《凱捷：2024年自智網絡研究報告-電信業為何及如何加速其自智網絡之旅（英文版）（80頁）.pdf》由會員分享，可在線閱讀，更多相關《凱捷：2024年自智網絡研究報告-電信業為何及如何加速其自智網絡之旅（英文版）（80頁）.pdf（80頁珍藏版）》請在三個皮匠報告上搜索。

1、Why and how the telecom sector should accelerate its autonomous networks journeyNetworks with intelligence#GetTheFutureYouWantWhat are autonomous networks?Autonomous telco networks are networks that,under most conditions,can self-configure,self-monitor,self-optimize,and self-heal.As there are varied

2、 definitions and terminologies in use,this report takes TM Forums autonomous network maturity model as the industry benchmark.This model defines six levels of autonomous networks:L0 Manual management:The system delivers assisted monitoring capabilities,which means all dynamic tasks must be executed

3、manually.L1 Assisted management:The system executes a certain prefigured repetitive sub-task to increase efficiency.L2 Partial autonomous network:The system enables closed-loop operations and maintenance(O&M)based on an artificial intelligence(AI)model for certain units in certain environments.L3 Co

4、nditional autonomous network:Building on L2 capabilities,the system can sense real-time environmental changes,and,in certain network domains,optimize and adjust its operation to the external environment to enable intent-based closed-loop management.L4 Highly autonomous network:Building on L3 capabil

5、ities,in a more complicated cross-domain environment,conducts analysis and makes decisions based on predictive or active closed-loop management of service-and customer-experience-driven networks.L5 Fully autonomous network:The goal for telco network evolution.The system possesses closed-loop automat

6、ion capabilities across multiple services,multiple domains,and the entire life cycle.2Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeySource:TM Forum,Autonomous Networks:Empowering Digital Transformation For

7、The Telecoms Industry,May 2019.FIGURE 1.TM Forum Autonomous Networks Levels and VisionPPPPPP/SPPPPSP/SPPPSSP/SP/SPSSSSP/SSSSSSN/ASelect scenariosAll scenariosP:Personnel,S:SystemsExecutionAwarenessAnalysisDecisionIntent/ExperienceApplicabilityLEVEL DEFINITIONAUTONOMOUS NETWORKS LEVELSL0:ManualOperat

8、ion and MaintenanceL1:Assisted Operation and MaintenanceL2:PartialAutonomousNetworkL3:ConditionalAutonomousNetworkL4:HighlyAutonomousNetworkL5:FullyAutonomousNetwork3Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks j

9、ourneyIn this report,the term telco is used synonymously with telecom operators or communication service providers,including,but not limited to,full-service operators(fixed,mobile,and broadband),mobile operators,broadband providers,cable broadband,and mobile service providers.Source:TM Forum,Autonom

10、ous Networks:Empowering digital transformation evolving from Level 2/3 towards Level 4,September 2023.ICTSmart cityZero-waitSwiftSimplifiedSelf-healingLeanAgilePreventiveZero-touchZero-troubleSelf-servingAs a ServiceAs a PlatformSelf-fulfillingSelf-assuringZERO-XEXPERIENCESELF-X Operations and Maint

11、enanceSmart industrySmart governmentSmart XX.LaunchDeliveryCarePlanningOrderingMarketingOperationDevelopment MaintenanceOrganizingOrchestration ConfigurationInfrastructureService BusinessMonitoringHealingOptimizingAUTONOMOUS NETWORKS VISION4Capgemini Research Institute 2024Networks with intelligence

12、:Why and how the telecom sector should accelerate its autonomous networks journey5Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyExecutive Summary Here are the key findings from our research on autonomous ne

13、tworks comprising a primary survey of 435 telco industry professionals,and in-depth discussions with 22 senior executives overseeing their organizations autonomous network initiatives(more details on research methodology are in the Appendix):TM Forums autonomous network levels current and future(202

14、8)9%42%42%6%1%0%0%End-2028(projected)End-2023Percent oftelcos in thecategory8%31%44%16%1%L0:Manual operation and maintenanceL1:Assisted operation and maintenanceL2:Partial autonomous networksL3:Conditional autonomous networksL4:Highly autonomous networksL5:Fully autonomous networks1.The telecom sect

15、or has just begun its autonomous networks journey,with a majority of telcos aiming to attain Level 3 autonomy or above by 2028.Today,a large majority of telcos(84%)are either at Level 1 or Level 2 autonomy for their overall networks Source:Capgemini Research Institute,Autonomous Telco Networks Surve

16、y,December 2023January 2024,N=113 large CSPs.6Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyExecutive Summary Leadership:Fewer than one in five organizations(15%)has appointed a dedicated leader for autonom

17、ous networks,and 31%are in the process of finding one.Roadmap:50%of telcos have a roadmap that covers only the next 1-2 years.3.Telcos that are moving faster on their autonomous networks journey are reaping significant benefits.The benefits from autonomous networks are clear:Telcos have realized a 2

18、0%improvement in operational efficiency and 18%network operational expenditure(OpEx)savings,on average,through autonomous networks initiatives undertaken in the past two years.71%of telcos have reduced energy consumption in this period,and telcos expect to reduce emissions by 32%on average in the ne

19、xt five years.The financial benefits of autonomous networks are expected to reward investment.(including operations),as defined by TM Forum Autonomous Network levels(see table above).The majority of telcos(61%)aspire to reach at least Level 3 autonomy by 2028.The journey will be long and complex,how

20、ever,as only 1%expect to attain Level 5 and 16%to attain Level 4 overall(including operations).Use cases:Most of the autonomous networks use cases are at the proof of concept(POC)or pilot stage,and about one in ten telcos is deploying these use cases at partial or full scale.2.Several organizational

21、 and technological challenges threaten progress on autonomy.Key challenges:Telcos cite employee mindset,integration issues,and data sovereignty concerns as top challenges hindering adoption.Strategy:Only about one in five telcos(17%)has a comprehensive autonomous networks transformation strategy wit

22、h well-defined goals and target timelines.7Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyExecutive Summary Telcos expect to invest$87 million on average in autonomous networks over the next five years.Our a

23、nalysis suggests that OpEx savings from autonomous networks would be$150 million$300 million per organization over the coming five-year period.Estimated return on investment(ROI)of 1.7x3.4x and a payback period of 2.9 to 1.5 years in conservative and optimistic scenarios,respectively.Telcos with a c

24、omprehensive strategy and advanced implementation of autonomous networks are seeing larger benefits than organizations without.We classified surveyed organizations into leaders,fast-followers,and beginners based on the maturity of their autonomous networks strategy and roadmap,planned investments,cu

25、rrent level of network autonomy,and implementation of various use cases.*Over the past couple of years,telco leaders achieved a cumulative 19%reduction in network OpEx,outpacing beginners who managed a 13%reduction.*Leaders have also demonstrated a 22%improvement in time-to-market for new services d

26、ouble that of beginners(11%).4.Telcos accelerate and sustain their autonomous networks journeys by:devising a strategy and roadmap for the autonomous networks journey;formulating the right talent strategy to support the scale-up of network autonomy;investing in key technology and data/AI initiatives

27、;setting the pace of their transformation in line with networks maturity;and innovating continuously to remain at the cutting edge of autonomous networks development.8Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks

28、journeyWho should read this report and why?This report provides insights into the autonomous networks transformation to leaders in the telecom sector working in various fields of specialization including network,operations,engineering,data and IT,finance,and strategy.Among network leaders,chief tech

29、nology officers,chief network officers,and heads of autonomous network and digital networks,and their teams will find it useful.Leaders from finance and strategy and top management at telcos will find it useful.Additionally,given the central role of data and artificial intelligence(AI)in autonomous

30、networks,this report will inform,chief data officers,chief AI officers,heads of analytics/AI/data,and chief information officers.Finally,considering the potential of autonomous networks to drive energy efficiency and carbon footprint reduction,sustainability leaders will also find it useful.This rep

31、ort is based on the findings of an industry survey of 435 senior executives(director level and above)from telcos,network equipment providers(NEPs),and hyperscalers.All organizations had annual revenue above$1 billion.About 51%of executives were from network;24%from data/AI and IT;15%from top managem

32、ent;and 10%from engineering.See the Research Methodology at the end of the report for more details.9Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyIntroduction The telecom sector eagerly anticipates the emer

33、gence of the era of autonomous networks,with self-sufficiency to configure,monitor,and“heal”themselves.However,its arrival will not be straightforward.When telecom networks started out,they were largely reliant on manual systems and processes.They gradually built in software-driven rules and automat

34、ion,although these were mostly reactive in nature(see figure 2).Today,the increasing use of data analytics and machine learning(ML)allows the network to take certain low-risk actions proactively or prescribe a series of actions to a human operator.Given recent advances in the field of AI,this is exp

35、ected to change.The promise of a self-serving,self-fulfilling,and self-assuring network,that is able to react autonomously to changing conditions to maintain or enhance the customer experience,seems closer than ever.Demand for superior customer experience is putting increasing pressure on network re

36、sources.On one hand,over the past five years or so,the newer generation of technologies,such as 5G and superfast fiber broadband,have improved the reach and quality of network services.On the other hand,the increasing number of mobile phones and devices connecting to the network and exchanging ever-

37、increasing volumes of data have deepened the complexity of the telco network.This has rendered some network-management activities beyond the reach of manual operations,necessitating software-based automated operations.Autonomous networks can help manage this complexity while also chipping away at ri

38、sing network-operations costs.This convergence of rising connectivity levels,AI,and high-performance computing among other factors,goes beyond simply building and deploying telco networks in a new way but has implications across dimensions such as network operations and maintenance,network managemen

39、t systems,planning,and observability of networks.From a telco business perspective,this evolution of network technology has a profound effect on an operators strategy and operating model,skill acquisition,and build vs.buy decisions.10Capgemini Research Institute 2024Networks with intelligence:Why an

