《电信管理论坛：2024网络生命周期自动化和智能化发展报告-聚焦亚太地区（英文版）（31页）.pdf》由会员分享，可在线阅读，更多相关《电信管理论坛：2024网络生命周期自动化和智能化发展报告-聚焦亚太地区（英文版）（31页）.pdf（32页珍藏版）》请在三个皮匠报告上搜索。

1、April 2024|www.tmforum.orgAuthor:Charlotte Patrick,Contributing AnalystEditor:Dawn Bushaus,Contributing Editorfocus onsponsored by:driving intelligencein network lifecycleautomation:Asia-Pacificcontents03 the big picture06 section 1:the need for intelligence in network lifecycle automation11 section

2、 2:how much AI and machine learning are CSPs implementing today?14 section 3:examples of intelligence in network lifecycle automation22 section 4:make it happen strategies to take advantage of new intelligence 26 additional resourcesFind out more from our sponsor about AI-driven intelligence in netw

3、ork lifecycle management in Asia-Pacific:ServiceNow helps Telcos accelerate growth,maximize network investment,and reduce costs across the entire digital ecosystem all while delivering seamless experiences.Be it planning for 5G or fiber buildouts,managing data centers,launching and delivering new se

4、rvices,or automating network operations,our purpose-built solutions help make work easier across the entire service lifecycle with automation and AI.Built for the fast-changing world of Telecom,the Now Platform connects people and data for greater productivity and innovation.bigpicturethe3inform.tmf

5、orum.orgLarge CSPs in countries such as Australia,India,Indonesia,Japan,Malaysia,New Zealand,the Philippines,Singapore,South Korea and Thailand(note that this report does not include China)are seeing some early value from domain-specific deployments that allow them to build the skill sets needed to

6、work with intelligence,without the complexities of cross-domain data gathering or processes.Successful projects include anomaly detection for service assurance and task-specific models in the radio access network(RAN).Smaller CSPs,or those in developing markets,do not have as many intelligence proje

7、cts underway,and it is unclear whether they will invest heavily unless the use case solves a significant pain point or provides well-documented return on investment(ROI).TM Forum surveys conducted this year and last have found much enthusiasm and confidence in AI and machine learning among CSPs in t

8、he Asia-Pacific region although the number of respondents from the region is small compared to the full group of global respondents.Our AI Benchmark survey,for example,found that a full 85%of APAC CSP respondents believe their organizations Since the launch of ChatGPT in late November 2022,intellige

9、nce(defined in this research as the deployment of both AI and machine learning)has moved into the mainstream consciousness of communications service providers(CSPs),giving staff the ability to interact with a large language model(LLM)and discover its strengths and weaknesses for themselves.In the As

10、ia-Pacific(APAC)region,large operators are implementing a range of active intelligence projects,while smaller CSPs are moving more slowly.4inform.tmforum.orgIt is unclear whether smaller CSPs will invest in intelligence unless the ROI is clear.are well positioned to exploit the potential of AI and m

11、achine learning.Mark Sanders,Telstras Chief Architect,notes that many of his engineers are curious about the possibilities of AI,and he adds that they are running projects outside regular working hours.Wheres the value?The graphic at the top of the next page from our AI Benchmark report highlights a

12、n interesting aspect of this enthusiasm in that respondents expect high potential impact from adding intelligence in all three survey categories:customer experience,operations and new product development.Its unclear whether this means that there are equally exciting opportunities across the company

13、or that respondents dont have a firm idea of where the real value will come from yet.This issue of value recognition is one of several challenges for CSPs operating in Asia-Pacific.On the technical side,Level 3 onwards of TM Forums Autonomous Network maturity model,where machine learning begins to b

14、e deployed across the network,is much more challenging than earlier levels.5inform.tmforum.orgProblems include but are not limited to legacy infrastructure,data quality,lack of organizational skill sets,and several issues related to governance,data privacy and risk management.The survey conducted fo

15、r TM Forums Digital Transformation Tracker(DTT 7)gives some insight into the maturity levels of APAC CSPs,with about half not having a real plan or vision in place to tackle the coming complexity(see pie chart opposite).This research,therefore,looks at the scope of intelligence deployments that will

16、 be needed to meet corporate goals such as improving customer experience,saving operational costs and underpinning new revenue streams.Our aim with this report is to look at the progress to date in the Asia-Pacific region and consider the future steps necessary to support successful rollout of intel

