TMForum：2024年自智網絡產業白皮書（英文版）（64頁）.pdf

《TMForum：2024年自智網絡產業白皮書（英文版）（64頁）.pdf》由會員分享，可在線閱讀，更多相關《TMForum：2024年自智網絡產業白皮書（英文版）（64頁）.pdf（64頁珍藏版）》請在三個皮匠報告上搜索。

1、1 Back to ContentsObjective:to highlight a blueprint for progression toward Level 4 autonomous networks,with a particular focus on high-value scenarios.AN JOURNEY GUIDELevel 4 industry blueprint high-value scenariosNOVEMBER 2024NetworksAutonomous2 Back to Contents Contributing and endorsing companie

2、s3 Back to Contents Accenture Adriano Poloni,Hakan Ekmen,Tunc Yorulmaz,Carmelo CiceroAIS Boonchoung Tansuthepverawongse,Wasit WattanasapAsiaInfo Sen Bian,Ye Ouyang,Chen Hehe,Wang YingBUPT Wenjing LiCAICT Ma Junfeng,Xu Yunbin,Liu Zhiruo,Zhang Shiqi China Mobile Kaixi Liu,Yuan Yao,Gu Xianyao,Tian Tian

3、China Telecom Shao Xinhua,Jia Yan,Wang Yanchuan,Zeng Yu,Zhu Ting,Zhang Le,Yuan Jingjing,Ning Jingwen,Wang Zeyu,Wu ShanshanChina Unicom Li Hongwu,Zhang Taoye,Zhao Jingyi,Geng Yan,Yang Jianjian,Zhang Ding,Zhao Zhanchun,Zhao Yongjian,Yang Jieyan,Li Xinglong,Zhang Jie,Wang Rong,Sun HongChina Unitechs So

4、ng Gang,Liu Jun,Huang Jianfeng,Wang HanChunghwa Telecom Hao-Yu Kao,Kai Ming Teng,Tse-Han WangDetecon Esa VesterinenEricsson David Condon,Tony Salim,Jrg NiemllerHKT Derek Chen,Lloyd Chan,Michael Yue,Kam-Shing FungHuawei Jiang Li,Lin Yongming,Paul Linhua Deng,Paulo Shikun Wang,Tim Tianxi Song,Zhao Fen

5、ghua,Zheng Guangying,Jiaxu Huang,Liu Xianpeng,Frederick Mang,Zhang Yejun,Yunhe Wu,Fernando Camacho,Kevin McDonnell,Song Xiaodi,Wang Yining,Zhu Feng,Wang Jian,Yang Zhilong,Zang Longquan,Hao Xingyuan,Zhang Guogang,Ma Xiping,Zhao Yueliang,Xiao ChuanyiInspur Zhang Chuangang,Shen Linjiang,Chu Yufei,Xiao

6、Hongmei,Han Jianyou,Han Quanlei,Li Quan,Li Bin,Sun XuezhiMTN Daniel Smith,Lloyd Mphahlele,Mohamed Salah,Tinyiko Mabunda,Nkosinathi NzimaSelfX Now Min HeSingtel Timothy Hui Chee Kin,Rakesh V Raja Gopal,Kay Hong LimTelecom Argentina Adrian Grimaldi,Eduardo Panciera Molanes,Ruben Delgado,Juan Pablo Mon

7、arca,Nicolas Tellechea,Damian Del Campo,Ana Olivieri,Juan Martin Bonelli,Nicolas Visca,Mariano A.Tomas,Diego Martinez HeimannTelefnica Edi Fernando Viera,Pedro Garcia Parra,Hector Bahia Aguilar,Jose Maria Ramon PardoTelekom Malaysia Elias Bin Ahmed Kamal,Muhammad Qusairy Bin MohamadTelkomsel Ari Son

8、dang,Deasy Apriani,Roy Simatupang,Wahyudi C.Purnama,Citra Tayuwijaya,Jendrohartono,Dwi S Utomo,Hery Wijayanto,Asep Jajang NTM Forum Alan Pope,Dave MilhamUbiqube Herv Guesdon,Raja Ramamurthy,Amit RajUltraPower Xiaojun Wang,Peng Liu,Zhen Chen,Chao Wang,Guoxing CaiWhale Cloud Guo Yihua,Luo Qi,Lu Chunta

9、o,Yang ZizhenZTE Ju Manchang,Xu Kun,Jiang Xianzhong,Xie PengxiangCONTRIBUTING COMPANIES&INDIVIDUALSEDITORSTM Forum Olta Vangjeli,Andy Tiller,Dawn BushausAN Project members Dong Sun,Kevin McDonnellEditors and contributors4contentsIntroduction 6Section 1:How does TM Forum define autonomous networks?7A

10、N vision 8AN Framework 9Intent and autonomous domains 9KEIs 10Autonomous network levels 11Reference architectures 12AN map,strategic planning and journey 13Section 2:What is the AN Level 4 Industry Blueprint?15Defining Level 4 AN 16AN L4 Reference Architecture includes AI 16Identifying L4 high-value

11、 scenarios 17Target-state design for Level 4 high-value scenarios 18Measuring value with KEIs,KBIs and KCIs 21Evolving to Level 4 23Section 3:Where are we on the AN journey?24Pilot assessment results 25CSPs AN strategies and progress 27Section 4:Best practices for implementing AN high-value scenario

12、s 30China Mobiles Level 4 AN practice aims for lights-out factory 30China Telecoms AN business value assessment 33AISs AN Level 4 architecture evolution 34Telecom Argentinas target architecture development 36Telkomsels AN Level 3.5 architecture design 37Singtel Group and OpCos AN practice collaborat

13、ion 38Telecom Argentinas AN use case selection and development 39Section 5:Innovation in autonomous networks examples of AN use cases and scenarios 41Enabling verticals:CSPs join forces in autonomous computing force network Catalyst 41Improving CX:AISs CX-driven optimization evolution 42Improving CX

14、:China Unicom automates and improves all-optical private line services 44Efficient O&M:China Telecom uses its own LLM in network maintenance scenario 45Efficient O&M:China Mobiles digital twin-based intelligent network optimization closed loop 46Efficient O&M:China Telecom reduces IP network failure

15、s with online intelligent simulation 48Efficient O&M:MTN builds a high-efficiency IP+optical autonomous network 50Efficient O&M:Telecom Argentina optimizes Wi-Fi in the home 51Efficient O&M:Telkomsels CPRI optical link fault prediction and prevention 53Efficient O&M:stcs digital twin-based cloud-net

16、work topology visualization and analysis 54Efficient O&M:China Mobiles digital and intelligent convergence of dual-domain network O&M 55Green energy:HKTs network evolution aims to increase sustainability 57Section 6:AN standards update 59Section 7:Whats next for autonomous networks?61Sizing the oppo

17、rtunity 63GenAIs impact 63Next steps 64Additional resources 64TM Forum 2024.The entire contents of this publication are protected by copyright.All rights reserved.The views and opinions expressed in this white paper are provided in thecontributors personal capacities and may not reflect the views of

18、 their companies.While allcare has been taken in preparation of this paper,no responsibility for any loss occasioned toany person acting or refraining from any action as a result of any material in this publicationcan be accepted by the editors,contributors or publisher.5Autonomous networks(ANs)repr

19、esent a transformative shift for communications service providers(CSPs),enabling them to operate more efficiently,deliver better customer experience and innovate rapidly.TM Forum is at the forefront of industry-wide collaboration to drive the deployment and evolution of AN technology,processes and b

20、usiness models.During the past year,CSPs have taken significant steps forward in laying the foundation for autonomous networks by upgrading ICT infrastructure including wireless and fiber-optic networks.Some mobile operators are beginning to roll out 3GPPs 5G-Advanced standards,while optical backbon

21、e networks are entering the 400-Gbps era,enabling ubiquitous,high-quality 10-Gbps access for end users.At the same time,AI technology is evolving rapidly,with CSPs beginning to use GenAI and launch projects to develop telecom-specific language models.Telco operations are evolving,too,as operators ad

22、opt platform business models such as network-as-a-service(NaaS)to abstract the complexity of the underlying network using APIs such as the TM Forum Open APIs.Ultimately,CSPs need to be able to scale orchestration of zero-touch,zero-wait and zero-trouble services end to end across network domains(for

23、 example,radio access,fixed access,core,IP,optical transport and data center/cloud networks).This requires automation of the entire service lifecycle,from ordering to fulfillment,activation,orchestration,management,assurance,optimization and billing.Telcos have been working on this kind of automatio

24、n for years and have made progress.In fact,TM Forums Open API program resulted from these efforts.But end-to-end automation has remained elusive because of interoperability issues,architectural fragmentation and heavy customization.TM Forums Autonomous Networks Project aims to change this.The projec

25、t was formed in 2019 to define what it means for networks to be“autonomous”and to explain how to get there.Unlike automation,which relies on predefined rules,autonomy involves intelligent systems making independent decisions.Transitioning from automation to autonomy is a significant business and tec

26、hnology challenge,requiring straightforward implementation and integration patterns.To help CSPs manage the transition from automation to highly autonomous networks,we published our first AN white paper in 2019,and since then the publication has evolved into an AN journey guide this edition marks th

27、e sixth annual.Our goal is to build consensus around how to evaluate,design,implement and manage autonomous networks.GETTING TO LEVEL 4A key element of TM Forums AN work that is gaining traction across the industry is a six-level taxonomy CSPs can use to measure their progress implementing autonomou

28、s networks.Each AN level has a set of characteristics describing the evolutionary stage of the CSPs journey from manual to fully autonomous operations(see graphic).Introduction Back to Contents|IntroductionTM Forum,2024Autonomous network levelsFully autonomous network:The system has closed-loop auto

29、mation capabilities across multiple services,multiple domains(including partners domains)and the entire lifecycle via cognitive self-adaptation.Highly autonomous network:In a more complicated cross-domain environment,the system enables decision-making based on predictive analysis or active closed-lo

30、op management of service-driven and customer experience-driven networks via AI modeling and continuous learning.Conditional autonomous network:The system senses real-time environmental changes and in certain network domains will optimize and adjust itself to the external environment to enable,closed

31、-loop management via dynamically programmable policies.Partial autonomous network:The system enables closed-loop operations and maintenance for specific units under certain external environments via statically configured rules.Assisted operations and maintenance:The system executes a specific,repeti

32、tive subtask based on pre-configuration,which can be recorded online and traced,in order to increase execution efficiency.Manual operations and maintenance:The system delivers assisted monitoring capabilities,but all dynamic tasks must be executed manually.0143256 Back to Contents|IntroductionAccord

33、ing to TM Forums definition,Level 4 autonomous networks represent a major shift from traditional human-defined automation processes to true autonomous decision-making.This is a significant leap from Level 3,where machines assist humans in decision-making but still rely on human oversight.At Level 4,

34、the network can self-manage,self-optimize and handle complex tasks.To achieve this level of autonomy,CSPs need to translate their customers business goals,or“intents”,into network requirements that can be fulfilled without the customers needing to understand the underlying infrastructure.This requir