40、d how the telecom sector should accelerate its autonomous networks journey Introduction The potential of autonomous networks isnt limited to achieving business goals.Reducing energy consumption,intelligent workload management,and better resource utilization are already delivering sustainability bene

41、fits,particularly in relation to climate.Are these benefits sufficient to justify the investment?If so,which use cases are most widely applicable and what value can they drive?Are telcos that are ahead of others on the journey achieving greater benefits?And,if yes,how can fast-followers and beginner

42、s accelerate their transformation?This research report attempts to answer these questions.Source:Capgemini Research Institute analysis.FIGURE 2.Evolution of autonomous networksDegree of network autonomyPeople-drivenPredictivePrescriptiveClosed-loop automationProactive automationReactive automationOp

43、en-loop automationHuman-machineinterfacePreventive(self-driving network)Software-driven(KPI and rules-driven)Business impact and productivity11Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyThe telecom secto

44、r is just starting its autonomous networks journey0112Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyThe majority of telcos aim to have Level 3 or above autonomy by 2028Today,the majority of telcos(84%)have

45、either Level 1 or Level 2 autonomy,as per TM Forums taxonomy of autonomous networks(see figure 3).In the next five years,the majority of telcos(61%)aim to attain at least Level 3 autonomy.Beyond this,only 16%of telcos expect to reach Level 4 autonomy in the next five years,and only 1%expect to reach

46、 Level 5 overall(including operations).Telcos aiming to reach Level 4 and 5 believe technology will mature in the next five years,as more autonomous network use cases are implemented across network domains.Paul Kells,director of network strategy and engineering at the British media and telecom compa

47、ny Virgin Media O2(VMO2),tells us:“We have a three-year and a five-year network-transformation plan to have an intelligent network thats intent-based,using AI/ML.The design aims to minimize build and run cost per bit to mitigate traffic growth and the rise in input costs,in particular labor and ener

48、gy.We aim to reach Level 4 in five years.”Percentages represent share of organizations by overall level of network autonomy(according to TM Forums levels of autonomous networks).Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 3.T

49、he majority of telcos(84%)have attained Level 1 or Level 2 autonomySTATE OF OVERALL NETWORK AUTONOMY,CURRENT VS.EXPECTEDExpected,2028Current,20239%0%42%42%8%31%6%44%1%0%16%1%Level 0Level 1L1/L2(2023)84%61%Global averageGlobal averageL3+(2028)Level 2Level 3Level 4Level 513Capgemini Research Institute

50、 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyGlobally,European telcos lead progress in autonomous networks journeys(see figure 4A and 4B).For instance,Orange has targeted Level 4 autonomy for some processes by 2025.1 Laurent Lebouche

51、r,chief technology officer at Orange and senior vice president of Orange Innovation Networks,says:“Our ambition to reach Level 4 network autonomy(for some workloads as of 2025)plays a crucial role in the Groups strategic plan Lead the future which was announced in February 2023.Through enhanced use

52、of data and AI at Level 4 autonomy,our network will be,more agile,more effective,more resilient,and higher performing.”Nearly two-thirds of telcos in Sweden,Spain,Italy,and France have attained Level 2 autonomy.In 2021,Telefnica launched a program called Autonomous Network Journey(ANJ),encompassing

53、all its operations in Europe and the Americas.As well as network autonomy,ANJs goal considers the required management of data,architecture and enablers,security,sustainability,plus an organizational structure and reskilling staff to support the network.2 PAUL KELLSDirector,Network strategy and Engin

54、eering at Virgin Media O2(VMO2)“We have a three-year and a five-year network-transformation plan to have an intelligent network thats intent-based,using AI/ML.The design aims to minimize build and run cost per bit to mitigate traffic growth and the rise in input costs,in particular labor and energy.

55、We aim to reach Level 4 in five years.”14Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journey33%of telcos in the US have moved to either Level 2 or Level 3 autonomyThe head of telco analytics at a large US-based

56、cloud platform provider says:“European communication service providers(CSPs)are leading in autonomous network levels,possibly driven by the need to reduce O&M costs as average revenue per user is lower in Europe than in North America.”One-third of telcos(33%)in the US have moved to either Level 2 or

57、 Level 3 autonomy.US telco Verizon aims to build autonomous networks with monitoring and analytics capabilities.Verizon aims to use AI/ML to optimize network resources;detect network anomalies before service disruption;and automatically resolve issues within a continuous automated framework.3Percent

58、ages represent the share of organizations at specified levels of network autonomy for each country.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=111 large CSPs.FIGURE 4A.Nearly two in three telcos in Sweden,Spain,France and Italy operate at network

59、autonomy Level 2 the highest level across all countriesCURRENT STATE OF OVERALL NETWORK MATURITY,BY COUNTRYLevel 0Level 1Level 2Level 3Level 4Level 51%100%0%GlobalSwedenSpainItalyFranceGermanyJapanNetherlandsUKIndiaCanadaUSBelgiumAustralia6%42%71%67%63%62%8%25%40%14%17%36%33%50%60%40%83%17%25%25%50%

60、11%22%67%43%7%40%20%25%50%58%42%23%8%22%25%11%13%29%42%9%15Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyPercentages represent the share of organizations at specified levels of network autonomy for each cou

61、ntry.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.Asia,excluding China=Australia,Japan,India;Europe=Belgium,France,Germany,Italy,Norway,Netherlands,Spain,Sweden,UK;North America=Canada,US.FIGURE 4B.Globally,European telcos lead progr

62、ess in autonomous networks journeysCURRENT STATE OF OVERALL NETWORK MATURITY,BY REGIONLevel 0Level 1Level 2Level 3Level 4Level 51%1%6%14%3%51%51%13%25%56%6%33%12%23%64%42%42%9%GlobalNorthAmericaEuropeAsia100%0%16Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sect

63、or should accelerate its autonomous networks journeyBy domain,the picture looks similar,with a majority of telcos having all network domains at Levels 02(see figure 5).A challenge CSPs face in deploying higher levels of autonomy in network domain is how to monetize their investments.Going by the fig

64、ures,telcos are taking a fragmented approach to upgrading network autonomy.A European telco told us that they are fairly advanced in the transport domain,followed by RAN,and core in that order.Percentages represent share of organizations at specified levels of network autonomy for each network domai

65、n.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 5.For nearly 90%of telcos,network autonomy by domain is at Level 02LEVEL OF NETWORK AUTONOMY BY MAJOR NETWORK DOMAINS,2023Level 0Level 1Level 2Level 3Level 4Level 515%49%25%4%5%25

66、%46%20%5%4%9%6%50%56%34%29%7%7%1%0%100%2%3%ServicesTransportRAN/OpenRANCore17Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyMost of the autonomous networks use cases remain in the initial stages of developme

67、ntWe analyzed 18 autonomous network use cases across network domains to understand current level of implementation.We found most to be at the POC or pilot stage(see figure 6).Percentages represent share of organizations at each stage of autonomous networks use case implementation.Source:Capgemini Re

68、search Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 6.Most autonomous networks use cases are at proof of concept/pilot stageSTAGE OF IMPLEMENTATION OF AUTONOMOUS NETWORK USE CASESNot considering or planning/exploringPOC/Business casePiloting on a small

69、 scalePartial/full-scale implementationSubscriber churn/behavior prediction and take corrective actionPredictive maintenanceAdaptive/dynamic network policies for changing conditionsIntelligent orchestration,slice optimization and SLA assurance in core and data centerSlice optimization and SLA assura

70、nce in RAN/ORAN17%26%17%12%17%12%12%20%12%19%9%32%40%35%33%46%33%39%42%30%18Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeySource:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023Ja

71、nuary 2024,N=113 large CSPs.Top autonomous network use cases across domainsNetwork DomainRadio AccessNetwork(RAN)Slice optimization and service-level agreement(SLA)assuranceTraffic steering and quality of experience(QoE)/quality of service(QoS)optimizationIntelligent orchestrationDynamic capacity pl

72、anningxHaul network predictive analyticsNetwork failure prediction in multi-layer software-definednetworking(SDN)Predictive maintenanceAdaptive/dynamic network policies for changing conditionsNetwork design optimization and resource allocationSubscriber churn/behavior prediction and take corrective

73、actionPrediction of level of user satisfaction Core and datacentersTransportOperations andmaintenanceCustomer-facinguse casesTop use cases under experimentatio(POC/pilots)per our surveyDescriptionOptimization of network slicing.Network slicing allows multiple virtual networks over a shared network d

74、omainDefining an optimal path for network traffic to maintain or improve customers QoE/QoSIntelligent orchestration is automating interactions across multiple types of device,domain,and related systemsAssessing network utilization,traffic volume,and traffic type to identify problems on a real-time b

75、asisAnalyzing network problems and providing solutions using AI/MLUsing AI/ML to predict failure of controllers in SDN networks under certain conditionsPredicting faults in network components by collecting data from multiple network sourcesChanging network policies to adapt to changing conditionsMea

76、suring metrics such as latency and throughput,and optimally allocating resourcesUnderstanding subscriber behavior and taking corrective action to retain customersAnalyzing data to understand customers satisfaction level with telecom services19Capgemini Research Institute 2024Networks with intelligen

77、ce:Why and how the telecom sector should accelerate its autonomous networks journeyUse cases Verizon demonstrated the ability to sustain performance levels for mission-critical functions while passing video data over a network slice in a completely commercial 5G environment through Axon Fleet 3 and

78、Axon Respond services.It is built on Verizons cloud-native,standalone 5G core,offering service flexibility,and automated scalability,among other benefits.Consumers of the service have enhanced customer experience as they can access live maps,live streaming of mobile videos,and real-time situational

79、awareness.4 The network director at a UK-based telco reveals:“Over 70-80%of the tickets,events,or incidents in the network are more or less similar or the same.These incidents occur frequently,demanding a considerable amount of human effort for activities that are not particularly complex to resolve

