification to UUI torequest user intervention.Users can choose to modify the origina
299、l intent to increase the maximumbandwidth.6.2 Application Cases-China Unicom6.2.1 Intent-driven Autonomous Orchestration CaseChina Unicom has joined forces with manufacturers and provincial branches to adopt theconcept of user experience as the center to create the industrys intent-driven 5G endogen
300、ousintelligentorchestrationnetworkapplications,andinnovativelyimplementtheuserexperience-centered intent-driven approach based on network endogenous intelligence.Practicalapplications of the network.This solution is driven by user intents and relies on 5G-A-orientedendogenous intelligence.Through in
301、telligent means,it achieves dynamic network policies andresource configuration based on business requirements to meet the evolving demands of diversenetwork services in the future 12.In this example,the extrinsic intent from the servicemanagement layer is mapped to network execution policies at the
302、resource management layer,enabling intent-driven autonomous orchestration network applications.37/56Figure 21 Intent-Driven Autonomous Network OrchestrationCompared with traditional solutions,this solution boasts characteristics such as intent-drivenoperation,experience-driven design,endogenous inte
303、lligence,precise orchestration,and agileswitching.The solution can significantly improve user experience for 5G services such ashigh-definition live streaming,video calls,video ringback tones,and AR/VR.In the past,in areaswith poor 5G coverage and high interference,these services were prone to frequ
304、ent screen blurand freeze,leading to a drop in user perception.Intent-driven intelligent orchestration ensures thatnetwork resources can always meet the service needs of users and greatly improve userperception.The intent-driven autonomous network orchestration solution leverages the extrinsicintent
305、-driven knowledge base within 5G base stations,comprising both a business knowledge baseand a logical raster knowledge base,as shown in Figure 21.The role of the business knowledgebase is to assist in accurately identifying user business requirements and making precisejudgments regarding the user ex
306、perience in the source cell.Meanwhile,the logical rasterknowledge base aids in anticipating and precisely selecting the experience in the target cell,facilitating seamless and agile switching.Through the precise identification of user businessrequirements,precise judgment of the user experience in t
307、he source cell,anticipation and preciseselection of the experience in the target cell,and seamless and agile switching,this solutionimplements intent-driven flexible orchestration services,thereby ensuring differentiated userexperiences.6.2.2 Intent-Driven Autonomous Energy SavingAs an important fou
308、ndation for the development of the digital economy,the 5G network hasbecome a key driver in the national dual carbon strategy.Promoting the green development of5G is a fundamental requirement for ensuring the sustainable development of resources and theenvironment and is of great significance to the
309、 comprehensive green transformation of theeconomy and society.In order to implement the national dual carbon strategy and meet the38/56requirements for green 5G development,China Unicom has created a practical case of IDNs withgreen and low-carbon as its core.Figure 22 Intent-Driven Network Autonomo
310、us Energy-Saving SolutionIn traditional energy-saving solutions,it is difficult to achieve intelligent energy-savingacross multiple systems and maximize the overall network energy efficiency,which is animportant challenge in building green 5G.Compared with traditional solutions,this solution isdrive
311、n by the intent of autonomous network energy saving.It uses autonomous means to achieveintelligent energy saving in RANs in multiple modes and at multiple sites,meeting therequirements for intelligent,refined,and large-scale closed-loop management of networkenergy-saving.In this example,the energy s
312、aving intent from the resource management layer ismapped to energy saving execution policies at the network element management layer,enablingintent-driven autonomous energy saving.Driven by the intent of network energy-saving,the IDN autonomous energy-saving solutionrelies on various network energy-
313、saving scenario models,traffic prediction models,and networkdeployment methods.It combines the preset operator policy intent tendencies to achieve theautomatic and precise selection of energy-saving policies and optimization of energy-savingparameters.As shown in Figure 22,the solution forms a close
314、d-loop process of intent awareness,intelligent analysis,flexible decision-making,and automatic execution.This process aims toreduce energy consumption while ensuring network performance.6.2.3 Automatic Drive Test Intelligent Evaluation SolutionTraditional drive test mainly relies on manual intervent
315、ion,which is time-consuming,labor-intensive,and lacks intelligence.China Unicoms automatic drive test intelligent evaluationsolution makes full use of the capabilities of intent-driven autonomous networks.The integrationof testing,evaluation,calculation,and management revolutionizes the traditional
316、drive testmode and reduces the labor-intensive workload of drive testing and analysis.In this example,theintent of the automatic driving test and intelligent evaluation from the service management layer ismapped to execution policies at the resource management layer,enabling intent-driven39/56autono