35、es a formal ontology to express these intents in a way that enables seamless machine-to-machine communication.Generative AI can be used to interpret natural language inputs from customers and generate intent expressions that the network can act upon.To date,more than 60 companies have signed the Aut

36、onomous Networks Manifesto pledging to work rapidly toward Level 4 autonomy.Even more have participated in AN Catalyst proofs of concept.Across the telecoms industry,more than 120 AN standards and research projects have been initiated,with multiple standards-development organizations(SDOs)and open-s

37、ource groups working to define and deploy ANs.Within TM Forums AN Project,many CSPs have outlined their AN visions and strategies,and 24 have carried out Autonomous Network level(ANL)assessments to promote AN principles and practices.We look at the anonymized results of these assessments in Section

38、3 of the guide.IDENTIFYING LEVEL 4 HIGH-VALUE SCENARIOSA main focus of the guide this time is on creating a blueprint for Level 4 high-value scenarios.Scenarios are use cases that combine an operation flow and a network domain(for example,fault management on the RAN).Level 4 scenarios are those that

39、 are likely to have maximum impact on the operations and/or business of a CSP.We explore how telcos and their technology partners are developing these scenarios,with a goal of evolving the necessary standards and sharing knowledge to make it easier to reach Level 4 autonomy.In September 2024,TM Foru

40、m launched a special type of Catalyst proof of concept called the Autonomous Network Innovation Pioneer Project to study AN Level 4 high-value scenarios.That teams first implementations are being demonstrated at Innovate24 Asia.The goal for each scenario is first to identify gaps between the current

41、 level of automation in CSPs networks and Level 4 autonomy.Then,the teams design solutions,implement and test them,and finally verify the achievement of AN Level 4.Teams participating in the ongoing project are leveraging TM Forums Autonomous Network Framework,which supports intent-based management

42、and intelligent orchestration to simplify network management,automate service delivery,and enable self-optimizing and self-healing capabilities.We discuss this framework throughout the guide.To create the guide,TM Forum solicited contributions from companies and individuals working both inside and o

43、utside TM Forums AN Project.The guide provides an overview and summary of selected contributions.In the appendix you will find links to all contributions received,which have been published in their entirety on the AN Projects webpage.You can read the first five AN guides by clicking on the images be

44、low.To learn more about how to get involved in TM Forums industry-leading work on autonomous networks,please contact Alan Pope.1 Back to ContentsObjective(s):to share the latest updates of business requirements,architecture,capabilities,collaborative activities,best practices and reference implement

45、ations in the Autonomous Networks industry.A WHITEPAPER ONEmpowering digital transformation from strategy to implementationSEPTEMBER 2022 autonomous networks1 Back to ContentsObjective:to share the latest updates of business requirements,architecture,capabilities and collaborative activities,best pr

46、actices,and reference implementations on Autonomous Networks to enable value-driven transformation.AN JOURNEY GUIDEEmpowering digital transformation evolving from Level 2/3 towards Level 4 SEPTEMBER 2023NetworksAutonomous7CSPs operating mobile and/or fixed networks have adopted a common vision for t

47、heir businesses to deliver connectivity and other services via autonomous networks that use predictive analytics and AI,including generative AI(GenAI),to determine which network resources are used to fulfill a customers business goals,with all complexity hidden from network operations teams and cust

48、omers.As this guide shows,the vision is becoming reality.TM Forum is one of several SDOs and open-source groups that are defining and evolving autonomous networks concepts,standards,implementation methods,and use cases to build consensus across the telecoms industry.Achievements include:Agreement on

49、 concepts such as the vision for Zero-X experiences,Self-X capabilities,AN levels(ANL),and key effectiveness indicators(KEIs)to measure the business value and operations effectiveness of AN scenarios Development of a target architecture supporting single-domain autonomy,cross-domain collaboration an

50、d intent-driven,full-stack AI Collaboration among SDOs such as 3GPP,CCSA,ETSI,GSMA,IEEE,IETF,ITU,NGMN and TM Forum to unify AN standards(known as MDSO)Pledges from more than a dozen CSPs to aim for Level 4 autonomy in multiple domains between 2025 to 2027.TM Forum members have developed an AN target

51、 architecture,which includes guiding architectural principles,a business architecture,and technical and formal reference architectures.CSPs already leveraging TM Forums Open Digital Architecture(ODA)can now integrate autonomous capabilities into their existing frameworks,transforming ODA into compos

52、able,self-managing systems organized as autonomous domains.All these assets are comprehensively defined in TM Forums standards,and most of the concepts have been outlined thoroughly in previous AN guides.We recap the most important concepts and work here.How does TM Forum define autonomous networks?

53、Back to Contents|How does TM Forum define autonomous networks?Section 1CSPs have a common vision to deliver connectivity and other services via autonomous networks that use predictive analytics and AI to determine which network resources are used to fulfill a customers business goals.8 Back to Conte

54、nts|How does TM Forum define autonomous networks?AN VISIONThe vision for autonomous networks is to provide zero-wait,zero-touch and zero-trouble customer experiences for consumers and businesses through intelligent infrastructure and agile operations(see graphic above).This makes the end user experi

55、ence better,by providing higher availability of network services,quicker fulfilment of orders and easier description of orders as“intents”rather than detailed specifications.At the same time,ANs need to support self-service,self-fulfillment and self-assurance of telecom network infrastructures for i

56、nternal users across various departments including planning,marketing,and operations and maintenance(O&M).TM Forums AN visionSmart CitySmart IndustrySmart GovernmentSmart XZero-waitZero-touchZero-troubleAs a ServiceAs aPlatform Launch Delivery CareSwift Operation Development MaintenanceSimplified In

57、frastructure Service BusinessSelf-healingSelf-serving Planning Ordering MarketingLeanSelf-fulfilling Organizing Orchestration ConfigurationAgileSelf-assuring Monitoring Healing OptimizingPreventiveZero-X ExperienceSelf-X O&MTM Forum,2024The vision for autonomous networks is to provide zero-wait,zero

58、-touch and zero-trouble customer experiences for consumers and businesses through intelligent infrastructure and agile operations.9 Back to Contents|How does TM Forum define autonomous networks?AN FRAMEWORKTM Forums AN Framework(ANF)is an out-of-the-box implementation methodology to help CSPs plan a

59、nd implement AN in an efficient and systematic manner(see graphic above).It consists of four key elements:KEIs to measure the benefits of implementing AN use cases AN level evaluations to measure progress Reference and technical architectures to guide implementation An AN Map to plan capability depl

60、oyment.The framework also includes operations best practices along with a methodology and tools for assessing AN levels.We discuss level evaluation in section 4,while section 5 focuses on best practices.INTENT AND AUTONOMOUS DOMAINSTo provide Zero-X experiences and Self-X O&M,the AN target architect

61、ure uses intent-based,closed-loop management as illustrated in the graphic on the next page.Intent is a way to express users requirements,goals and constraints,and then interact with them in various domains in a way that allows a system to adapt its operation accordingly.Importantly,intents are abst

62、racted from the technical inner workings of the network.Put more simply,intents are the“what”not the“how”meaning,you tell the system what goal or outcome is required without having to tell it how to achieve the goal.This decoupling increases agility,allowing autonomous systems to fulfill the intent

63、using data,machine learning and AI,while enabling supplier innovation that can be easily integrated and adopted.Intent can be expressed at various operational layers(business intent,service intent and resource intent,for example).In the AN Reference Architecture(IG1251),intents guide the closed loop

64、 control process,which is a continuous process of monitoring,analyzing and adjusting performance to improve the quality of experience for end users and automate the full lifecycle of services.Autonomous domains can be defined based on factors such as service types,network technologies or deployment

65、locations.From an infrastructure perspective,examples include closed-loop systems within an operators access,metro backbone,core,edge and customer networks.At the service operations layer,autonomous domains may also be organized around specific services such as SD-WAN,Voice over LTE(VoLTE)or content

66、 delivery networks(CDNs),with collaboration between domains to manage and optimize service delivery.The AN framework is an out-of-the-box implementation methodology to help CSPs plan and implement AN in an efficient and systematic manner.TM Forum Autonomous Network FrameworkTM Forum,202410 Back to C

67、ontents|How does TM Forum define autonomous networks?Each autonomous domain operates independently,guided by the operators business objectives,while hiding implementation details from customers through an API abstraction layer.When multiple autonomous domains are defined and deployed,the upper layer

68、 service operations leverage intent-driven interactions between domains to manage and fulfill the entire service lifecycle.TM Forum members have collaborated in several Catalyst proofs of concept to develop the Intent Management API(TMF921),which communicates between autonomous domains to manage clo

69、sed-loop operations across business,service and resource layers.High-level descriptions of intent are defined in TM Forums Autonomous Networks Business Requirements and Framework(IG1218),and the technical work is published in Intent in Autonomous Networks(IG1253).KEIsKey effectiveness indicators,or

70、KEIs,measure the business value and operations effectiveness of autonomous networks.The AN KEI system included in the AN Framework is based on:the balanced scorecard theory(a widely used KEI creation tool);CSPs practical experience with KEIs;and TM Forums indicator framework as described in TM Forum

71、 Metrics Framework(GB935),TM Forum Metric Definitions(GB988)and other standards documents.This value-driven system focuses on both effectiveness and efficiency.TM Forum members have also created the AN KEI Framework(explained in Autonomous Networks Effectiveness Indicators IG1256),which allows CSPs

72、to add,delete or modify indicators to develop their own KEI system.We discuss KEIs more in the next section.Intents are the “what”not the“how”meaning,you tell the system what goal or outcome is required without having to tell it how to achieve the goal.This decoupling increases agility.Intent-based,

73、closed-loop managementTM Forum,202411 Back to Contents|How does TM Forum define autonomous networks?AUTONOMOUS NETWORK LEVELSAs explained in the introduction,TM Forum has established a six-level taxonomy to categorize the progression of autonomy in networks,ranging from Level 0(no automation)to Leve

74、l 5(full autonomy).This model is gaining acceptance industry wide,with other SDOs such as the China Communications Standards Association(CCSA)using the AN levels to define autonomous networks and measure CSPs progress in adopting them.The ANL evaluation methodology is published in Autonomous Network

75、 Levels Evaluation Methodology(IG1252),which describes the evaluation approach and operation flows,task-evaluation criteria,scoring method,etc.An evaluation tool(GB1059)enables CSPs to measure their operations current level of autonomy.The first step is identifying“evaluation objects”,which are dete

76、rmined by selecting an operational use case based on services domain,network technology domain and operational flows(see graphic on the next page).Autonomous LevelsL0:Manual Operation&MaintenanceL1:Assisted Operation&MaintenanceL2:Partial Autonomous NetworksL3:Conditional Autonomous NetworksL4:High

77、Autonomous NetworksL5:Full Autonomous NetworksExecutionPP/SSSSSAwarenessPP/SP/SSSSAnalysisPPP/SP/SSSDecisionPPPP/SSSIntent/ExperiencePPPPP/SSApplicabilityN/ASelect scenariosAll scenariosP People(manual)S System(autonomous)TM Forum AN level taxonomyTM Forum,2024TM Forums AN level taxonomy is gaining