80、.Automating these can save significant time and resources.”Orange uses AI and advanced data-management platforms to detect abnormal network behavior or degradations in call quality or data flows.These trigger actions within intelligent automation and service-orchestration systems to ensure a consist

81、ent service level across distributed IT environments.5 Emmanuel Lugagne-Delpon,director of technical networks,outlined examples of how Orange has improved its network operations using AI tools:“AI can enable Orange to build smarter networks,improve operational efficiency,and transform customer inter

82、action and experience,all while following AI guidelines.”Simon Norton,head of digital networks and OSS(Europe and international networks)at Vodafone,says:“We have introduced autonomous networks use cases in anomaly detection and predictive maintenance,we aim to replace more than 30%of our reactive t

83、ickets with predictive tickets.This enables us to fix issues before they ever impact a customer and at a lower cost relative to a classic reactive intervention.”20Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks jour

84、neyEMMANUEL LUGAGNE-DELPONDirector of technical networks,Orange“AI can enable Orange to build smarter networks,improve operational efficiency,and transform customer interaction and experience,all while following AI guidelines.”21Capgemini Research Institute 2024Networks with intelligence:Why and how

85、 the telecom sector should accelerate its autonomous networks journeyWhat is holding back the industrys progress on autonomy?0222Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyTelcos point to cultural issues

86、,integration,and lack of technological maturity as top challenges to transformationAs figure 7 shows,half of telcos cite cultural challenges(e.g.,inappropriate employee mindsets and behavior)and technology integration issues as the top barriers to the transformation of autonomous networks.Business c

87、ulture develops and is reinforced longitudinally.Managing cultural transition can be more difficult than deploying the technology itself.As the director of product management at a US-based network equipment provider says,“Its culture/traditional operations vs.a mindset change.The way you operate net

88、works,or the kind of upskilling required now has tremendously changed compared to a few decades ago.For instance,do I choose the traditional ITIL functions,or the advanced NetDevOps?”“We need to continue the transformation of telecoms culture,to make it more inclusive,more customer-focused,and more

89、innovative,”says Kate Johnson,chief executive officer at Lumen Technologies,a global US-based telecom company,“There are enormous opportunities for the industry to solve technology challenges across sectors including retail,healthcare,manufacturing,and government agencies.We need to focus on underst

90、anding the human element and our customers first.”6 Moreover,employees are not usually enthusiastic about change.A recent Harvard study found that,“Willingness to support enterprise change collapsed to just 43%in 2022,compared with 74%in 2016.In 2022,the average employee experienced 10 planned enter

91、prise changes such as a restructure to achieve efficiencies,a culture transformation to unlock new ways of working,or the replacement of a legacy tech system up from two in 2016.”7MABEL POUS-FENOLLAR Global head of digital and zero-touch operations at Vodafone“We focus on the entire life cycle of ne

92、tworks as part of our operational strategy.Instead of solely concentrating on the operational side,there is value in considering the design,build,and operation phases and how they interconnect.Our intent is to integrate technologies from the beginning of network development to ensure a seamless and

93、autonomous network operation.”23Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyAutonomous network implementation is a complex,systematic transformation involving many ecosystem partners.Implementing solution

94、s across different layers is challenging.Percentages represent the share of organizations.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 7.Telcos encounter several obstacles to their transition towards higher network autonomyTOP

95、 BARRIERS TO ADOPTION OF HIGHER LEVELS OF NETWORK AUTONOMY51%48%28%25%24%21%21%38%32%33%Insufficient business cases or implementationexamples to draw onInsufficiently mature standards for implementationLack of confidence in hyperscalers/NEPs to provide higherlevels of autonomous networksLack of requ

96、ired skill sets among internal workforce AI regulations Data protection regulations e.g.,GDPRApplicable technologies not yet matureData sovereignty concernsIntegration issuesCultural challenge in terms of employee mindset24Capgemini Research Institute 2024Networks with intelligence:Why and how the t

97、elecom sector should accelerate its autonomous networks journey51%of telcos cite employee mindset as a top barrier to the adoption of autonomous networksHighlighting the importance of acknowledging technology integration issues,Juan Luis Mulas,head of telco cloud,OSS and automation at Orange Spain,s

98、ays:“To transform to Level 4 network autonomy,its not only us who have to change.Our ecosystem,including vendors,system integrators,hyperscalers,and software providers,will have to reskill or upskill people to drive this change forward.”Several regulatory challenges around data sovereignty,data prot

99、ection,and AI,also appear among top ten challenges perceived to be hindering telcos autonomous networks journeys.We will see in the later sections that leading telcos are bothered much less about these as compared to telcos that are at initial stages of their transformation.Moreover,large organizati

100、ons must also grapple with their legacy systems.Significant capital investment and employee training are required to switch over to new autonomous infrastructure.Hence,we also see CSPs initially focusing on automating their existing procedures,rather than overhauling entire operations.As we will see

101、 further,telcos are inadequately prepared to tackle these challenges and advance on their journey as there are several gaps in strategy and implementation.25Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyOnl

102、y about one in five telcos has a comprehensive strategy for autonomous networksAs figure 8 shows of the surveyed telcos,only 17%have a well-defined autonomous networks transformation strategy,while nearly seven in ten are working on implementing autonomous networks without a structured strategy.Mabe

103、l Pous-Fenollar,global head of digital and zero-touch operations at Vodafone,comments:“We focus on the entire life cycle of networks as part of our operational strategy.Instead of solely concentrating on the operational side,there is value in considering the design,build,and operation phases and how

104、 they interconnect.By leveraging technologies such as SDN(software-defined networking)and OpenRAN,we aim to automate and close the loop across various phases.Our intent is to integrate these technologies from the beginning of network development to ensure a seamless and autonomous network operation.

105、”JUAN LUIS MULASHead of global telco cloud,OSS and automation at Orange Spain“To transform to Level 4 network autonomy,its not only us who have to change.Our ecosystem,including vendors,system integrators,hyperscalers,and software providers,will have to reskill or upskill people to drive this change

106、 forward.”26Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyJuan Manuel Caro,director of autonomous networks at Telefnica,explains:“We look at our autonomous networks strategy as a set of five drivers:agility

107、 faster processes;efficiency conducting operations in a better way;sustainability reducing energy consumption and emissions;quality improving customer satisfaction;and perception and intelligence making our networks self-reliant and optimizable.”Source:Capgemini Research Institute,Autonomous Telco N

108、etworks Survey,December 2023January 2024,N=121 large CSPs.FIGURE 8.The large majority of organizations are working without a dedicated strategy for autonomous networksMATURITY OF AUTONOMOUS NETWORKS TRANSFORMATION STRATEGIESHave attained a thorough strategy for transforming to autonomous networks,co

109、mplete with clear goals and target timelinesEngaged in autonomous network efforts,but lack a dedicated,organization-level strategyNot yet transformed,but actively pursuing various transformation initiatives and working towards a comprehensive strategyCurrently not transformed,but in the process of d

110、eveloping a strategy for autonomous network transformation within the next 25 yearsHavent transitioned to autonomous networks but considering the possibility within the next 13 yearsHave not transformed to autonomous networks and currently have no plans moving40%36%1%5%2%17%27Capgemini Research Inst

111、itute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyFewer than one in five organizations has appointed a dedicated leader for autonomous networks,and 31%are in the process of finding oneBy appointing a dedicated autonomous networks lea

112、der,organizations can signal the priority they are assigning to autonomy initiatives.Yet fewer than one in five telcos has done so to date(see figure 9).About half have added advancing autonomous networks to the responsibilities of an existing leader About one-third(31%)have yet to identify a leader

113、 who will further their organizations autonomous networks agenda For its leading its ANJ program,Telefnica appointed Juan Manuel Caro Bernat as head of autonomous networks by giving him a dedicated responsibility of this strategic initiative.Source:Capgemini Research Institute,Autonomous Telco Netwo

114、rks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 9.Nearly one-third of organizations are yet to appoint an autonomous networks leader ORGANIZATIONS THAT HAVE APPOINTED CLEAR LEADERSHIP FOR LEADING AUTONOMOUS NETWORKS INITIATIVES15%52%31%2%No,we have not appointed/identified a leader yet

115、and have not started the process of identifying oneNo,we are in the process of identifying a leaderYes,an existing leader has been entrusted with additional responsibility for autonomous networksYes,a new leader with a dedicated responsibility for autonomous networks28Capgemini Research Institute 20

116、24Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeySource:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 10.More than one-third of organizations are yet to develop a roadmap for

117、 their autonomous networks journeysORGANIZATIONS THAT HAVE A CLEAR ROAPMAP FOR THEIR AUTONOMOUS NETWORKS JOURNEY9%50%39%2%No,we dont have a roadmap yet and are not developing oneNo,were in the process of developing a roadmapYes,we have a roadmap,however,only for the next 1-2 yearsYes,we have a roadm

118、ap for the entire journeyNearly half of the telcos have a roadmap for only the next 1-2 yearsIt is clear that a transformation of this size and scale will entail a multi-year journey.However,few telcos recognize the degree of planning required.Our survey found that only half have a roadmap for the n

119、ext couple of years and just about one in ten have one for the entire journey(see figure 10).The global head of autonomous networks at a large,multinational telecom operator based in Spain comments:“We have set targets for 2026 and 2030.Our focus is to manage networks comprehensively,efficiently,and

120、 zero-touch,all while warranting plausible bandwidth.The goal is to have intelligent and automated networks,with effective traffic management.”29Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyLaurent Lebouch

121、er,talking about Oranges road map,comments:“To improve network lifecycle management process,our plan is to focus on workloads where it matters the most and where the need to be agile is felt the most.We need to improve the time to market,the time it takes to bring new features to the network,and we

122、need to do it in a safe and agile way.”JUAN MANUEL CARODirector of autonomous networks at Telefnica“We look at our autonomous networks strategy as a set of five drivers:agility faster processes;efficiency conducting operations in a better way;sustainability reducing energy consumption and emissions;