317、mous driving test.Figure 23 Automatic Drive Test Intelligent Evaluation Solution Flow ChartAs shown in Figure 23,the solution is deeply empowered by IDNs.It automatically conductsproblem identification,root cause analysis,intelligent decision-making,automatic execution,andclosed-loop evaluation of e
318、xecution effects,thereby enabling end-to-end management ofproblematic road sections and achieving Level 3 autonomy level in this scenario:(1)Problem sensing:The solution automatically collects massive network MR/MDT data,uses AI models to fit key indicators of road conditions,and achieves automatic
319、collection,processing,and geographical presentation of four major categories of indicators,includingcoverage,quality,voice sensing,and data sensing.Based on this,problematic road sections areautomatically identified.(2)Root cause analysis:The solution conducts root cause analysis of identified probl
320、ematicroad sections,quickly presents the analysis results of problematic road sections,and shortens theanalysis time from 2 hours required by traditional methods to 15 minutes,significantly improvinganalysis efficiency.(3)Intelligent decision-making:Based on the analysis results of problematic road
321、sections,the solution uses the decision-making AI model to provide optimal solution suggestions,includingspecific optimization measures and parameter tuning suggestions.(4)Automatic execution:The system can flexibly connect to the ticket system,directly pushroot cause analysis results,handling sugge
322、stions,and key indicators to the front line,helping thefront line to quickly solve problems.The system automatically evaluates the effectiveness of the optimization results after a ticketis submitted,and judges whether the problem is solved,and how effective the solution is,therebyachieving end-to-e
323、nd closed-loop control of problematic road sections.At present,this scenariohas been applied on a large scale in 31 provinces,with a total of 5.21 million kilometers of 4G/5Gtest-free mileage,equivalent to a test cost of CNY 109 million.After actual verification of existingnetworks in many provinces
324、,the matching accuracy of road quality problems found in traditionalmanual road tests is over 90%.6.3 Application Cases-China Mobile6.3.1 Intent-driven Hierarchical Service Assurance CaseCurrent network service assurance faces several challenges:poor user experience due to thehigh skill requirements
325、 of traditional operations,long response time caused by numerous O&Mparameters and complex processes,lengthy feedback cycles lacking timeliness,rigid schedulingpolicies unable to flexibly meet diverse business needs,and the inability to provide continuousassurance covering the entire lifecycle of se
326、rvices.To address these issues,meet the differentiated40/56business needs of users,and enhance the flexibility of business services,we have adopted anintent-driven service assurance method.This method aims to promptly and efficiently address thediverse needs of different users,thereby improving the
327、overall business experience.In thisexample,the intent for hierarchical service assurance from the business management layer ismapped to differentiated scheduling policies at the service management layer,implementingintent-driven differentiated service assurance.Intent-driven hierarchical service ass
328、urance involves recognizing intents,matching them withdifferent levels of assurance capabilities,and providing differentiated guarantees for networkservice quality.These include the following different hierarchical assurance capabilities:(1)Gold/Level 1 assurance:strictly restricted usage.This level
329、 of assurance supportsunlimited wireless resource preemption,supports intelligent scheduling policy optimization,andenables all high computing power overhead functions.(2)Silver/Level 2 assurance:only targets key service assurance.This level of assurancesupports limited wireless resource preemption,
330、supports intelligent scheduling policy optimization,and enables high computing power overhead functions on demand.(3)Copper/Level 3 assurance:on-demand usage.This level of assurance does not supportwireless resource preemption,but supports intelligent scheduling policy optimization,and enablescomput
331、e-intensive functions on demand.Multi-dimensional differentiated services are provided for the preceding different hierarchicalassurance capabilities:1)Differentiated scheduling priorities:Diverse priority levels are supported based ondifferent 5QI,NSSAI,and types of services.This includes facilitat
332、ing wireless resourcepreemption for higher-priority services.2)Differentiated scheduling policies:Distinct schedulingparameter configurations,such as configurations for the target BLER and uplink power control,are supported based on different 5QI,NSSAI,and service types.3)Differentiated schedulingal
333、gorithms:High computing power overhead functions such as enhanced LDPC iterations andPUCCH format1 IRC can be enabled for specific 5QI,NSSAI,and types of services.In practical terms,the intent-driven hierarchical service assurance system follows thefollowing process:(1)Intent input:The user inputs sensing service logic in natural language,including theregion,app,time,and assurance level.(2)Intent