78、acceptance industry wide,with other SDOs using it to define autonomous networks and measure progress.12 Back to Contents|How does TM Forum define autonomous networks?The first step in AN level evaluation is identifying evaluation objects,which are determined by selecting an operational use case.Then

79、,CSPs can evaluate AN levels for these scenarios based on an assessment questionnaire that evaluates cognitive activities in the network like intent handling,awareness,analysis,decision and execution(IAADE),each of which contains a set of tasks to be measured.At Level 3 and below,users intents can b

80、e achieved using policy-driven operations and existing interfaces.At Level 4,AI generates rules so that the network can optimize itself.Well discuss AN level assessment more in Section 3.REFERENCE ARCHITECTURESIn addition to standards for ANL assessment and KEIs,members of the AN Project are working

81、 on reference,business and technical architectures as part of the AN Framework.While TM Forum is focusing on creating domain-agnostic AN standards,other SDOs such as 3GPP and IETF are defining technology-specific autonomous domain architecture.TM Forum is working with these groups to align standards

82、.We discuss standards more in Section X.AN evaluation objectsTM Forum,2024AN MAP,STRATEGIC PLANNING AND JOURNEYFinally,as part of the AN Framework TM Forum members have developed a set of best practices for planning and deploying AN.These include the AN Map,AN Strategic Planning(ANSP)and AN Journey(

83、see graphic above).The AN Map identifies the scope and priority of AN planning and deployment,serving as a“sandbox”for evaluating and promoting AN work.Creating the map began with defining AN implementation scenarios.The team initially identified the 11 high-value scenarios below.We discuss these sc

84、enarios and others in the next section,where we look at the AN Level 4 Industry Blueprint.1.Service marketing2.Service delivery3.Service assurance4.Complaint handling5.Network planning6.Network deployment7.Network change8.Fault management9.Quality optimization10.Energy-efficiency optimization11.Reso

85、urce management13 Back to Contents|How does TM Forum define autonomous networks?CSPs business strategiesHigh-level requirementsIndustry standardsCutting-edge technologiesKEIsAN mapTarget architectureLevel standardsANSPAN strategyExecutives commitmentAN blueprintIndustry contributionsInnovative techn

86、ologiesAN JourneyAssessmentEvaluation and objective formulationImplementationInnovation pilot and replicationGap analysisProcess walk-through and transformation itemsSolution designSystem capability breakdownAN MapAutonomous operations practicesTM Forum,2024The AN Map identifies the scope and priori

87、ty of AN planning and deployment,serving as a sandbox for evaluating and promoting AN work.14 Back to Contents|How does TM Forum define autonomous networks?ANSP outlines corporate level AN strategies and executives commitments while also providing guidance on implementing the four key elements of th

88、e AN Framework.It also specifies strategies for contributing AN strategies to the wider telecoms industry.The AN Journey(sometimes called AN practice)refers to the overall transformation process,guiding CSPs from their current state to the desired future state of higher network autonomy.It is an ite

89、rative,four-step process based on high-value scenarios for telcos to plan and execute their AN strategies:1.Assessment for each high-value scenario,CSPs evaluate the live network based on ANL and KEIs to establish baselines and set improvement goals based on service requirements and the investment p

90、lan.2.Gap analysis CSPs analyze the gaps between the baselines and goal to identify the root causes of breakpoints and weaknesses.3.Solution design based on their target architectures,CSPs convert capability requirements into functional requirements for their solution providers to follow and then te

91、st new developments in pilot solutions.4.Implementation based on the results of the pilots,CSPs launch the solutions within a designated operating company and then replicate in other companies to improve the groups overall ANL and KEIs.Next,we look at the development of TM Forums new AN Level 4 Indu

92、stry Blueprint and the role that AI plays in autonomous networks.The AN Journey(sometimes called AN practice)refers to the overall transformation process,guiding CSPs from their current state to the desired future state of higher network autonomy.15 Back to Contents|Section 2:What is the AN Level 4

93、Industry Blueprint?What is the AN Level 4 Industry Blueprint?Level 4 autonomous networks require AI.Indeed,at this level networks are transformed to become based on knowledge and inference so that they can behave with human-like intelligence and achieve an autonomous closed loop that encompasses con

94、text awareness,action planning,problem solving and decision-making.TM Forum sees Level 4 autonomous capabilities as critical to helping CSPs reach their business goals,or“value outcomes”.These include:Monetization of the network achieving time to market of one week or less for new services,plus real

95、ization of network-as-a-service(NaaS)and deterministic service level agreements(SLAs)Operational efficiency creation of digital employees to increase the number of devices maintained per person and reduce the onsite maintenance workload by thousands of person-years,while ensuring secure network oper

96、ationsResource efficiency reducing energy consumption and carbon-dioxide(CO2)emissions,and increasing return on network investmentExperience excellence reducing the rate of faults and complaints,providing minute-level service delivery and fault rectification,and improving SLA compliance.To guide Lev

97、el 4 AN deployment,TM Forums AN Project,in collaboration with other standards bodies,is developing the AN Level 4 Industry Blueprint.Based on the Autonomous Network Framework,the blueprint serves as a foundation for industry-wide collaboration on,and improvement of,AN capabilities.Section 2The AN Le

98、vel 4 Industry Blueprint explains that Level 4 requires end-to-end automation of NOC(network operations center)operations in high-value scenarios and a reduction in the number of site visits by engineers or technicians to just one.Four key elementsAN Level 4 Industry BlueprintVision:Zero-X/Self-XInd

99、ustry assessment and certificationService experienceNetwork capabilityProcesses and methodsAN map3ScopeCooperationArchitecture principlesTarget architecture2Autonomous Networks FrameworkinstantiationOperations practicesOrganizationOrganizational transformationOrganizational capabilitiesANStrategyPla

100、nAN JourneyAssessmentImplementationGap analysisSolution designClosed-loop evaluationL1 LxScenario-based51L4 Vision?4Value measurement2ArchitecturesFull-stack AI/Gen AI architectureRole-oriented copilots/operation scenario-oriented agentsAI/Gen AI drive AN L4 advancementDefine phased AN L4 roadmap to

101、 guide the pace of standards establishment and live networks deploymentPhase 1:2025-2027,Focus on selected 20 value scenariosPhase 2:2028-2030,More and E2E complex scenariosPhased evolution of AN L4 RoadmapSource:IG1326-AN wp 5.0EvolutionLevel characteristicsAN level1BenefitValue propositionsKBI/KEI

102、/KCI4Define L4 key valueTailored to value scenarios,E2E automation of NOC operations and self-service in one-time site visits for site operations.Implement 20 value scenarios preferentiallyDefine the L4 target status of high-value scenariosValue scenarios lead AN L4 implementation3Value ScenariosMea

103、sure L4 value and effectivenessKBI-KEI-KCI:Three-tiered index systemBVM Model:Measurement the business value of AN L4TM Forum,2024TM Forum AN Level 4 Industry Blueprint16 Back to Contents|What is the AN Level 4 Industry Blueprint?The blueprint includes:An expanded definition of Level 4 AN A referenc

104、e architecture that includes enabling technologies such as AI Summaries of high-value scenarios A measurement system including KBIs,KEIs and KCIs(key business,effectiveness and capability indicators)to assess scenarios A phased evolution roadmapDEFINING LEVEL 4 ANTM Forums AN level evaluation method

105、ology defines Level 4 AN like this:“In a complex,cross-domain environment,the system enables decision-making based on a predictive analysis or active closed-loop management of service-driven and customer experience-driven networks via AI modeling and continuous learning.”The AN blueprint goes furthe

106、r,explaining that Level 4 requires end-to-end automation of NOC(network operations center)operations in high-value scenarios and a reduction in the number of site visits by engineers or technicians to just one.NOC operations encompass all production tasks except those requiring onsite visits during

107、network planning,construction,maintenance,optimization and operations(though in principle AI-driven robots would handle site visits at Level 5).This includes analysis,prediction,solution development,decision-making,command and dispatch,and remote operations.The idea is to facilitate predictive and p

108、roactive O&M driven by enhancement of services and customer experience.This is achieved through predictive analysis based on real-time analysis of vast amounts of data,automated decision-making and automated use of tools.Site operations include activities that require onsite visits(for example,netwo

109、rk device installation,service provisioning,hardware fault rectification and board replacement).The goal is to achieve independent onsite operations performed by installation and maintenance engineers with the assistance of copilots so that they do not need additional human guidance.This approach en

110、ables all production tasks to be completed in one go,significantly reducing labor and O&M costs.AN L4 REFERENCE ARCHITECTURE INCLUDES AIThe TM Forum AN Reference Architecture introduces the concept of full-stack AI in autonomous networks,operating across layers of the AN Framework(see graphic on the

111、 next page).Heres what happens at each layer:Business operations layer AI focuses on implementing intelligent applications for business processes such as precision marketing,translation of customers intents,customer experience,and flexible combinations of products and offerings.Service operations la

112、yer AI focuses on implementing intelligent applications for service processes such as service intent translation,end-to-end resource orchestration and scheduling,SLA policy creation,and end-to-end fault and performance management.Resource operations layer AI focuses on local,lightweight development,

113、model retraining and AI asset management to optimize local intelligence and support intelligent perception,analysis,decision-making and execution on local networks.In network elements,AI can comprehensively detect network status in real time and perform inference.Effective cross-layer AI collaborati

114、on makes it possible to complete increasingly complex tasks.The goal is to support AI-based,automated closed-loop network operations and implement intelligent automation in different service scenarios to reach higher levels of AN.GenAI will play an important role,working together with traditional AI

115、/ML.ML can be used to identify patterns in data,predict results based on historical information and make decisions.GenAI is particularly well suited to tasks that require high creativity and innovation,such as intelligent Q&A for routine O&M tasks and complex network configuration.Combined,tradition

116、al AI and GenAI will be able to identify constantly changing network requirements and generate optimal network configurations.The TM Forum AN Reference Architecture introduces the concept of full-stack AI in autonomous networks,operating across layers of the AN Framework.17 Back to Contents|What is

117、the AN Level 4 Industry Blueprint?GenAI can be used at any operational layer in the AN reference architecture and can work in tandem with other types of AI to develop two primary kinds of agent applications:role-oriented copilots and scenario-based agents capable of acting on behalf of users or syst

118、ems.This integration significantly enhances autonomous capabilities across all operational layers(business,service,and resource operations)as depicted in the graphic above.The AN reference architecture inherently distributes intelligence and knowledge across all three operational layers,with each au

119、tonomous domain containing intelligence specific to its functions.As AI continues to evolve,the formal reference architecture remains robust and adaptable,accommodating advancements without the need for structural change.We discuss AI agents more at the end of this section.In December 2023,TM Forum

120、set up the GenAI for AN Standards working group in the AN Project and launched a discussion about a recently published document called Embracing Generative AI in Telecom:Amplifying Autonomous Network Evolution(IG1345).The group plans to expand on this work.IDENTIFYING L4 HIGH-VALUE SCENARIOSSelectin