123、quality improving customer satisfaction;and perception and intelligence making our networks self-reliant and optimizable.”30Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyThe potential of AI/ML and generativ

124、e AI for autonomous networksThree out of five telcos are exploring the potential of generative AI(Gen AI)for autonomous networksAI and ML,along with generative AI,will help CSPs to achieve Level 3 and above autonomy over the next 25 years.AI/ML-based methodologies and generative AI have the potentia

125、l to help CSPs in use cases such as:complex event processing;dynamic bandwidth selection and path selection;network capacity planning;and network provisioning and optimization.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=284 telco tech executives,1

126、13 large CSPs.FIGURE 11.One in ten telcos has implemented generative AI for networks at a partial scaleGENERATIVE AI MATURITY FOR AUTONOMOUS NETWORKS FOR YOUR ORGANIZATION2%13%21%57%0%7%We are not interested in integrating generative AIWe are yet to begin exploring generative AI capabilities We have

127、 started exploring the potential of generative AIWe have begun working on some pilots of generative AI initiativesWe have enabled generative AI capabilities in some of networks/regions/functionsWe have enabled generative AI capabilities in most/all of our networks/regions/functions31Capgemini Resear

128、ch Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journey“Gen AI cannot be disassociated from automation anymore.While automation was mostly heuristic in the beginning,operators now constantly deploy telemetry,network function data a

129、mong others and build ML algorithms,”says head of telco analytics at a large US-based cloud platform provider.Eight out of 10 telcos are either currently using or plan to use generative AI in complex event processing.Seven out of 10 respondents are either currently using or plan to use generative AI

130、 in dynamic bandwidth and path selection.Next in terms of use-case implementation are network provisioning and optimization,network capacity planning,and network slicing(half of respondents).Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=284 telco te

131、ch executives,113 large CSPs.FIGURE 12.More than half of telcos are currently using or planning to use generative AI in several use case in the next 12 yearsCURRENTLY USING OR PLANNING TO USE GENERATIVE AI IN NEXT 1-2 YEARS2%79%69%58%58%53%Network slicingNetwork capacity planningNetwork provisioning

132、 and optimizationDynamic bandwidth and path selectionComplex event processing32Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyComplex event processingComplex event processing aggregates data and analyzes cau

133、se-and-effect relationships between events on a real-time or near-real-time basis.It finds patterns in event relationships and draws insights from analysis of those patterns,upon which users can then act.It also makes it possible to spot anomalies in the data sooner,with lower latency.Developments i

134、n both closed and open-source large language models(LLM)have improved generative AIs ability to understand natural language.The use of LLMs has the potential to help field engineers to ask questions of data and correlate problems with their possible causes more efficiently than before.Dynamic bandwi

135、dth and path selectionGenerative AI and AI/ML algorithms can be used by CSPs networks to optimize resource usage and manage changes in network traffic in real time,by identifying bottlenecks and optimizing traffic routes.The algorithms analyze large amounts of traffic data to allocate bandwidth dyna

136、mically,enhancing network efficiency,and reducing energy costs and carbon emissions.It also ensures important and critical applications receive greater bandwidth during peak usage periods.The algorithm can also predict network congestion by analyzing historical and real-time data,allowing CSPs to ac

137、t to prevent network congestion.Generative AI can help generate synthetic data to help train AI models for a variety of scenarios and make them more robust to handle problems in future.33Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its

138、autonomous networks journeyMore granular and specific generative AI use cases to improve efficiency and effectiveness across network operations include:1.Translation:Engineers within network operations centers(NOCs)are tasked with scrutinizing network data and deriving insights to inform decision-ma

139、king.Generative AI can swiftly translate this data into plain language text,enhancing operator efficiency and effectiveness.2.Fraud resolution:Traditional methods of fraud detection and resolution often suffer from delays,typically requiring 24 weeks for operators to identify and issue legal notices

140、 to perpetrators.This timeframe affords the fraudster time to abscond,leaving the network vulnerable.Integration of generative AI can speed up drafting and dispatching fraud-related documentation,allowing more expeditious legal intervention.YVONNE KUIMBAHead of AI and data at TM Forum“By harnessing

141、synthetic data to train new network models and enable self-evolution of networks,telco systems can become adept at navigating unforeseen scenarios.”34Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journey3.Model tr

142、aining:Generative AI demonstrates efficacy in model training through the generation of synthetic data.This synthetic data,in turn,facilitates the training and refinement of novel network models,contributing to advancements in network evolution and performance.Yvonne Kuimba,head of AI and data at TM

143、Forum shares:“By harnessing synthetic data to train new network models and enable self-evolution of networks,telco systems can become adept at navigating unforeseen scenarios.”Paul Kells at VMO2 says:“We plan to use Gen AI in network configuration and fault analysis,correlating and analyzing custome

144、r complaints with actual metrics on the network in the locality and providing more informed insight into where reactive and proactive maintenance is necessary.”35Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journ

145、eyTelcos that are moving faster on autonomy are reaping clear benefits0336Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyAutonomous networks offer clear,significant benefitsTelcos with higher autonomy receiv

146、e more business and operational benefits,make greater technological advancements,offer better end-customer experiences,and enjoy greater sustainability(see figure 13).In the past two years,telcos have,on average,achieved a 20%improvement in operational efficiency and an 18%reduction in network OpEx

147、through autonomous networks initiatives.More than half have witnessed a marked improvement in fault detection,network scalability,and flexibility,along with a notable decrease in network downtime and outages.Yvonne Kuimba at TM Forum,says:“Better resource allocation,increased operational efficiencie

148、s,and enhanced scalability are compelling benefits of autonomous networks.Telcos are proactively rolling out new services,optimizing network resources,and prioritizing customer experience by minimizing downtime and enhancing satisfaction.Their focus is on achieving predictability and efficiency in n

149、etwork management,and autonomous networks play a pivotal role in realizing these objectives.”Julia Martinez Arenas,customer experience manager at Telefnica,underscores end-customer benefits:“Having a personalized and real-time customer satisfaction index is key to fine-tuning some new services in 5G

150、,such as network slicing and allows zero-touch automation to solve or mitigate any issues that may impact our customers experience.”8Highlighting the qualitative benefits,Simon Norton of Vodafone notes:“There are a lot of operational excellence benefits we get from automation.The mean time to resolv

151、e has decreased from many hours to just minutes.Weve observed a 95%improvement in the time required,from incident detection and work order generation to dispatching the field engineer.”Furthermore,71%of telcos have been able to reduce significantly the energy consumption of their networks in the pas

152、t two years through autonomous network initiatives.We explore this in more detail in the upcoming section on sustainability.37Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeySource:Capgemini Research Institut

153、e,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 13.Autonomous networks have yielded significant benefits Type of benefitPercent of CSPs that have realized significant improvements in these areas20%Improved scalability and flexibility in network management55%51%48

154、%47%58%58%71%Improved fault detection Reduced network downtime and outagesOptimized resource allocation and utilizationEnhanced network coverage and connectivityImproved reliabilityImproved energy savings18%Technological benefitsEnd-customerbenefitsSustainability benefits18%17%Faster time-to-market

155、for new servicesCost savings through reduced network CapExCost savings through reduced network OpExIncreased operational efficiencyEXTENT OF BUSINESS AND OPERATIONAL BENEFITS REALIZED THROUGH AUTONOMOUS NETWORKS INITIATIVES IN THE LAST 1-2 YEARS38Capgemini Research Institute 2024Networks with intell

156、igence:Why and how the telecom sector should accelerate its autonomous networks journeyThe benefits of autonomous networks are expected to outweigh investmentAccording to our survey results,telcos expect to invest$87 million per organization in autonomous networks over the next five years.Our analys

157、is suggests that network OpEx savings from autonomous networks alone amount to$150 million$300 million for an average telco in our survey over this period(see table below).For mid-sized telcos(typically with annual revenue of$11billion$29 billion),potential OpEx savings are estimated at$300 million,

158、and$500 million for large-sized telcos(typically over$30 billion in annual revenue).More details on these calculations are available in the appendix.SIMON NORTONHead of digital networks and OSS(Europe and international networks)at Vodafone“There are a lot of operational excellence benefits we get fr

159、om automation.The mean time to resolve has decreased from many hours to just minutes.Weve observed a 95%improvement in the time required,from incident detection and work order generation to dispatching the field engineer.”39Capgemini Research Institute 2024Networks with intelligence:Why and how the

160、telecom sector should accelerate its autonomous networks journey*Investments in autonomous networks include hardware,software,services,talent,among other elements.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.Capgemini Research Instit

161、ute analysis.OpEx growth between 202228 is assumed to be same as GDP growth estimates from IMF at 3%annually.Average investment dedicated to autonomous networks over the next five years*(survey data)$87 million$10.2 billion68.5%$7 billion17%$1.2 billion$6.7 billion9%4.5%$300 million(optimistic scena

162、rio)$150 million(conservative scenario)3.4x(optimistic)1.7x(conservative)1.5 years(optimistic)2.9 years(conservative)Average revenue per operator(survey data)Total telco OpEx as a share of total revenue9Total telco OpEx per operator in our survey(estimated)=B*C Network operations OpEx as a share of

163、total OpEx10 Network operations OpEx(estimated)=D*ETotal network OpEx over the next five years(estimated)=Sum of projected network OpEx in 202428 Expected network OpEx reduction by telcos enabled by autonomous networks in the next 25 years(survey data),optimistic scenarioExpected network OpEx reduct

164、ion by telcos enabled by autonomous networks in the next 25 years(survey data),conservative scenarioCorresponding network OpEx savings from autonomous networks use cases(estimated)=G*H and G*IReturn on Investment(estimated)=J/APayback period(estimated)=A/K/5ABCDEFGHIJKLTable 1.Analysis of OpEx savin

165、gs from autonomous network use cases40Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyLeading telcos are significantly outperforming their peers To gain a sense of how telcos progressing faster on their auton