121、g high-value scenarios helps CSPs specify the practice scope of autonomous networks.It also helps them improve their return on investment and achieve potential business or operational improvements.The AN Level 4 Industry Blueprint contains methods and references for identifying high-value scenarios,

122、on which the entire telecoms industry is now collaborating to refine.The identification method is as follows:1.CSPs clarify the scope of key services and networks and match them with the 11 end-to-end network operation processes to form a complete set of operation scenarios.2.Then,they comprehensive

123、ly analyze the cost and human resources needed,the impact on cybersecurity operations,and the technical feasibility of each scenario to select the scenarios they want to promote that is,Level 4 high-value scenarios.TM Forum,2024Read this research report to learn more about how CSPs are using AI:Full

124、-stack AI/GenAI in the AN architecture18 Back to Contents|What is the AN Level 4 Industry Blueprint?TARGET-STATE DESIGN FOR LEVEL 4 HIGH-VALUE SCENARIOSAccurately designing the Level 4 AN target state for high-value scenarios enables CSPs and suppliers to achieve technical breakthroughs,drive applic

125、ation innovation and accelerate the adoption of Level 4 practices.Here,we look at three examples:fault management(in multiple domains),wireless network quality optimization and complaint handling in home broadband services.In each case,the operator follows four steps:1.Set business and operations ef

126、fectiveness objectives establish L4 objectives for high-value scenarios by referring to AN KEIs2.Define AN capability objectives identify critical AN capabilities needed to achieve Level 4,focusing on the intent-awareness-analysis-decision-execution closed loop3.Break down AN capability requirements

127、 use the three-layer AN architecture(the business,service and resource operations layers)to sort key technologies and guide technical breakthroughs and solution design4.Define the Level 4 target process streamline the end-to-end business process,identify workload distribution and AN capability weakn

128、esses,and combine this with the AN capability breakdown to clarify development directions for key processes.Identification of AN Level 4 high-value scenariosTM Forum,2024OperationOptimizationPlanDeploymentMaintenanceOthers:e.g.Cloud19 Back to Contents|What is the AN Level 4 Industry Blueprint?Huawei

129、 Proprietary-Restricted Distribution3Level 4 target state:fault managementKBI objectivesReduced annual opex:$XXX,000Number of employees who have transitioned XXNumber of major faults per year:0KEI objectivesMTTR XX minutesNumber of major risks per month:0Faults and risks that can be remotely rectifi

130、ed:automatic rectification(manual review is required for certain ones)Faults and risks that require onsite operations:automatic and accurate diagnosis and automatic recovery solution recommendationAutomation rate:proportion of the number of sub-scenarios that have been automatically processed to the

131、 total number of sub-scenarios 90%Service operations layerCross-domain fault agents:Fault aggregation analysis across vendors and network domains as well as cross-domain fault diagnosis and rectification are automatically completed.Cross-domain fault copilots:Single-domain copilots are integrated as

132、 the unified entry for role-oriented copilots.Faults and risks are manually rectified based on the rectification guide provided by the copilots.Resource operations layer(EMS+NE)Single-domain fault agents:Faults and risks in a single domain(single vendor)can be automatically identified,diagnosed and

133、rectified.Additionally,the copilots are used to notify O&M personnel for manual takeover(if necessary).Single-domain fault copilots:Single-domain copilot APIs can be integrated with and invoked by cross-domain fault copilots.1.Network quality improvement:Analysis based on a single data source(alarm)

134、Analysis based on multiple data sources(alarms/performance/logs),quickly identifying and handling potential faults and risks,achieve proactive maintenance.2.New handling mode:Manual handling for all faults Automatic handling for certain faults by fault agents3.Scale of tickets decreased and new tick

135、ets dispatch method:Alarm based ticket dispatch Incident based ticket dispatch with accurate onsite handling End-to-end process from fault/risk identification to fault/risk rectification in single network domains(such as wireless,core,transport and IP networks)and cross-network domain scenarios(such

136、 as wireless+transport and IP+transport)Level 4 characteristicsFor single-domain/cross-domain faults and risks,all other phases are fully automated,except for necessary site visits(such as onsite hardware operations and high-risk command review).Scenario1.KBI/KEI objectives2.Autonomous capability re

137、quirements3.AN capability requirement breakdownCross-domain(cross-network domain and cross-vendor)fault agentFMESingle-domain(single-vendor)fault agentFault/Risk identification(A)Impact analysis,demarcation and locating(A)Rectification solution generation(D)Rectification and verification(E)Alarm col

138、lection/filtering/parsing/standardization/correlationAlarmAlarm-based ticket dispatchService impact analysisDemarcation and locatingRectification solution generationRectification and verificationTicket closureQuery/Analysis assistanceNOCBackend officeAs-isTo-beSingle-domain fault identification&risk

139、 prediction Demarcation and locating Rectification solution generation Rectification and verification in some scenariosAlarmPerformanceLog.Fault copilotsManual takeover of operation ticketsRectification and verificationTicket closureQuery Q&AFMEAccurate onsite handling suggestions11223344Cross-domai

140、n fault correlation Demarcation and locating Rectification solution generation Rectification and verification in some scenariosIncident-based ticket dispatchIncident reporting2Cannot be remotely rectified4.AN L4 major transformation itemHighlightsThe E2E automatic closure rate of routine optimizatio

141、n is over 60%.The total network optimization time is reduced by 15%.The network optimization issue handling duration is reduced by 25%.High-precision real-time awareness of poor network quality and user-/service-level experience deterioration100%automation of software commissioning optimization100%c

142、overage of three types of optimization sub-scenariosInteractive assistance and automatic optimization advice generation for optimization analysis and onsite optimizationNetwork optimization platformFormulates routine optimization appraisal objectives and ticket dispatch rulesOMCAutomatically analyze

143、s performance and experience issues，Automatic single-domain optimization or outputs optimization advice.Base station1）Accurately identify application-and user-level poor-QoE2）remote and automatic adjustment of digital antennas.HighlightsNetwork native intelligence will be introduced to implement pre

144、cise awareness of poor network quality and user-/service-level experience deterioration.Automatic system monitoring will replace manual monitoring to identify poor network/service quality and automatically generate tickets.More than 60%tickets can be automatically closed based on the agent.The optim

145、ization copilot will provide accurate adjustment and DT operation guides for live network optimization.4.AN L4 major transformation itemAs-isProvincial branch/Subsidiarycentralized monitoring teamProvincial branch/Subsidiarycentralized optimization teamMunicipal branchThird-party field engineersOnsi

146、te survey and handling(more than 50 times/100 sites)Ticket dispatch for all issuesManual monitoring andpoor-QoE issue identificationAnalysis and solution revisionTicket dispatch to citiesAnalysis and solution confirmationOn-site ticket dispatch3.Onsite optimization copilotTo-beCentralized monitoring

147、Third-party field engineersAnalysis and solution confirmation(MTTR reduced by 20%)1.Optimization agent2.Network optimization analysis copilotIntelligent service boardAutomatic system monitoring and poor-QoE issue identificationProvincial branch/Subsidiarycentralized optimization teamMunicipal branch

148、Analysis and solution revision(analysis ticket decreased by 20%)Onsite survey and handling(Efficiency improved by over 25%)Expert involvementMunicipal confirmationTicket combination to minimize site visitsGroup/SubsidiaryGroup/SubsidiaryEstablish optimization objectives of network quality/user exper

149、ience KPIs.Establish service objectives of network quality/user experience.L4 target state:wireless network quality optimizationLevel 4 characteristicsScenario Description Mainly including three sub-scenarios：1）Performance optimization involves coverage,rate,interference,high load,and low traffic.2)

150、User experience optimization involves top data services such as web browsing,video,gaming,payment,and instant messaging,and voice-related experience optimization for call connection,call drop,and voice quality.3）Typical scenes involve tidal traffic,burst traffic,and compensation for cell outage.The

151、E2E quality optimization process includes poor network quality/experience deterioration identification,root cause locating,optimization solution generation and decision-making,automatic optimization implementation,and optimization effect evaluation.24/7 proactive optimization is provided,and 100%aut

152、omation is achieved for routine wireless network optimization.More than 60%optimization tickets require no site visit.Self-service site visits are enabled to resolve issues at one time.1.KBI/KEI objectives2.AN capability requirements3.AN capability requirement breakdownNetwork Quality Deterioration

153、Identification(A)Impact analysis,demarcation,and locating(A)Optimization solution generation and decision-making(D)Rectification and verification(E)Level 4 target state:fault managementLevel 4 target state:wireless network quality optimization20 Back to Contents|What is the AN Level 4 Industry Bluep

154、rint?Onsite HBB complaint handling rate:45%Average duration of onsite complaint handling:65%Full ticket dispatch for faults encountered by specific usersInterception rate of faults encountered by specific users:35%2.Fault rectification assisted by digital customer service personnelSwitching to anoth

155、er Wi-Fi frequency band/Device reboot/Removing and inserting the network cableCopilotOnsite ticket handling rate:automatic rectification).Network department/NOCOSSNMS+NECopilotIncompatible Wi-Fi settings,roaming problems of sticky users,and a large number of user connectionsCall centerUsersInstallat

156、ion and maintenance engineersAgent:HBB experience assuranceOnsite ticket handling duration:45 minutesAs-isTo-be(Level 4)Highlights3.HBB faults encountered by specific users are rectified by frontline installation and maintenance engineerswithout NOC intervention.4.AN L4 major transformation itemLeve

157、l 4 target state:HBB service complaint handlingL4 target state:MBB customer complaint handling1、KBI/KEI objectives1)First call resolution 2)Auto complaint handling rate 3)Repeated complaint rate 2、Capability objectives1)Network complaint prediction and prevention 2)Auto complaint issue demarcation 3

158、)Auto network issue diagnosis 4)Auto execution and verification 3、Break down AN capability requirements End userIdentify user receiving poor quality with network root cause to support proactive customer care and self-service checkingCall agentUse LLM to interact with affected user to provide more in

159、formation online and close more complaints without network teams2nd tier/domain teamIntelligent complaint issue diagnosis and solution output to support auto-execution and verificationProactive user experience-oriented network optimization4、End-to-end business flowKey transformation pointFrom reacti

160、ve complaint handling to proactive and predictive From report complaint issue by end user to self-service checking and verify suggestion From manual complaint issue analysis and decision to intelligent outputFrom manual complaint issue resolution to auto-execution and verificationScenarioThe mobile

161、broadband(MBB)customer complaint handling process is important for routine management.It consists of tasks such as complaint prediction,analysis and decisioning,network issue execution,and complaint issue verification until closure of the complaint ticket.Call center,NOC and radio optimization teams

162、 use various tools to complete those processes.The main objective is to quickly analyze root cause and output precise resolution and suggestion to improve customer satisfaction.As-isTo-beComplaint awarenessComplaint analysisOutput solutionSolution execution&verificationReactive complaint receiveCrea