166、omous networks journey are achieving better outcomes,we classified surveyed organizations into leaders,fast-followers,and beginners,based on the following dimensions:Maturity of autonomous networks transformation strategy and roadmap Investments dedicated to autonomous networks transformation Curren

167、t level of network maturity Stage of implementation of various autonomous networks use cases across domainsA telco that has made significant progress across all four dimensions is designated a leader.Additional information regarding this analysis can be found in the appendix.Based on this analysis,l

168、eaders acquire significant benefits currently unavailable to beginners(see figure 14).Over the past 12 years,telco leaders achieved a 19%cumulative reduction in network OpEx,outpacing beginners,who managed a 13%reduction.Leaders have also demonstrated a notable improvement in time-to-market for new

169、services,achieving a 22%cumulative enhancement twice that of beginners,while fast-followers achieved a 17%improvement.This outperformance could be attributed to the presence of a structured strategy,a clear roadmap for the entire journey,higher levels of network autonomy and experimentation,and scal

170、ed implementation of several use cases across network domains.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs,N=19 Leaders,N=74 Fast-followers,and N=20 Beginners.Capgemini Research Institute analysis.FIGURE 14.Leaders outperform beginne

171、rs in substantial benefitsEXTENT OF BENEFITS REALIZED IN THE LAST 12 YEARS,LEADERS VS.OTHERSFaster time-to-market for new servicesCost savings through reduced network OpExCost savings through reduced network CapExIncreased operational efficiencyBeginnerFast-followerLeader11%17%22%13%19%19%16%18%20%2

172、1%19%19%41Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyTelcos at Level 2/3 have realized significantly larger benefits than those at Level 0/1Telcos operating at Levels 23 have reported notable reductions

173、in operational costs,faster time-to-market for new services,and increased overall efficiency(see figure 15).Telcos with partial or conditional autonomy(Level 23),utilizing data,AI capabilities,and real-time insights to enable intent-based closed-loop management in certain network domains,have achiev

174、ed an 11%improvement on average in operational efficiency.Conversely,those with no or minimal autonomy(Level 01)have realized a modest 4%improvement on average in operational efficiency.Likewise,telcos with higher network autonomy are also saving more,with 16%reduction on average in network OpEx.Ora

175、nges Juan Luis Mulas sums it up:“Implementing fully autonomous use cases has reduced the mean time to repair from hours to minutes,delivering enormous benefits to our customers.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=112 large CSPs,N=58 at Lev

176、el 0/1 overall network autonomy,N=54 at Level 2/3 overall network autonomy.Capgemini Research Institute analysis.FIGURE 15.Telcos at higher levels of network autonomy have realized greater cost savings and faster time-to-market for new servicesEXTENT OF BENEFITS REALIZED IN THE LAST 12 YEARS,TELCOS

177、AT L0/L1 VS.L2/L34%11%11%16%11%6%11%9%Increased operational efficiencyCost savings through reduced network OpExCost savings through reduced network CapExFaster time-to-market for new servicesLevel 0/1Level 2/342Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom secto

178、r should accelerate its autonomous networks journey20%improvement on average in operational efficiencyInvestments in autonomous networks include hardware,software,services,talent,among other elements.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=108

179、 large CSPs.Capgemini Research Institute analysis.FIGURE 16.Telcos anticipate significant OpEx savings and ROI of 24x as they advance through each level EXPECTED INCREMENTAL SAVINGS AS TELCOS ADVANCE TO HIGHER LEVELS OF NETWORK AUTONOMY$202$203$310L1 to L2L2 to L3L3 to L40$100$200$300$4000.01.02.03.

180、04.05.04.53.52.51.50.5Expected incremental savings through reduced network OpEx over the next 5 years(in$millions)ROI2.02.33.9Incremental savings($millions)Return on InvestmentWe also analyzed the potential savings from reduced network OpEx for telcos over the next 5 years as they progress each leve

181、l of network autonomy.With each advancement,significant savings are anticipated at high ROI(see figure 16).Telcos planning to move from Level 1 to Level 2 within the next five years expect an average savings of about$202 million at an ROI of 2x,while those transitioning to Level 3 anticipate an addi

182、tional$203 million in savings.Further progression to Level 4 is estimated to bring an additional$300 million in total savings at an ROI of 4x,revealing that gains gather speed as telcos move to Level 4.43Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector shoul

183、d accelerate its autonomous networks journeyThe sustainability benefits of autonomous networksThe network accounts for 90%of energy use for an operator on average and the radio access network(RAN)represents more than 80%of this.11 Consequently,energy efficiency has become a core strategic priority a

184、cross the telecom ecosystem,offering both financial and environmental advantages.Nearly all(97%)of executives in our survey believe that embracing higher levels of autonomous networks will optimize energy usage,concurrently reducing carbon emissions within their organizations.Telefnica Group has red

185、uced its energy consumption by 7.2%between 2015 and 2022 even as the traffic managed by its networks increased more than seven-fold in the same period.12 The company has reduced its overall emissions by 51%in seven years.A recent trial conducted by China Mobile yielded a saving of 475 million kWh of

186、 electricity and a reduction of 373,000 tons in carbon emissions over a 12-month period.13 97%of executives believe that embracing higher levels of autonomous networks will optimize energy usage,concurrently reducing carbon emissions within their organizations44Capgemini Research Institute 2024Netwo

187、rks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeySource:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.Capgemini Research Institute analysis.Refer to appendix for more details on the analys

188、is.FIGURE 17.Telcos expect up to 5%reduction in their GHG emissions through autonomous networks within the next five yearsEXPECTED ROLE OF AUTONOMOUS NETWORKS IN REDUCING GHG EMISSIONS OF TELCOS,OPTIMISTIC SCENARIOEXPECTED ROLE OF AUTONOMOUS NETWORKS IN REDUCING GHG EMISSIONS OF TELCOS,CONSERVATIVE

189、SCENARIOCurrent overall GHG emissions per telcoTargeted overall GHG emissions per telco in the next 2-5 years-32%Expected overall emissions reductionExpected reductionin networkemissions byautonomous networksExpected reductionin networkemissions byautonomous networks5.65 Mt,CO2-eq3.84 Mt,CO2-eq5.2%C

190、urrent overall GHG emissions per telcoTargeted overall GHG emissions per telco in the next 2-5 years-32%Expected overall emissions reduction5.65 Mt,CO2-eq3.84 Mt,CO2-eq2.6%Our telco cloud study last year found that telco networks presently account for 42%of greenhouse gas(GHG)emissions of a telcos o

191、verall emissions.The telcos surveyed for this study aim to reduce their overall emissions by 32%overall within the next 5 years.Telcos project a 7.515%reduction in their networks carbon emissions as they transition to higher levels of autonomous networks in conservative and optimistic scenarios resp

192、ectively.Our analysis projects that autonomous networks have the potential to cut down telcos overall emissions by 2.65.2%in conservative and optimistic scenarios,respectively(see figure 17).2.65.2%expected reduction in telcos overall emissions enabled by autonomous networks45Capgemini Research Inst

193、itute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeySeveral autonomous networks use cases have the potential to improve network energy efficiency and reduce telcos overall carbon footprint,without compromising customer experience,such

194、as:AI-enabled energy savings in RAN,core,and data centers Predictive analytics and maintenance for network devices and infrastructure Helping enterprise customers cut down Scope 1 carbon emissionsSimon Norton of Vodafone adds:“With AI/ML and automation,we can dynamically optimize energy usage during

195、 low-traffic hours without compromising customer satisfaction.By analyzing performance data and considering geographical factors,we can precisely adjust energy-saving features on our sites,potentially doubling energy savings.”Mabel Pous-Fenollar of Vodafone further emphasizes:“Automation plays a piv

196、otal role in our sustainability efforts.Our focus extends beyond energy to support customers worldwide.Autonomous networks are crucial for advancing sustainability,addressing both industry and global challenges,particularly in the face of climate change.The frequency of incidents,such as floods and

197、fires,has significantly increased 46Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeydue to climate change.Our vision for autonomous networks,combined with weather data,positively impacts not only energy but a

198、lso on a broader scale concerning the planet.”Only 5%of organizations currently involve sustainability experts in teams committed to advancing their autonomous networks,while about 15%recognize a gap in expertise.Bridging this talent gap would also be essential for realizing the full potential of au

199、tonomous networks in terms of fast-tracking telcos environmental sustainability.SIMON NORTON Head of digital networks&OSS(Europe&international networks)at Vodafone“With AI/ML and automation,we can dynamically optimize energy usage during low-traffic hours without compromising customer satisfaction.B

200、y analyzing performance data and considering geographical factors,we can precisely adjust energy-saving features on our sites,potentially doubling energy savings.”47Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks jo

201、urneyHow can telcos accelerate and sustain their autonomous networks journeys?0448Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyDevising a strategy and roadmap for the autonomous networks journeyAs we saw p

202、reviously,only about one in five organizations(17%)has a comprehensive autonomous networks strategy in place.We also find that having a robust strategy and roadmap is a key differentiator,allowing those in possession to reap the highest rewards of this transformation.It can also serve as a key enabl

203、er to scale and sustain the transformation through its entire journey.17%of telcos have a comprehensive autonomous networks strategy in placeElaborating on their autonomous networks strategy and its origins,Juan Manuel Caro,director of autonomous networks at Telefnica said,“Though our strategy for a

204、utonomous networks started with operations a few years ago,we soon realized that the opportunity and challenges are much bigger than operations and span the entire technical organization.Thats when we decided to make it a company-wide slogan to give it a broader reach than pure automation in operati

205、ons.”Leading telcos prioritize a comprehensive strategy involving most or all of the following elements(see figure 18):Strategy and roadmap:A comprehensive strategy backed by business case,and internal as well as external collaboration.People:Reskilling and upskilling the workforce to adapt their wa

206、ys of working with a higher degree of automation,sometimes in closed-loop environments.Also deals with changing mindset and culture to be more open to delegating control of networks to software and cloud.Technology:Including virtualization and cloudification of networks and having the infrastructure