163、te complaint TT Analyze complaint based on information from call agent and network informationManual solution designGuide domain troubleshooting and suggestionDrive test surveyRemote or physical adjustmentCall back to customerEnd UserCall agent/Non-human touchpoint2nd tierDomain teamProactive poor-q

164、uality customer identification and customer careSelf-service for complaint issue handlingLLM to close more complaints based on network information and summary information from customer Intelligent diagnosis and effectiveness solution outputAuto handling and verificationProactively user experience-or

165、iented network optimizationLevel 4 target state:MBB customer complaint handlingThe mobile broadband(MBB)customer complaint handling process is important for routine management.It consists of tasks such as complaint prediction,analysis and decisioning,network issue execution,and complaint issue verif

166、ication until closure of the complaint ticket.Call center,NOC and radio optimization teams use various tools to complete those processes.The main objective is to quickly analyze root cause and output precise resolution and suggestion to improve customer satisfaction.ScenarioScenario Description21 Ba

167、ck to Contents|What is the AN Level 4 Industry Blueprint?MEASURING VALUE WITH KEIs,KBIs AND KCIsIn the Level 4 blueprint,TM Forum has expanded the KEIs outlined in the AN Framework to include a three-layer AN Indicator Framework that also uses KBIs and KCIs to measure the value of use cases.We expla

168、ined KEIs in Section 1(see page 11).KBIs measure the business value of AN scenarios and can be used to demonstrate this value to executives.KCIs are used to set capability development objectives and guide all departments and operating companies in developing them.The graphic below shows the AN Value

169、 Indicator Framework.KEIsKey effectiveness indicators For outcomes of scenarios presented to domain executives KCIsKey capability indicators Capability development and improvementKBIsKey business value indicators For overall AN value presented to CXOs and VPsValueoutcomeCapabilitydevelopmentKBIsKEIs

170、KCIsRevenue from operationsReduced O&M costsNumber of major faultsNumber of employees who have transitionedRevenue from capabilitiesDevice power-saving rateSaved energyProvisioning,complaint,fault and optimization duration.Service delivery timeComplaint resolution timeMean time to recoveryQuality op

171、timization timePrecision marketing success rateService delivery timeTotal stimulated traffic.NaaS API countNumber of NaaS API calls.Number of major risksMTTR of critical faultsOne-time success rate of major changes.Automation rate and intelligence rate/levelTrouble shooting,quality optimization,netw

172、ork change,energy-efficiency optimizationservice provisioning,complaint handling,etc.Proportion of inefficient resources.Service availability ratioTime to marketComplaint resolution timeO&M cost reductionOperations revenue increaseHighly stable networksAN Value Indicator FrameworkTM Forum,2024L4 tar

173、get state:RAN energy efficiency optimization1.KBI/KEI objectivesBase station energy-saving gain increases by 10%annually2.Capability objectivesData acquisition:one point access to the national base station data,resources,instructions,AI computing power,etc.Accurate prediction：Realize the intelligent

174、 prediction of the tide law of the base station business and determine the energy-saving timeSolution generation:Automatically generates a complete energy-saving solution,including energy-saving period and energy-saving method.3.Break down AN capability requirements Business operationBusiness intent

175、 translated into energy-saving policy requirementsService operationIt should support the automatic analysis capability of base station characteristics;the ability to generate energy-saving schemes that maximize the reduction of base station energy consumption and minimize network impact;the automati

176、c generation and delivery of instructions;and the optimization and adjustment of energy saving strategiesResource operation(OMC+NE)Realize the automatic collection of base station data,resources,instructions,AI computing power,etc.;can automatically execute instructions and feedback optimization res

177、ults4.End-to-end business flowKey transformation pointData analysis:Transition from historical data analysis to automatic collection and intelligent predictionEnergy-saving solution generation:Transition from expert experience to automatic generationStrategy configure:Transition from manual configur

178、ation to automatic scheduling and remote configurationFeedback adjustment:Transition from periodic manual adjustment to automatic parameter optimization through machine learningAs-isTo-beScenariosUsing intelligent technology to achieve accurate service prediction of the base station,generate persona

179、lized energy saving schemes and automate execution can effectively improve the energy-saving effect and reduce operating expenses.This includes collection of network data,analysis of service characteristics,generation of energy-saving schemes,dispatch of energy-saving instructions and provide feedba

180、ck optimization of execution results.Network data acquisitionIntelligent predictionAutomatic energy saving strategies generationAutomatic instruction schedulingResult feedback optimizationbase station data,resources,instructions,AI computing power,etcthe tide law of the base station businessRealize

181、the coordination of soft and hard close of base station energy savingUsing reinforcement learning techniques,parameter thresholds in policy configuration are optimizedThe remote control of the base station is realized through the command online arrangement and command automatic scheduling moduleHist

182、orical data analysis Configure strategy deliveryfeedback and optimization periodicity Energy saving plan generationCollect and analyze the historical traffic data to determine the traffic low periodBased on the analysis results and experience,expert make specific energy-saving plansDeliver the confi

183、guration strategy in the network management system or base station deviceRegularly evaluate the energy-saving effect,adjust and optimize according to the actual situationLevel 4 target state:RAN energy efficiency optimizationUsing intelligent technology to achieve accurate service prediction of the

184、base station,generate personalized energy saving schemes and automate execution can effectively improve the energysaving effect and reduce operating expenses.This includes collection of network data,analysis of service characteristics,generation of energy-saving schemes,dispatch of energy-saving ins

185、tructions and provide feedback optimization of execution results.ScenarioScenario Description22 Back to Contents|What is the AN Level 4 Industry Blueprint?TM Forum has also developed a Business Value Measurement(BVM)model to analyze the correlation between KEIs and KBIs and measure how changes in KE

186、I affect KBI improvements.For example,a CSP can look at the correlation between customer satisfaction(a KEI)and the customer churn rate(a KBI)to determine how to increase revenue.The table below shows KEIs and KBIs for select Level 4 scenarios,which we look at in more detail in Section 4,which focus

187、es on best practices for implementing these scenarios.AN Business Value Measurement modelKEIKBIAverage FactorScale Factor(Optional)Financial Factor1.Saved energyReduced O&M costs-Less electricity fees=Annual power consumption of a single device x Power saving rate of a single devicexDevice scalexEle

188、ctricity price2.Average duration per ticket-Lower outsourcing costs=Average maintenance duration per ticketxTotal number of ticketsxOutsourcing labor cost per unit timeNumber of employees who have transitioned-Shift of employees towards higher-value work=Average person-hour per ticketxTotal number o

189、f ticketsAnnual labor hours per person3.Service Delivery TimeRevenues from operations-Increased revenue from early provisioning and early sales=Number of provisioning days reduced for each userxAnnual average service provisioning volumexARPU(Revenue per user/month)4.Service availability Ratio-Increa

190、sed revenue from shorter interruption duration and less revenue loss=Reduced interruption duration per userxNumber of usersxARPU5.Total stimulated traffic-Increased revenue from experience improvement and stimulated traffic=Average stimulated traffic per areaxNumber of stimulated areasx Traffic reve

191、nue per unit6.Service availability Ratio-Increased revenue from quality improvement and SLA monetization=Availability x Annual durationxNumber of usersxARPU7.Precision marketing success rate-Increased revenue from precision marketing/Sales=Precision marketing success ratexTotal number of marketing u

192、sersxARPU8.Number of NaaS API callsRevenues from capabilities-Increased revenue from capability openness=Number of new APIsxAverage number of API callsxAPI unit price/call9.Number of major risksNumber of major faults-Major faults caused by major risksNumber of major risks10.MTTR of major faults-Majo

193、r faults that are not handled within the specified duration=Number of times that major faults are not handled within the specified duration11.One-time success rate of major changes-Major faults caused by major network changes=(1 One-time success rate of major changes)xNumber of major changesHigh-val

194、ue scenarios&KEIs#mappingService marketing7Service delivery 2/3Service assurance 2/4/5/6Complaint handling 2/4Network planning Network deploymentTrouble shooting2/4/9/10Network change11Quality optimization2/5/6Energy efficiency optimization1Resource managementTM Forum,2024Fault management23 Back to

195、Contents|What is the AN Level 4 Industry Blueprint?Evolution path to L4 autonomous networksTM Forum,2024Getting to Level 4 AN essentially means that CSPs operations move from an environment where systems assist humans to one where humans assist machines.EVOLVING TO LEVEL 4TM Forum recommends that th

196、e journey to Level 4 autonomous networks be divided into two phases as shown in the graphic below.The objectives,typical scenarios and key technologies of each phase have been defined in this preliminary model as a starting point for discussion and improvement by the entire telecoms industry.Getting

197、 to Level 4 AN essentially means that CSPs operations move from an environment where systems assist humans to one where humans assist machines.Many large operator groups are currently operating at or near Level 3 in at least some domains(see Section 3).This means they are deploying predictive AI and

198、 digital twin technology and are experimenting with GenAI.In the first phase of Level 4 deployment,operators will apply agents to ensure that the network makes decisions based on awareness information and instructs corresponding modules to take further action without manual intervention.Agents can b

199、e leveraged to solve complex problems,including network self-optimization and self-healing,without a high degree of human intervention.In phase 2,multiple cross-layer and cross-domain agents will collaborate to complete tasks like optimization of user experience and networks.Multi-agents can be diff

200、erent functional entities,such as network elements,autonomous domain management and control systems,and application programs.By collaborating,they facilitate distributed decision-making and a high degree of flexibility and scalability.Until Level 5 AN is reached,human-machine collaboration will be r

201、equired to support smooth network operations and ensure network security.The main objective of such collaboration is to smoothly switch between network control modes and minimize the impact of the sudden changes caused by the switching.AN Level 4 cannot cover all scenarios.For example,the system may

202、 not know how to resolve or mitigate conflicts between multiple automatic closed-loop objectives.At Level 5 when no humans are involved machines will generate their own intents and could decide to prioritize customer experience versus carbon footprint.But some operators may stop short of implementin

203、g Level 5 in their production networks because they may not be willing to exclude humans completely from service delivery and assurance processes.Next,we look at where CSPs are on their AN journeys.TYPICALSCENARIOSKEYTECHNOLOGIESL3(2024)L4 PHASE 1(20252027)L4 PHASE 2(20282030)Autonomous closed-Loopw

204、ithin a domain by agentEnd-to-end closed loop of complex scenarios by multi-agents collaborationPredictive AI are deployed at scale and copilots appliedCopilot will be deployed at scale and agents appliedAgents will be deployed at scale andapplications of multi-agent explored Digital twin Predictive

205、 AI Copilot LLM,etc.Network-convergent sensing Tool invoking AI agent Complex task planning AI agent,etc Decision-intelligent and responsible AI Network group task collaboration multi-agents Simulation verification/prediction,etcFocus on end-to-end complex scenarios E2E complaint handling and troubl

206、eshooting E2E network service/business operation Wireless network optimization and planning collaboration,etc.Focus on query,assistand analysis scenariosAssist analysis and query of typical faults,networkperformance monitoring and analysis,and signaling analysis Service assurance and complaint handl

207、ing for individual services,home broadband and private lines Fault management of network specialties Network change for core,IP and transport networks Quality optimization in wireless,IP and transport networksFocus on 20 selected high-value scenarios24 Back to Contents|Where are we on the AN journey

208、?Where are we on the AN journey?Many CSPs are making progress implementing autonomous networks.Several operators have stated publicly their ambition to achieve Level 4 AN in at least some domains and use cases over the next two years.The table on pages 28-30 provides a snapshot of AN implementation

209、based on contributions received from CSPs involved in the AN Project.Attained levels noted in the table are self-reported by the CSP and typically reflect a value for the operators network overall.But this doesnt offer a true comparison of operators progress.To enable comparison,the way AN levels ar

210、e assessed needs further standardization.TM Forums Autonomous Network Levels Evaluation Methodology(IG1252)helps CSPs self-assess their progress,but detailed scoring for specific domains and scenarios hasnt been fully standardized.The AN Project has been working to change this by enabling normalized

211、 comparison of operators AN levels.In the first half of 2024,TM Forum together with CSPs including AIS,Orange and Telefnica and suppliers such as Detecon,Ericsson and Huawei held a series of workshops to further define AN level evaluation methodology and develop a self-service assessment tool to gau

212、ge AN deployment.The group began with fault management in the RAN and core network,which CSPs have identified as key scenarios.As noted in Section 1,AN level scoring is based on an assessment questionnaire that evaluates cognitive activities in the network like intent handling,awareness,analysis,dec

213、ision and execution(IAADE),each of which contains a set of tasks to be measured.The questionnaire is based on the professional grading standards of SDOs such as 3GPP and ETSI with added weighting and detailed scoring criteria.The graphic below shows an example,with questions based on 3GPPs published

214、 standards in the RAN and core.Section 3AN level evaluation sample questionHigh-Value ScenarioCognitive Activity(IAADE)Service CapabilityWeightQuestionOption AOption BOption COption DAnswerRAN Fault ManagementAwarenessData collection5%Does the wireless network fault management system support automat

215、ic data collection in various fault scenarios?Fault scenarios include NE disconnection or out-of-service,cell faults,and fronthaul network faults.Yes.The system can automatically collect data(alarm,configuration,and performance data etc.)Partially.Data collection rules or templates are manually defi

216、ned.The system can automatically collect data(alarm,configuration,and performance data etc.)Partially.Manually select data collection items.The system can automatically collect data(alarm,configuration,and performance data etc.)No.People use the system to collect dataBAlarm filtering5%Does the wirel

217、ess network fault management system support automatic alarm filtering in various fault scenarios?Fault scenarios include NE disconnection or out-of-service,cell faults,and fronthaul network faults.Yes.The system can automatically filter out invalid or redundant alarms.Partially.Manually define alarm

218、 filtering rules.The system automatically filters out invalid or redundant alarms.Partially.People use the system filter out invalid or redundant alarms.No.People filter out invalid or redundant alarms.ATM Forum,2024The way AN levels are assessed needs further standardization,so TM Forums AN Project

219、 has been working on an evaluation tool to enable normalized comparison of operators AN levels.25 Back to Contents|Where are we on the AN journey?PILOT ASSESSMENT RESULTSDuring 2024,two dozen CSPs used the tool to assess their AN implementations of fault management in the RAN and core network,with p

220、lans to expand scenarios to include fault management and quality optimization in IP networks and network change in RAN and core networks.Some of the companies participating in the evaluations included AIS,Antel Uruguay,China Telecom,Elisa,Globe Telecom,IOH Indonesia,MTN Group,Orange,stc,Telecom Arge

221、ntina,Telefnica,Telkomsel,among others that wish to remain anonymous.MTN,Orange and Telefnica evaluated the scenarios in multiple countries.Preliminary AN level evaluation results were presented in June 2024 at DTW-Ignite after the first phase of the project.At this stage it was clear that the quest

222、ionnaires used to carry out the evaluations needed refinement to enable normalized scoring.Changes were made during phase 2 of the project,resulting in the publication of two standardized AN Level Evaluation Tool(ANLET)questionnaires for fault management on the RAN and core networks,respectively.Mos

223、t of the phase 1 participants and some new CSPs repeated evaluations using the new questionnaires.Anonymized results from both phases are presented below and on the next page.Two dozen CSPs have used TM Forums pilot assessment tool to evaluate their AN implementations.AN level evaluations for fault

224、management in the RANSRILANSRILAN KAKAPHILIPPINESPHILIPPINESAsiaEuropeLATAMMEA0.001.002.003.004.00TM Forum,20245.00Asia 1Asia 4Asia 5Asia 6Europe 1Europe 2Europe 4Europe 5Europe 6Europe 7Europe 8Europe 9Europe 10Europe 11LATAM 2LATAM 3MEA 1MEA 2MEA 3MEA 4MEA 5MEA 8MEA 7MEA 9n n AN level(phase 1)n n

225、AN level(phase 2)26 Back to Contents|Where are we on the AN journey?The average ANL score for fault management in phase 2 was 2.5 in the core network and 2.8 in the RAN.Core network stability was also assessed,with a higher average ANL of 3.5.Overall,AN levels achieved remain relatively low consider

226、ing the commitment of some of these operators to reach Level 4 by 2025,with most scores clustered between Level 2 and Level 3.Leading operators have outlined clear business objectives,capability goals and the evolution path toward Level 4 for high-value scenarios(see Section 7).Other operators can r

227、eference these examples to accelerate their development of Level 4 capabilities.AN level evaluations for fault management in the core networkSRILANSRILAN KAKAPHILIPPINESPHILIPPINESAsiaEuropeLATAMMEA0.001.002.003.004.00TM Forum,20245.00Asia 2Asia 3Asia 5Asia 6Europe 1Europe 2Europe 4Europe 5Europe 6E

228、urope 7Europe 8Europe 10LATAM 2LATAM 3MEA 1MEA 2MEA 3MEA 4MEA 5MEA 6MEA 10LATAM 1n n AN level(phase 1)n n AN level(phase 2)27 Back to Contents|Where are we on the AN journey?CSPs AN strategies and progressCSPTM Forums ODA,AN framework and AN level taxonomy are the foundation of AISs Cognitive Techco

229、 strategyAN strategyLevel 4 high-value focusProgress and plans Core network fault management Wireless network optimization:customer experience driven Optical and transport service assuranceIs participating in TM Forums AN level assessment pilot;achieved L3 in the two targeted processes(RAN and core

230、fault management);results include:CX optimized by 10%Traffic loss reduced by 90%in extreme scenarios Network efficiency improved by+20%Is extending AN practices to network optimization,configuration,planning,change and other processes.China Mobile was the first CSP to set a goal to reach L4 on a lar

231、ge scale in its production network;technology function is centralized,and the company is using end-to-end service operation layer across domains to drive overall development of technical AN capabilities Service delivery Network change Energy-efficiency optimization Service-quality optimization Troub

232、leshooting Complaint handlingAchieved overall AN L3.2 in 2023;results include:Significant improvement in customer satisfaction 80%reduction in the number of major network faults 20%increase in average number of sites maintained per person Savings of 3,200 person-years of human labor and 4 billion Kw

233、h of electricity Contribution to 70 AN technical standards and 15 whitepapers Has launched a telecoms LLM with 13.9 billion parameters,trialing applications like data self-service and AI chatbots.China Telecom is guided by 3-ultra Vision,considers autonomous cloud-network operations as an important

234、stepping stone for the transformation toward network+cloud 5G core network:complaint handling IP network:network changes(intelligent online simulation)Innovative practice of network LLM in network maintenance scenarioIs participating in AN level assessment pilot.Achieved overall AN L3+in 2023;result

235、s include:Return on equity up from 6.41%in 2022 to 6.93%in 2023 Savings of 1 billion kilowatt-hours of electricity annuallyIn 2024 aims to achieve L4 in 40%of high-value scenarios and will promote large-scale application of its LLM in the network,creating thousands of digital employees.China Unicom

236、aims to implement autonomous systems across all stages of planning,development,maintenance,optimization and operations to enable digital operations,intelligent scenarios and agile services.Optical private lines service assurance Intelligent planning and construction Intelligent energy savings in bas

237、e stations Closed-loop control of optical cable risksIs performing ANL evaluations for 5 networks across 8 provinces focusing on high-value scenarios using GenAI;results include:1.5 million customer complaint diagnoses per month with an 80%one-time problem resolution rate Intelligent energy-saving s

238、ystems cover 98%of base stations,saving over 40 million kilowatt-hours monthly Open platform aggregates 2,800 capabilities,utilized over 38 billion times per monthWill focus on additional high-value L4 AN scenarios in the next two years.28 Back to Contents|Where are we on the AN journey?CSPs AN stra

239、tegies and progressCSPChunghwa Telecom has been working on closed-loop automation for specific network scenarios since 2019.AN strategyLevel 4 high-value focusProgress and plansSince 2023,the company has expanded AN operations management to enhance customer experience and reduce operational costs;re

240、sults include:Conducted ANL assessments for fixed and mobile broadband services,achieving L3 autonomy Developed an automated fixed broadband service rate upgrade mechanism,benefiting 1.16 million usersPlans to apply AN to mobile broadband and IP transport networks,aiming to achieve L4 by 2025.Fault

241、management in RAN and IP transport domains Energy-efficiency optimization in the RAN MTN is implementing AN as part of its Ambition 2025 Strategy,which includes a technology framework called PACE(Platform,Agile,Connectivity and Experience).All MTN operating companies aligned with the companys AN str

242、ategy and target architecture,which is based on TM Forums AN framework.Cross-domain network fault management IP and optical network autonomous O&MParticipated in the AN level assessment pilot and will continue to align to the assessment for RAN,core,IP and change management.In 2023,MTN released its

243、AN Blueprint 2.0,establishing ANL standards and developing a comprehensive use case library.In 2024,the company plans to design Version 3 of the blueprint,identifying high-value scenarios to finalize assessment criteria,effectiveness targets,processes and architecture for AN L4.Singtel initiated its

244、 AN program in 2023,fostering group-level collaboration to benchmark,share best practices and apply them across the group.Designated fellow Bridge Alliance operator AIS as an AN coach to standardize the four key elements of the AN Framework and support operating companies in L3 and L3+use-case deplo

245、yment.In 2024,will create its AN L4 Blueprint to transition from automation to autonomy.Fault management and customer-complaint management RAN energy saving and network optimizationHKT aims to bring the right digital solutions to society to accelerate growth and create a sustainable future for all.W

246、ireless network optimization focusing on energy usage according to data traffic using AI technology Optical and transport service to minimize energy consumption and improve energy efficiency through AI technologyIs participating in TM Forums AN level assessment pilot.Focus in 2023 was on reducing 20

247、%of HKTs total energy consumption(compared with 2018 baseline),and HKT plans to further cut energy consumption this year and in the future using innovative methods over its AN.29 Back to Contents|Where are we on the AN journey?CSPs AN strategies and progressCSPTelefnica launched a program called Aut