207、 for data collection,analysis,and AI.49Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journey Pace of transformation:Including identifying and selecting use cases that make the most sense from a network-maturity pe

208、rspective.Innovation:Embracing continuous innovation,including the latest technologies,such as generative AI and OpenRAN.We will explore each of these areas in greater details after we delve into how telco leaders develop a robust autonomous networks strategy.Source:Capgemini Research Institute anal

209、ysis.FIGURE 18.Key elements of a robust autonomous networks strategyDevising a strategy and roadmap for the autonomous networks journeyInvesting in key technology and data/AI initiativesInnovating continuously to be in tune with the latest developmentsStrategyPeople TransformationTech and data/AIPac

210、eInnovationFormulating the right talent strategy to support the scale-up of network autonomy Calibrating the pace of your transformation in line with networks maturity50Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous network

211、s journeyEstablishing the business case backing the investmentSecuring financing for an initiative such as autonomous networks is a key part of a robust strategy.Some organizations make a global/group-level financial commitment.However,to sustain such a transformation,it is critical that telcos iden

212、tify areas that can deliver value in the short as well as the longer term.It not only justifies the initial investment but pays for itself over the journey,while also generating associated business benefits.Juan Manuel Caro explains how Telefnica justifies investment in autonomous networks:“Fundamen

213、tally,were investing in upgrading our network with next-generation technologies and preparing it for new customers and services.Its a key enabler for our business and customer experience and does not require a short-term pay back.Nevertheless,we are getting efficiencies from automation and autonomy

214、that allow us to deliver our services faster,better,and cheaper.And,going forward,it will enable the monetization of new services and the massification of network APIs(application programming interfaces).”It is also a tendency in the sector to tie in investments like these with large technology refr

215、esh cycles that happen every five years or so.However,it need not be done this way.As we saw in section 3,the level of investment necessary for autonomous networks is not very high(1%of annual revenue over five years on average per organization).The high ROI from autonomous networks initiatives(1.42

216、.7x)is likely to be recouped within five years.Laurent Leboucher of Orange shares how CAPEX needed for the transformation are relatively small and should be prioritized:“The CAPEX required for autonomous networks is small compared with overall network CAPEX.This is why we encourage all our countries

217、 to make those investments.”JUAN LUIS MULAS Head of telco cloud,OSS&automation engineering at Orange Spain“A program like this demands investment from a strategic perspective.Once thats done,a few use cases can provide cost savings that can be ploughed back in as investments.For instance,in use case

218、s that were executing,there are significant CapEx savings.”51Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyOften,the business case stems from selected use cases.Some telcos find it better to execute proof-o

219、f-value experiments on a set of use cases,derive benefits,and use those findings to make a case for further investment in autonomous networks.Juan Luis Mulas at Orange Spain explains such an approach:“A program like this demands investment from a strategic perspective.Once thats done,a few use cases

220、 can provide cost savings that can be ploughed back in as investments.For instance,in use cases that were executing,there are significant CapEx savings.”Growing and collaborating as an ecosystemAll players in telecoms ecosystems recognize the need for sectorial transformation.Orange,for instance,eng

221、ages with TM Forum and open-source initiatives such as CAMARA from GSMA and the Linux Foundation.Apart from Orange,24 other major operators are also part of the GSMA global initiative.Through open network APIs,operators enable an ecosystem of developers to contribute to building network capabilities

222、,share learnings,and progress faster together.14The need for greater reliability and faster time-to-market for services are the top technological and business drivers,respectively,for transitioning to autonomous networks.To deliver on these objectives while accommodating the rising complexity of the

223、 network and the burden of legacy equipment,telcos will need the support of ecosystem players.As figure 19 shows,the majority of telcos(60%)prefer to lead the transformation of the network core themselves,but are much more open to their partners leading the transformation of other domains hyperscale

224、rs in transport(53%)and hyperscalers(43%),and software vendors(49%)in services.52Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyOur survey found that 38%of telcos consider data sovereignty concerns and 32%co

225、nsider data protection regulations such as GDPR to be a barrier for their organization in moving to higher levels of autonomous networks.However,leaders seem to have already found ways to make progress in compliance with regulations as much fewer of them perceive these challenges(16%and 26%)vis-vis

226、beginners(40%and 55%,respectively).As such,it becomes all the more important for telcos and their partners in the industry to closely work with government,policy institutions,and standards development organizations to ensure that regulation serves as a catalyst for the growth of the industry,not as

227、a hurdle.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 19.Telcos preferences for leading partners for autonomous networks transformationOurselves(telcos)Network equipment providersHyperscalersSystem integratorsSoftware vendorsC

228、oreRANTransportServices60%35%18%15%10%27%42%40%34%14%12%28%53%24%35%23%4%43%29%49%OF TELCOS THAT PREFER A PLAYER AS A LEAD PARTNER FOR AUTONOMOUS NETWORKS TRANSFORMATION53%of telcos are open to hyperscalers leading the transformation of transport network53Capgemini Research Institute 2024Networks wi

229、th intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyAaron Chase,senior director,network architecture strategy at Infinera,speaks about helping telcos overcome the interoperability challenge:“Telcos and CSPs are evolving their legacy equipment and closed sy

230、stems towards an open,disaggregated system.Additionally,separate domain controllers for each layer of networks that have never talked to each other before now need to.Network providers need to bridge the fiber optic network to other layers while enabling simple plug and play capability.This allows C

231、SPs to get alarm correlation,even from equipment that comes from different vendors.It helps them maintain the status quo for the next-generation tech cycle and save precious CapEx.”Building a global/group-level strategy with strong governance at global and local levelsFor global telcos,a key challen

232、ge to overcome is connecting global/group-level strategy with local strategies and initiatives at operating companies(OpCos)or country-level operators.Juan Manuel Caro told us about the strong governance model Telefnica has set up:“For each operating business,we have a program structure with a local

233、 leader and a group-level sponsor.The program comprises nine projects across all major areas and countries covered in the autonomous networks program.Every year,in Q1,we sync KPIs at both levels and commit our objectives to the Executive Committee together.Thereafter,we report our progress every thr

234、ee months for each of the projects.”He adds,“This robust governance has taken some time to establish,but it has been worth it,as it has created a common language and a shared sense of purpose across various internal teams.”Formulating the right talent strategy to support the scale-up of network auto

235、nomy Bridging the skills gap in areas such as AI and multi-disciplinary skillsA recent European Commission study showed that information and communication technologies(ICT)is one of the top three sectors in terms of new job vacancies,seeing a sharp rise of 4.1%in 2022.15 The transformation to autono

236、mous networks requires a step change in terms of how telco workforces manipulate and use data.This AARON CHASE Senior director,network architecture strategy at Infinera“Telcos and CSPs are evolving their legacy equipment and closed systems towards an open,disaggregated system.Network providers need

237、to bridge the fiber optic network to other layers while enabling simple plug and play capability.This allows CSPs to get alarm correlation,even from equipment that comes from different vendors.It helps them maintain the status quo for the next-generation tech cycle and save precious CapEx.”54Capgemi

238、ni Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeychange will require AI,ML,and data analytics skills at an unprecedented level.Nearly 3 in 5 telcos see the greatest skills shortage lying in AI and ML(58%).Telcos will

239、 require new strategies to acquire and retain data and AI experts.The second-largest skills gap is in multi-disciplinary or cross-domain/transversal skills.This area requires individuals to possess a combination of skills to a high level,for instance,a network engineer with data analytics skills or

240、a data expert with network domain knowledge.This is a tough ask,as these skills have historically been developed in specialized silos.During the implementation of autonomous networks,planners should consider both future and current demand.During the initial phases of the journey,more data/AI and aut

241、omation-related skills are required,whereas,during the advanced,more extensively automated stages of the journey,monitoring,problem-solving,and decision-making skills will be more heavily in demand.Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=335 t

242、elco executives.FIGURE 20.AI/ML and multi-disciplinary skills show the biggest gaps for autonomous networks transformationBIGGEST GAP IN TERMS OF SKILLSETS CURRENTLY REQUIRED FOR MOVING TO HIGHER LEVELS OF AUTONOMOUS NETWORKSSustainability skillsCybersecurity skillsNetworkingAI safety and ethics(for

243、 addressing issues linked to AI explainability and transparency)Automation Multi-disciplinary skills(e.g.,network engineer with data analytics skills)AI/ML58%44%42%40%16%34%15%55Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomou

244、s networks journeyinnovation professionals entrusted with implementing autonomous functionalities.Mabel Pous-Fenollar of Vodafone shares her strategy for overcoming this challenge:“It is critical that we bring all stakeholders along on this journey.We need supporters,but we also need everyone who is

245、 impacted by the transformation,whether they support it or not.A way to gain detractors support is to demonstrate the benefits to them on a smaller scale,winning their trust,and allowing them to test and ask questions.Theres no better way to organizational change than to win over detractors.”Investi

246、ng in key technology and data/AI initiatives Technology integration issues(48%),lack of maturity of AI and ML for networks(32%),and lack of standards(21%)are the top three technological barriers to moving to higher levels of autonomy.To ensure that there are no roadblocks from a technological perspe

247、ctive,telcos will need to make strategic investments in key technology and data/AI initiatives,such as:Upskilling and reskilling of the workforceTelcos are faced with the dual challenge of retaining the knowledge and experience of resigning or retiring staff,while also attracting and retaining new w

248、orkers.Moving internal systems and processes to higher levels of network autonomy allows manual tasks(and those beyond manual capabilities)to be automated.Survey data shows that there is potential to automate an estimated one-third(30%)of work traditionally done by in-house network engineers/archite

249、cts and administrators.Our survey also finds that over one-quarter of these workers(27%)must be reskilled to allow them to work with newer technologies and processes and undertake more value-adding jobs.AT&T invests more than$200 million annually in its internal training organization,branded as AT&T

250、 University.In addition,the company also teams up with universities and online-learning platforms.16 Preparing your organization for a mindset and culture changeTransformation calls for steadily relinquishing a level of control over the network to software and decision-making systems,such as AI and

251、ML models.As a result,it involves reorganization and reimagining of systems,processes,and tools,from manual or semi-autonomous ways of working to software-led or fully autonomous processes.To be fully effective,an upheaval of deep-rooted practices and processes requires a mindset change among the wo

252、rkforce.As we saw in section 2,our survey found that the cultural challenge(i.e.,changing the employee mindset)stands as the single biggest barrier(51%)to telcos moving to higher levels of autonomy.Further analysis shows that fewer organizations who have advanced on their journeys and developed a ro

253、admap consider this a challenge(43%)compared with those who are yet to build a roadmap for their transformation(63%).Juan Luis Mulas of Orange Spain concurs:“Changing peoples mindset has been the most challenging aspect of this journey so far.Having a strong leadership and vision has helped rally pe

254、ople around this program.Identifying appropriate trainings and customizing learning journeys depending on individual needs have helped as well.”A cultural divide also exists between operations teams,whose key priority is to optimize network operations and customer experience,and the data and AI team

255、,and 56Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyTaking stock of current and future data landscapesOften telcos lack an end-to-end view of their data landscape.Large organizations typically prioritize d

256、ata-related investments in customer-facing domains,leaving operations and back-end technology on the back burner.This obstructs the required extraction and distribution of data.For instance,lack of availability of a real-time data-collection,analysis,and feedback loop severely limits use cases such

257、as outage and threat detection,and dynamic policy management.The head of AI at a large Australian telco says:“From a data landscape perspective,mobile networks are fairly software-driven and have a simpler path to value.Fixed networks,in contrast,carry a lot of legacy that makes it harder to pull da

258、ta,clean it,and use it in models.”He adds:“Due to the complexity of KPIs,devices,exchanges,and so on,one of the biggest learnings in recent years has been that,instead of deploying a use case in one of the domains(such as in mobile or fixed domains),it is better to break down the data landscape into

259、 logical segments,solve for a use case such as anomaly detection in individual segments,and then scale those up.”AIS,one of Thailands top mobile operators,uses AI-enabled network digital maps to increase the efficiency of its IP-RAN network planning.17 It has automated collection and visualization o

260、f near-real-time data,including inventory,fault alarms,and linking KPIs,among others.It also highlights anomalies such as link outages and high latency,making network planning,as well as assurance,very effective.As a result,the fault-identification timeframe has reduced to less than one hour and the

261、 time-to-market for new services has reduced by 60%.Mabel Pous-Fenollar of Vodafone shares an example of how better data visibility can lead to better outcomes:“The crucial nature of visibility and availability of good-quality anonymized,aggregated data cannot be overstated.In countries where we hav

262、e a mature data landscape,for instance,for inventory data,it has allowed us to go beyond simple service automation to predicting and preventing equipment and service failures.”57Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomou

263、s networks journeyAdvancing network automation as part of telco cloud transformationTelco cloud transformation is one of the key levers of autonomous networks.The virtualization and containerization of network functions enables faster rollout of network functions,automation,and services to end-custo

264、mers.Our 2022 research on telco cloud transformation found that nearly half of telco network capacity will be cloud-native in the next 35 years.Yet,network automation is a low investment priority for telcos.As figure 21 shows,among top investment areas in the telco cloud value chain,network automati

265、on stands lowest,with only 27%of early adopters of telco cloud prioritizing it.27%of early adopters of telco cloud prioritize investment in network automationSource:Capgemini Research Institute,Telco Cloud Executive Survey,November-December 2022,N=114 telco executives.Percentages represent the share

266、 of executives selecting top three areas of investment across their telco cloud value chain.FIGURE 21.Network automation stands at the lower end of telco cloud investment prioritiesR&DCloud management software and toolsProduct/service innovation(e.g.,internet of things,edge computing)Cloud infrastru

267、cture(public/private/hybrid)Network automationEarly adopters of telco cloudFollowers in telco cloud adoptionLaggards in telco cloud adoption100%44%9%65%43%45%38%10%9%27%22%18%36%0%27%WHERE ARE YOU INVESTING THE MOST ACROSS THE TELCO CLOUD VALUE CHAIN(BY ORGANIZATION CATEGORY)?58Capgemini Research In

268、stitute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyEstablishing robust data governance and data frameworks for effective data management and processingA key pillar of the technology strategy for autonomous networks is the governance

269、 and management of data.Data governance sets standards and policies for how data is collected,stored,and analyzed in an organization.It is important that these policies observe the latest data security,privacy,and ethical guidelines before any data source is employed in an autonomous network use cas

270、e.The ever-increasing complexity of telecom networks makes it a challenging endeavor to collect data generated from a variety of network equipment,often through different protocols and standards,and then using it effectively to derive insights.Due to the proprietary nature of network equipment,the a

271、pplications and services that are bundled with them often produce data that is not interoperable and translatable across devices and domains.This leads to a highly fragmented and siloed data landscape that not only hinders use case implementation but also makes some autonomous networks use cases pra

272、ctically infeasible.The network transformation lead at a large,American multinational telecom company says:“Getting to Level 4 automation requires a solid foundation in real-time data analytics.Streaming analytics is a complex terrain,and its not just about collecting data but also swiftly analyzing

273、 it,and taking remedial actions with strong decision support systems powered by AI.The challenge lies not just in technology but in ensuring interoperability across diverse devices and ecosystems.”To overcome these challenges,telcos must put in place data frameworks that help with aggregating,normal

274、izing,and correlating heterogeneous network data.A data framework like this makes network data more open,accessible,documented,and trusted,serving as the foundation of further processing.It opens doors to making this data available for ready consumption by services or machine learning and AI-based m

275、odels that can be built on top of this framework.Further,MLOps/AIOps tools can be introduced to efficiently develop these models and manage their lifecycle.Beatriz Ortega,telco business development manager at Red Hat,talks about the importance of an end-to-end view of data:“Telcos have a goldmine of

276、 data sitting across various systems.However,when it comes to having the data ready for one use case,its almost always a challenge.Often,BEATRIZ ORTEGA Telco business development manager at Red Hat“Telcos have a goldmine of data sitting across various systems.Often,data is not interoperable,not of a

277、cceptable quality,or not available in real time for enabling real-time and near-real-time use cases.So,this is a key challenge that we try to solve by bringing in a network of partners that can bring together the right data,in the right place,at the right time.”59Capgemini Research Institute 2024Net

278、works with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyCalibrate the pace of your transformation depending on the maturity of autonomous networks To begin,telcos must select which network domain and use cases to prioritize for transformationTelcos at

279、the beginning of their autonomous networks journeys should prioritize one or two network domains among core,transport,access,or operations.Once the choice of domain is clear,several factors need to be considered:The level of risk that the organization is willing to take during experimentation and sc

280、aling(if selected use cases pose a risk to end-users/consumers,a network change going wrong can shut down connectivity,negatively impacting customer experience)The difficulty level of implementation The required level of investment The level of software-based network functionsdata is not interoperab

281、le,not of acceptable quality,or not available in real time for enabling real-time and near-real-time use cases.”She adds:“So,this is a key challenge that we try to solve by bringing in a network of partners that can bring together the right data,in the right place,at the right time.”Telefnica Brazil

282、 streamlined the integration,configuration,and provisioning of its network transport and mobile access layers,saving close to 6,000 workdays,reducing network alarms by 25%,and reducing rework by 30%.18 This was made possible by centralization of data from all key areas and customizing access to it.6

283、0Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeySource:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.FIGURE 22.Operations and core are top-priority

284、domains for autonomous networks36%31%20%7%3%TOP PRIORITY FOR TRANSITIONING TO AUTONOMOUS NETWORKS,BY NETWORK DOMAINCoreOperations(including OSS and BSS)RANTransportEdge/services Availability of good-quality network data Availability of domain experts Figure 22 shows network domains that are the top

285、priority for transitioning to autonomous networks.Juan Manuel Caro told us Telefnicas priorities for its transformation:“Were fairly advanced in terms of autonomous use cases for transport,followed by RAN.And Assurance is by far the process where we have advanced more until now.For Core,we have just

286、 implemented the first AI closed-loop use case,although we are being cautious with the level of autonomy there.It is important to gain trust before going farther due to the high impact of any potential problem of this domain.”The network director at a UK-based telco echoes this sentiment:“RAN takes

287、precedence due to frequent incidents.With numerous data centers and thousands of base stations,theres a significant opportunity for workload reduction.In critical areas like data and voice in the core,operators maintain a conservative approach due to higher risks.Mistakes in the core can impact a su

288、bstantial portion of the network,a risk many avoid.This cautious approach is common among many European operators.The focus on RAN is driven by lower risk and greater automation opportunities,resulting in significant savings.”61Capgemini Research Institute 2024Networks with intelligence:Why and how

289、the telecom sector should accelerate its autonomous networks journeyThe selection of use cases for beginner telcos can be low-hanging fruit:low-risk,high-benefit use cases to explore potential and prove value on a small scale.An agile test-and-learn strategy is crucial here,so that engineers can tri

290、al multiple approaches.Once there is sufficient confidence in technology and partners,the most successful can be scaled up to larger levels,across layers,domains,and cities/countries.Telcos midway through their autonomous networks journeys must double down on their investment and focus on scaling Te

291、lcos midway through their transformation journey have already proven value in some of the use cases and possibly partially scaled some of them.Given their relative success to date,it makes sense to redouble their investments and efforts on top use cases.They can do this by breaking silos across doma

292、ins and optimizing resource allocation among them,scaling tech hardware and software,and making these available across regions and domains.Tools should be made available across the organization,and best practices identified and implemented across functions,among other such scaling initiatives.HAN CH

293、EWAPAC head of network planning and optimization at Nokia“Our team invests in developing digital twins of physical networks.Creating a virtual replica of the network enables us to optimize network design,enhance diagnosis,and conduct simulation and what-if analysis.It also allows us to simulate and