248、onomous Network Journey(ANJ)in 2021 to increase autonomy and transform its architecture,management of data,and organizational structure to support AN;ANJ is being applied in all operating companies in Europe and Latin America,but deployment progress varies among OpCos.AN strategyLevel 4 high-value f

249、ocusProgress and plansTelkomsel adopted AN as part of its digital innovations,aiming to reach L4 by 2025.Ensuring high service quality during peak events.Transport network(fronthaul optical link):fault predictive maintenance Wireless network:fault management(cross-domain)Wireless network:optimizatio

250、nIs participating in AN Level assessment pilot.In 2023,Telkomsels cross-domain experts,guided by the Networks Quality Digitalization Department,made significant progress towards L3.Notably,the self-assurance use case for massive events was recognized at TM Forum DTW2023-Ignite and deployed across 37

251、 sites in Indonesia.Is participating in the AN level assessment pilot.Anticipates several key improvements in achieving AN Level 4 including:AI-powered operations assurance,where closed-loop automation will prevent service disruptions.Real-time optimization and decision-making,where automated system

252、s will manage network configurations,including dynamic features like network slicing and NaaS capabilities.Efficient planning and deployment,where real-time inventory management,scenario-based planning(e.g.,what-if analysis)and AI-driven optimization will streamline deployment processes.TM Forum,202

253、4Is focusing on getting to AN Level 4 for multiple processes(O&M,optimization/configuration,testing,deployment and planning)in multiple domains(radio access,transport,IP,core,IT cloud,and fixed access networks).The company is placing high emphasis on O&M in all domains and on RAN optimization,IP dep

254、loyment and transport planning.Telecom Argentina drives its strategy of developing ANs and contributing to TM Forums best practices through its Autonomous Network Transformation Office(ANTO).Wi-Fi optimization Core network:network optimizationIs participating in AN level assessment pilot.Has committ

255、ed to achieving its goals through a two-stage approach,with the first focusing primarily on developing technical use cases and the second aiming for more balance between technical and business domains.stc has a goal to achieve ANL 4 in 2025 in some scenariosIs participating in TM Forums AN level ass

256、essment pilot.2024 focus was on automatic network change and AIOps capabilities including:Digital twin-based cloud-network visualization Intelligent signaling storm simulation and emergency assistance Pipeline-based network change and deployment automationResults included:Fault management-CNF and cl

257、oud health status unified visualization,moving from 5+portals to 1 and cutting fault-demarcation time from 2 hours to 15 minutes Network changes and deployment:ultra-security(automatic execution by unified pipeline reducing 80%of manual breakpoints);high-efficiency(deploying 200+test cases in minute

258、s);and full scenarios for both telco-cloud and core networkIn 2025,stc plans to introduce foundational models for alarm analysis,an intelligent copilot and a complaint handling expert.Core network:fault management,network changes,design and deployment30 Back to Contents|Best practices for implementi

259、ng AN high-value scenariosBest practices for implementing AN high-value scenariosSection 4In 2022 as CSPs began to gain some experience implementing autonomous networks,we added a section to the journey guide to highlight their best practices.Early on,these focused primarily on developing an AN visi

260、on and target architecture and then progressed to include development of KEIs and methods for evaluating AN levels.Many of the best practices have been incorporated into TM Forums AN standards.Indeed,the previous section outlines some best practices for designing the target-state for Level 4 scenari

261、os,which have been added to the AN Level 4 Industry Blueprint.This year we received best practice contributions from AIS,China Telecom,China Unicom,Singtel Group and its operating companies,stc,Telecom Argentina,and Telkomsel.All are available to read in their entirety on the AN Project website.Here

262、,we highlight some in the following areas,which are important to implementing AN:Business value measurement CSPs main goal in adopting AN is to increase business value and operational efficiency.To do that figuring out how to evaluate and quantify the results of implementing AN is key.Reference arch

263、itecture development some CSPs are using TM Forums AN standards to define their own target network architectures aiming for AN Level 4.Collaboration between a telco group and its operating companies telco operating companies need to be able to quickly implement AN strategy communicated by the group,

264、and this requires efficient collaboration.Use case selection selecting AN Level 4 use cases and scenarios is an iterative process.CHINA MOBILES LEVEL 4 AN PRACTICE AIMS FOR LIGHTS-OUT FACTORYChina Mobiles networks and operations are vast,with 31 local operators serving 1 billion mobile customers and

265、 309 million wireline broadband customers as of June 2024.An autonomous networks pioneer,China Mobile has set a new Level 4 goal focusing on high-value scenarios,with an aim to create an end-to-end,fully automated“lights-out factory”pipeline for digital and intelligent O&M.The core concept of the li

266、ghts-out factory is to drive end-to-end full automation in every high-value scenario identified by China Mobile.Traditional O&M processes typically involve multiple steps,where each is partially automated through individual modules.These modules are then combined to complete the entire task.In the l

267、ights-out approach(shown below)each step is automatically completed,and data flows seamlessly between steps,enabling efficient collaboration without any human intervention.This enhances overall operational efficiency,reduces human-induced errors,improves service quality and lowers operational costs.

268、3199%Timeliness of complaint handling 95%O&M workforce structure Continuously optimizingReplacing human labor with machines 5,000 person-yearsReducing volume of frontline dispatch tickets 5%-10%Reduction of equipment energy consumption 5 billion kWhReducing network maintenance costs Continuously red

269、uceLegend:Automated/intelligent stage Onsite construction stage Automatically transfer to the next stageInstallation and maintenance stationRequirements reviewCustomer response centerResource allocationCustomer response centerData productioninputService commissioningDelivery reviewInstallation and m

270、aintenance stationUser demandCustomer satisfactionoutputThrough automation of O&M processes,China Mobile aims to enhance safety and quality,strengthen customer experiences,and improve efficiency and effectiveness:Enhancing safety and quality requires continuously improving customer satisfaction and

271、reducing the number of faults.Strengthening customer responsiveness involves continuously improving the timeliness of service activation,enhancing the timeliness of complaint handling and continuously optimizing the workforce structure to transition more staff towards customer service support roles.

272、Improving efficiency and effectiveness,while ensuring the stable operation of the network,involves gradually replacing human labor with machines,reducing the volume of frontline dispatch tickets,maximizing the reduction of equipment energy consumption,and lowering outsourcing costs.The graphic below

273、 shows expected improvements for each of these goals.China Mobiles approach to end-to-end,fully automated O&MTM Forum,2024TM Forum,2024China Mobiles expected O&M improvements32 Back to Contents|Best practices for implementing AN high-value scenariosThe sheer size of China Mobiles network means its O

274、&M scenarios are diverse.This makes reaching Level 4 AN a huge and complex undertaking requiring a systematic approach that can be measured.The graphic below shows the step-by-step approach China Mobile is taking,and a description follows.Autonomous networksDigital transformation of network O&MMeasu

275、re 1:Value drivenEffectiveness planningSelected high-value scenariosMeasure 2:Concept and process optimizationConcept upgradeCreating valueSimplify processLeveraging capabilities in advanceEliminate process breakpointsMeasure 3:AutomationMeasure 4:AI+innovationMeasure 5:Data drivenKnowledge manageme

276、ntEndogenous network elementNetwork management automationLarge model innovationSmall model promotionResource data governanceMetrics data governanceOrganizational change and talent developmentChina Mobiles step-by-step approach to the lights-out factoryTM Forum,20241.Driven by value with dozens of bu

277、siness types and five major network domains,China Mobiles first step was to comprehensively re-mapped the full panorama of O&M scenarios,which includes more than 100 end-to-end scenarios.Next,it established a three-tier value evaluation system and selected high-value scenarios based on their value r

278、anking.A“pilot-first-then-promote”approach was adopted,ultimately leading to comprehensive implementation across all scenarios.2.Concept and process optimization China Mobile aims to shift its O&M philosophy from being device and management centric to become customer and business centric.The key is

279、to use end-to-end processes as the breakthrough point,integrating new concepts and delivering new value.Through continuous process optimization,China Mobile plans to further simplify ineffective steps,reduce resource waste and time consumption,and eliminate process breakpoints and bottlenecks.This e

280、nsures that data and tasks flow automatically,fully leveraging capabilities in advance to avoid resource idleness and waste and optimizing resource allocation.33 Back to Contents|Best practices for implementing AN high-value scenariosWhen implementing autonomous network scenarios,CSPs must consider

281、what business value they can deliver and whether they can lead to a return on equity.3.Automation to build an end-to-end,fully automated O&M production pipeline,it is necessary to continuously improve the O&M knowledge system and expert rules for automation.It is also essential to continuously enhan

282、ce the inherent capabilities of network elements and network management systems.This involves constructing a cross-domain,cross-vendor and cross-system production line that is fully automated end to end.This improves the efficiency and service quality of high-value scenarios.4.AI+innovation with AI

283、technology as the innovation engine,China Mobile aims to transform O&M from an auxiliary tool used enhance quality and efficiency into an indispensable core capability.The company is working to strengthen the collaborative and integrated innovation of both large and small AI models,which involves vi

284、gorously promoting benchmark applications of small AI models to fully realize their value at scale,as well as innovatively developing O&M AI agents.These agents will unify the scheduling of large and small models and various network management system capabilities,reshaping key production processes i

285、n network operations and maintenance,and enhance the intelligence level in high-value scenarios.5.Data driven to meet the requirements for data accuracy,completeness and consistency in network O&M scenarios,it is essential to focus on strengthening network data governance across the entire network,e

286、mbedding high-quality data into production processes to support operational analysis and decision-making.This will allow China Mobile to shift O&M from being experience driven to become data driven.CHINA TELECOMS AN BUSINESS VALUE ASSESSMENTWhen implementing autonomous network scenarios,CSPs must co

287、nsider what business value they can deliver and whether they can lead to a return on equity(ROE).China Telecom has been considering this in its cloud-network operations.ROE reflects the ability of an enterprise to use its net asset value to generate profits.The idea is to focus on the quality of cap

288、ital return and the business value of improved operating performance,and to pay more attention to input-output efficiency.China Telecom used the DuPont analysis method,which is a framework for analyzing a companys performance originally developed at the DuPont Corporation.China Telecom started by de

289、composing the key factors that affect ROE.Then,the company clarified three business value directions:Lean cost management Cloud-network quality improvement and revenue generation Efficient asset operation.The value directions were then decomposed into six specific requirements as shown in the graphi

290、c on the next page,and 16 indicators were added.Some examples of the indicators include smart electricity meter collection coverage,automatic fault identification rate,identification rate of potential pitfalls,service-provisioning time,automation rate of service inspection,automatic customer complai

291、nt-handling rate.34 Back to Contents|Best practices for implementing AN high-value scenariosTM Forum,2024China Telecoms ROE process for cloud-network operationPromising results China Telecoms results are promising.The company reports boosting ROE from 6.41%in 2022 to 6.93%in 2023.This improvement su

292、pported significant revenue generation from B2B 5G services,home broadband and private lines,while saving over 1 billion kilowatt-hours of electricity annually.Now,China Telecom is beginning to evaluate potential ROE improvement from the introduction of its network large language model(LLM),which wi

293、ll create thousands of digital employees in various roles.The company believes it will bring major changes in O&M,operations processes,systems integration and cloud-network capabilities by transforming NOC personnel,reducing the workload of external on-site installation and maintenance,and reducing

294、labor costs.Overall,China Telecom is aiming to achieve L4 AN in 40%of high-value scenarios in 2024,such as on-site O&M,cross-domain troubleshooting,hazard prediction and end-to-end service assurance.AISS AN LEVEL 4 ARCHITECTURE EVOLUTIONThailands AIS is embracing a“Cognitive Techco”strategy to enhan

295、ce network intelligence and automation in alignment with TM Forum standards.The company is developing an AN Level 4 target architecture that focuses on two processes customer complaint management and network fault management to improve customer satisfaction and operational efficiency.AIS has partici

296、pated in TM Forums AN level assessment pilot,achieving L3 in both processes.AISs AN L4 architecture shown on the next page embodies:A service layer with a“1+2+N”model that enhances cross-domain digital twin and LLM capabilities and facilitates cross-domain collaboration A resource layer that include

297、s MCAID(manager,controller,analyzer,AI inference/LLM and digital twin)for single-domain autonomy Multi-vendor compatibility,convergent data integration,DevOps methodology,openness,near real-time analysis,simulation and intelligence Effective single-domain autonomy and cross-domain collaboration35 Ba

298、ck to Contents|Best practices for implementing AN high-value scenariosFeatures of AISs Level 4 target architecture include:Service layer addresses challenges posed by siloed systems,provides converged data subsystems and supports multiple centers.The convergent data platform enables data access thro

299、ugh APIs,fostering business layer applications.Service layer interfaces with vendor-specific element and network management systems(EMS/NMS),centralizing data and enabling cross-domain collaboration based on digital-twin and LLM.Resource layer enhances data analysis with AI/LLM capabilities,achievin

300、g closed-loop processes based on the digital twin within a single domain and simplifying integration with the service layer.Digital twin and LLM make networks visible,perceivable and controllable,reshaping network O&M and enabling AIS to evolve from automation(+AI)to autonomy(AI+).Promising resultsA

301、IS has reported encouraging results overall from its evolving AN practices including a 10%optimization of customer experience,a 90%reduction in traffic loss in extreme scenarios and a+20%network efficiency improvement.For example,for RAN burst traffic caused by crowd gathering such as at sporting ev

302、ents or in vehicle traffic jams,the system quickly detects traffic bursts in minutes and balances traffic between burst cells and neighboring cells through adjust antenna parameters,improving the edge user experience by 10%.Antenna parameter adjustments in neighboring cells also make it possible to

303、provide minute-level cell outage detection and compensation,which has helped AIS reduce traffic loss by 90%.MService LayerBusiness Layer2C2H2BResource Layer(vender X)Resource Layer(vender Y)AnalyzerControllerManagerDomain APIs(Service-oriented API)GUIAPI GatewayServiceOrch.Fault MgmtWork ForceResour

304、ceMgmtConfigMgmtQualityMgmtN CentersCenterDevOpsAI Engine(Machine Learning&LLM)Convergent Data2 Platforms1 EngineACAPIsGUIRANIPOpticalCoreAccessRANIPOpticalCoreAccessAI InferenceML/LLMIScenario CapabilityBasic Orchestration CapabilityETLData martCore 360RAN 360Data CatalogDigital TwinsDigital TwinsD

305、igital TwinsAPI&DataAPI&DataData Governance?Resource Layer(Vendor X)?Service Layer?N Centers2 Platforms1 EngineBusiness Layer2C2H2BAPI&DataRANIPOpticalCoreAccessAPI&DataAISs Level 4 target architecture evolutionTM Forum,202436 Back to Contents|Best practices for implementing AN high-value scenariosF

306、inally,AIS is leveraging AI to improve the accuracy and efficiency of fault identification and root-cause diagnosis,and to improve customer-complaint prediction,resulting in efficiency improvement of more than 20%.TELECOM ARGENTINAS TARGET ARCHITECTURE DEVELOPMENTTelecom Argentina has established a

307、AN Transformation Office(ANTO)to ensure success of its AN evolution.This team is evolving the companys AN framework,roadmap and target architecture.ANTO has committed to achieving its goals through a two-stage approach.The first stage focuses primarily on developing use cases from a technical perspe

308、ctive,while the second stage aims for greater balance between the technical and business domains,plus other domains within the company where comprehensive use cases are being addressed.See page 40 for more on the companys best practices for selecting use cases.Telecom Argentinas AN target architectu

309、re(shown below)seeks to streamline operations and offer a completely autonomous network ecosystem.It includes three layers:The business layer focuses on providing an optimized user experience for telecom services through network exposure.The service layer focuses on service orchestration and automat

310、ion,enabling multi-domain network interoperability.In the resource layer,an autonomous domain exposes a range of business services based on the capabilities of the objects it governs;it is self-managed,self-scaling and manages its own lifecycle.?ANDEEXAW?ANDEEXAW?ANDEEXAW?ANDEEXAW?ANDEEXAW?ANDEEXAW?

311、ANDEEXAW?ANDEEXAW?ANDEEXAW?AI/MLExecutionAwaresessAnalizeDecideAvailablePartial/Analysis?Telecom Argentinas reference architectureTM Forum,202437 Back to Contents|Best practices for implementing AN high-value scenariosTELKOMSELS AN LEVEL 3.5 ARCHITECTURE DESIGN Telkomsel is using TM Forums AN Framew

312、ork and best practices to accelerate the development of autonomous networks in Indonesia.The company,which has a goal of achieving AN Level 4 by 2025,outlined its AN Level 3.0 architecture last year,focusing on fault management.Now,Telkomsel is highlighting its AN Level 3.5 architecture design,which

313、 introduces a Resource Engine to attain autonomy within a single domain and enhance collaboration across domains at the service level.The graphic below shows the architecture with the Resource Engine at the bottom bolstering the Digital Engines closed-loop capabilities.An explanation of some of the

314、key elements of the Resource Engine follows.?Telkomsels AN Level 3.5 architectureTM Forum,2024Scenario applications the objective in each scenario is to increase and improve intelligence and automation.For example,intelligent analysis performed on alarm data has the ability to discern underlying ala

315、rms and offer intelligent recommendations for resolving them.Network management,controller and analysis utilizing a built-in AI Engine(see below),control components execute automated operations and conduct intelligent governance and analysis on data gathered across various scenariosData awareness ne

316、twork elements provide the element management system(EMS)with a variety of data,including performance and alarm information,via standard southbound interfaces,such as the PMS and YANG interfaces.Built-in AI Engine and Domain API standard southbound interfaces and real-time data collection allow the

317、Resource Engine to visualize network-level data online.By reporting output to operations support systems(OSS)via the northbound API,resource-layer applications can improve automatic root-cause locating and intelligent cross-domain issue diagnosis.38 Back to Contents|Best practices for implementing A

318、N high-value scenariosSINGTEL GROUP AND OPCOS AN PRACTICE COLLABORATION Singtel Group owns or has a stake in several operating companies(OpCos)in the Asia-Pacific region including Optus in Australia,Bharti Airtel in India(and Africa),Globe Telecom in the Philippines and AIS in Thailand.Singtel and i

319、ts OpCos are also part of the Bridge Alliance,a mobile alliance with 35 member operators collectively serving more than 1 billion customers in Asia-Pacific,the Middle East,Africa and Europe.Based on TM Forums AN Framework and methodology,Singtel and its OpCos have developed an innovative approach to

320、 group-level collaboration that helps to create benchmarks,exchange best practices and experiences among OpCos,and apply best practices across the group.Singtel Group selected AIS,which is a leader among the OpCos in AN practice for incident management,as AN coach to systematically lead the other op

321、erators to align on the definition of the four key elements of the AN Framework:the AN Map,AN levels,KEIs and target architecture.AIS also supports the OpCos on their AN journeys comprising AN level assessment,gap analysis,solution design and implementation as shown in the graphic below.An explanati

322、on of the collaboration process follows.OpCos/Associates：AN JourneyAssessment&PracticeGroup：4 Key Elements Standard Definition&GuidanceAN MapProcess Selectionxx Value ProcessKEI Framework Definition/SampleANL Definition&SampleTarget ArchitectureDefinition&Samplexx Value ProcessImplementationUC Trial

323、&DeploySolution DesignDevelopANLAssessmentContinuousEvolutionGap Analysis Singtel Groups approach to collaborationTM Forum,2024The teams started by collaborating on alignment of the four key elements with guidance from the AN coach:Value process selection a common high-value scenario for the inciden

324、t management process is identified and agreed on as a pilot scenario across the group to achieve AN Level 3.ANL definition the OpCos worked together to define what Level 1 to Level 4 looks like for each process and then adopted these agreed definitions as a unified blueprint across the group.Target

325、architecture the group defined a three-layer target architecture,while the OpCos developed their own architectures based on definition and AN strategy.KEI framework a KEI framework was defined collaboratively and adopted as a blueprint across the group to guide OpCos in formulating business value im

326、provements through enhancement of AN practices measured by KPIs.39 Back to Contents|Best practices for implementing AN high-value scenariosThen,the teams can progress through AN journey with support from the AN coach:ANL assessment the OpCos perform their assessment based on their agreed ANL definit

327、ion.Here the AN coach moderates the ANL scores based on the supporting reasons presented by each OpCo to ensure fair and consistent assessment.Gap analysis and solution design benchmarking is done across the group to perform the gap analysis for each OpCo against the AN coach of each operation task.

328、Best practice and experience from the AN coach and any OpCos that score well in their AN level assessments are exchanged with other OpCos as capability references so that they can plan their target AN level,identify gaps and design solutions.Solution Implementation based on the reference design,the

329、lead OpCo pilots the solution deployment and shares the results with other OpCos,who can then use the pilot insights to speed up their own deployment.Through this collaboration approach,OpCos can benefit from the groups collective experience and apply proven techniques to achieve quick wins.As an ex

330、ample,AIS shared their implementation of a use case on RAN CPRI risk prevention and prediction as a reference to the group.(CPRI,which stands for Common Public Radio Interface,defines the key internal interfaces between radio equipment and radio equipment controllers in base stations.)Telkomsel then

331、 picked up this use case for implementation using the reference from AIS.As a result,Telkomsel was able to shorten its solution design and implementation time significantly.TELECOM ARGENTINAS AN USE CASE SELECTION AND DEVELOPMENTTelecom Argentinas AN Transformation Office,ANTO,has adopted an Agile a

332、pproach to accelerate network transformation.The implementation of this methodology aims to promote the deployment of prioritized use cases developed collaboratively among various departments in the company.Telecom Argentinas process involves an iterative design of minimum viable products(MVPs)and t

333、heir increments.The stages are shown in the graphic on the next page,and a description of each follows.Through a collaboration approach,OpCos can benefit from the groups collective experience and apply proven techniques to achieve quick wins.40 Back to Contents|Best practices for implementing AN high-value scenarios1.Discovery in the first phase the owners of the different technological domains co