294、train AI/ML algorithms and explore optimization scenarios to validate network implementation without the risk of impacting end user experience.The network digital twins will play a crucial role in telcos transition to more complex networks,with more sites/sectors,frequency and technology enabling th

295、e network to increase autonomy.”62Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyLaurent Leboucher told us about an initiative thats paying dividends for Orange:“We have a Telco Cloud Factory that is a platf

296、orm on which to build and deliver telco cloud instances for different countries and then operate them.We have delivered 40 instances to date across the group and plan to scale this to 60.”He adds:“On top of this layer,we have created a Network Integration Factory and a tooling zone that oversees onb

297、oarding and testing of each network function before they are put into production.Images and test plans are shared with countries,which can customize and deploy functions in their specific environments.This reduces a significant amount of process and tech overhead,leading to efficiencies.”Virgin Medi

298、a O2 is working to advance its autonomous networks journey by progressively defining and implementing“one use case at a time,”using data,analytics,and AI.19 Reflecting on the development,Paul Kells at VMO2,remarks:“Most of the use cases are presently at Level 2,and we are scaling up.In the past six

299、months,weve executed around 20 material use cases,each yielding material benefits.The key is strong governance coupled with robust change management.This ensures a smooth transition and guards against reverting to old operational methods after the initial disruptive phase.”Embrace continuous innovat

300、ion in network architecture and operationsAdvance initiatives to adopt open,disaggregated,and software-driven networksPursuing open and disaggregated networks throws the door open to innovative use cases some of them on autonomous networks have the potential to be big value drivers for telcos.It als

301、o paves the way for innovative services that are more in-tune with the networks behavior called network-aware apps/services.OpenRAN technology that introduces a set of open standards for radio access network(RAN)hardware and software is one of telco industrys key initiatives to decouple RAN hardware

302、 and software and reduce dependence on proprietary equipment.The disaggregation of network hardware and software that OpenRAN enables,removes the LAURENT LEBOUCHER Group chief technology officer at Orange and senior vice president of Orange Innovation Networks“The CAPEX required for autonomous netwo

303、rks is small compared with overall network CAPEX.This is why we encourage all our countries to make those investments.”63Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyneed to procure them from a small set o

304、f suppliers,boosting vendor diversity,and supply chain resilience.OpenRAN has the potential to be a key catalyst for innovation as it opens doors to new and more effective solutions,and advanced network functionalities that accelerate autonomous workflows.Recently,Vodafone started the rollout of Ope

305、nRAN at 2,500 mobile sites across Wales and the southwest of England.20 Trials have demonstrated that this transformation allows Vodafone to replace legacy technology with network equipment that exceeds 4G and 5G call success rate,as well as download and upload speeds compared to that of the legacy

306、tech.CAMARA is a telco global API alliance that seeks to:simplify the complexity of telco networks for making user-friendly APIs;adhere to data-privacy and regulatory requirements;and facilitate the integration of applications to network.2164Capgemini Research Institute 2024Networks with intelligenc

307、e:Why and how the telecom sector should accelerate its autonomous networks journeyTelco APIs have the potential to make telco networks much more programmable by enabling the extraction of data/information from the network.They also allow configuring the network by means of programmed actions taken b

308、ased on AI/ML model outputs.Exposing network APIs also makes it easier for developing network-aware apps and services a top 2 technological driver for autonomous networks for a majority of telcos(65%).Experiment with emerging technologies such as generative AI,metaverse,and digital twins to enhance

309、network operators efficiencyOur recent cross-industry research on generative AI revealed that generative AI is on the boardroom agenda for 96%of organizations.22 While generative AI is in its infancy in terms of scaled adoption and implementation,nearly 60%of executives globally say their leadership

310、 is a strong advocate for generative AI and only 39%are taking a“wait-and-watch”approach to adoption.As discussed earlier(see“The potential of generative AI for autonomous networks”),generative AI has a significant potential for various network use cases,well beyond its potential for customer engage

311、ment uses.As such,more exploration and experimentation is needed to harness the full potential of generative AI for telecom networks.Some telcos are using generative AI to help field technicians with a summary of past incidents and their resolutions,helping reduce the mean time to resolve(MTTR)for t

312、heir ongoing investigation significantly.23With immersive interfaces and metaverse,network engineers working on complex networks or machinery can access support from experts situated elsewhere.Similarly,technicians can scan components of machines they are working on through their AR headsets or AR-e

313、nabled smartphones and read detailed information about these components.The network transformation lead at a large,American multinational telecom company remarks:“We recognize the potential of innovation in driving new value.Therefore,we focus on technology convergence and particularly invest in eme

314、rging technologies such as gen AI,5G,edge computing,immersive technologies,etc.to drive autonomous networks.”Digital twins act as digital replicas of physical networks allowing simulation of network behavior with real-time or historical data.The insights can be used to optimize network performance a

315、nd predict future outcomes.Han Chew,APAC head of network planning and optimization at Nokia,asserts:“Our team invests in developing digital twins of physical networks.Creating a virtual replica of the network enables us to optimize network design,enhance diagnosis,and conduct simulation and what-if

316、analysis.It also allows us to simulate and train AI/ML algorithms and explore optimization scenarios to validate network implementation without the risk of impacting end user experience.The network digital twins will play a crucial role in telcos transition to more complex networks,with more sites/s

317、ectors,frequency and technology enabling the network to increase autonomy.”By leveraging these emerging technologies,telcos can further boost their operations efficiency and effectiveness.65Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate i

318、ts autonomous networks journeyThe operating model of telecom networks is steadily changing from one managed by human operators to an autonomous one where software aided by data and AI increasingly makes the network smarter,faster,and more efficient.Its a fundamental shift in terms of how networks ar

319、e designed,developed,and operated,since the shift from analog to digital networks a few decades earlier.As previous waves of similar transformations have shown,the transformation is imminent,and it is the pace of change that will define the difference between winners and also-rans.Conclusion66Capgem

320、ini Research Institute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyAppendixWe assessed the surveyed organizations using the aforementioned criteria.Organizations falling within the range of Mean+1 to Mean+3 were classified as Leaders

321、;those within Mean 1 to Mean+1 were categorized as Fast-followers;”while those below Mean 1 were categorized as Beginners.I.Classification of leaders,fast-followers,and beginnersModel inputs Strategy and roadmapEstablishment of a comprehensive autonomous networks transformation strategy,with well-de

322、fined goals and target timelinesEstablishment of a clear roadmap for the entire journey Telcos investing more than the global average and those investing less than the global averageTelcos state of network maturity based on TM Forums autonomous network levels.Current stage of implementation of vario

323、us autonomous networks usecases across domainsInvestments(including hardware,software,services,talent,among others)Current level of network autonomyImplementation of use casesFactors on which the analysis for leaders vs.fast-followers vs.beginners was based67Capgemini Research Institute 2024Networks

324、 with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyII.Autonomous networks business case calculations for mid-size and large-size telcosa.b.c.d.e.f.g.h.Benefit/KPIMid-size telco(typically$10-$29bn in annual revenue)Large telco(typically over$30bn in ann

325、ual revenue)Average investment dedicated to autonomous networks over the next five years*(survey data)$118 million$11.9 billion$2.02 billion$11.4 billion5%$293 million2.5x2 years$149 million$24 billion$4.08 billion$23 billion4%$502 million3.4x1.5 yearsTotal telco OpEx(estimated)=Average annual reven

326、ue*Total telco OpEx as a share of total revenue24 (68.5%)Network operations OpEx(estimated)(17%of total OpEx)Total network OpEx over the next five years(estimated)=Sum of projected network OpEx in 202428Expected network OpEx reduction by telcos through autonomous networks in the next 25 years(survey

327、 data),conservative scenarioCorresponding network OpEx savings from autonomous networks use cases(estimated)=d*eReturn on Investment(estimated)=f/aPayback period(estimated)=a/f/5*Investments in autonomous networks include hardware,software,services,talent,among other elements.68Capgemini Research In

328、stitute 2024Networks with intelligence:Why and how the telecom sector should accelerate its autonomous networks journeyIII.Expected role of autonomous networks in reducing GHG emissions of telcosRemarks/referencesI.II.III.IV.V.VI.VII.0.9684 Telco companies are responsible for 1.62%of global carbon e

329、missions25 Global GHG emissions in 2022 is 53.8 Gt CO2-eq26(=1.8%*53.8)66%5.6542%1.970.29550.1478(III=I*II/Number of operators in the study(113)(Survey data)(V=III*IV)(VII=(VI/III)15%(Survey data),optimistic scenario(VI=V*15%)7.5%(Survey data),conservative scenario(VI=V*7.5%)Telecom sector emissions

330、(2022),Gt CO2-eqShare of global market represented by telcos in our studyAverage emissions per telco,Mt CO2-eqCurrent share of GHG emissions of telco networks in organizations overall GHG emissionsCurrent emissions from telco networks per telco on average,Mt CO2-eqProjected reduction in network emis

331、sions enabled by higher levels of autonomous networks,Mt CO2-eqContribution of autonomous networks to target achievement5.2%(optimistic scenario)2.6%(conservative scenario)Source:Capgemini Research Institute,Autonomous Telco Networks Survey,December 2023January 2024,N=113 large CSPs.The optimistic s

332、cenario assumes that a half of the total reported gains(30%)are ultimately realized at the organizational level,whereas the conservative scenario assumes a fourth of the gains are ultimately realized.69Capgemini Research Institute 2024Networks with intelligence:Why and how the telecom sector should

333、accelerate its autonomous networks journeyResearch methodologyWe surveyed 435 industry executives,from large CSPs,network equipment providers(NEPs),and hyperscalers.Of these,335 executives belong to CSPs,50 to NEPs,and 50 to hyperscalers.They are based in 14 countries across North America,Europe,and Asia-Pacific.The distribution of respondents and their organizations is provided below.26%10%10%10%