17、ligence.Read the AI Benchmark to learn more about telcos AI strategies:Where will AI/ML most positively impact the telecoms business?(APAC responses)TM Forum,20240%Customer experience/CRMOperations/network operationsNew product development10%80%20%30%40%50%60%70%APAC CSPs approach to OSS/BSS automat

18、ionTM Forum,2024n We have an architectural roadmap/vision which we adhere to strictly n We are pragmatic and seek to build automation into processes as needed47%53%n Lown Mediumn Highthe need for intelligence in network lifecycle automation section 16inform.tmforum.orgNetwork lifecycle automation(NL

19、A)is the use of intelligence and automation to create end-to-end processes across multiple domains within a CSPs environment such as the RAN,IP transport network and mobile core network,and across multiple vendors solutions.Adding intelligence and automation results in many new capabilities such as

20、fault resolution for the self-healing network and prediction to allow better network planning.The six main NLA processes are shown in the graphic on the next page with a rough timeline of intelligence deployment depicted by the arrow.Rollout typically starts with descriptive intelligence used in ser

21、vice assurance scenarios and progresses downward and to the right to become prescriptive,encompassing many operational processes.The black text shows use cases with the need for machine learning while pink text highlights future use cases for Generative AI(GenAI)or other LLM tools.A description of e

22、ach process follows.Service assurance,trouble management and self-healing.The assurance process includes the deployment of AI in operations(AIOps)and other intelligence to handle the high volumes of data generated in 5G networks and to provide complex anomaly detection,correlation and causal analysi

23、s.Automation using AIOps will also be necessary to move to a self-healing network.Network management.This category contains all processes for managing and optimizing network domains,including specialized machine learning use cases in the RAN and power management.It also includes the development of i

24、ntent-based networks 7inform.tmforum.org(TM Forum defines intent as“the formal specification of all expectations,including requirements,goals,and constraints given to technical systems”.)Network deployment and maintenance.This process includes all activities enabling the deployment,commissioning,sel

25、f-configuration,testing and validation of networks including RAN-specific actions such as neighbor relations.Service orchestration.Service orchestration designs,creates,delivers and monitors service offerings,requiring automation that spans order handling,service or resource provisioning,network orc

26、hestration,policy management,and assurance.It aims to move from early automation in order handling towards the automated deployment of services on the(increasingly hybrid and complex)network and automated assurance of those services against customers service level agreements(SLAs)and intents.Adding

27、intelligence and automation results in many new capabilities.8inform.tmforum.org Network planning/capacity management.This process includes the strategic design of the network to meet current and future needs,including topological design,network dimensioning and simulations.Capacity management,requi

28、ring the automation of traffic analysis and diagnostics,forecasting of future capacity needs,and automated configuration of the network also happen here.Intelligence to support operational teams.This category brings together a range of intelligence needed to support humans working in network-related

29、 areas,including the network operations center(NOC)and service operations center(SOC),service help desk,and field services.These capabilities will often be delivered as part of a solution in one of the areas above,but Gen AI-supported use cases and other AI solutions that help humans with specific t

30、asks are included here.According to Anthony Rodrigo,Group Chief Information Officer,Axiata Group,building machine-learning models has been a“hobby”development for CSPs over the last few years.However,intelligence has been creeping into the telco in waves,driven by the successful development of indiv

31、idual models for a particular task.These tasks include:Anomaly detection,where problems that have complex data sets and invisible issues cannot be easily solved with rules-based models;in this area,machine learning has been applied first in assurance and testingTM Forum 2024(source:Charlotte Patrick

32、 Research)Examples of intelligence deployed in network lifecycle automationDescriptiveDiagnosticPredictivePrescriptiveGeneration of 3D coverage maps Selection and display of complex network configuration suggestionsML to stitch together multi-layer topology graphsNetwork data management Management o

33、f digital twinRoot cause analysis(automated)Identification of network configuration that breaches policyTroubleshootingNetwork diagnostics to find poorly performing network elements Prediction of future issues and their impactPrediction of future failure points and best resolutionPrediction of trend

34、s affecting network configurationPredictive maintenanceCapacity forecastingAutomated raising and remediation of trouble ticketsProvide instructions to move network workload to most appropriate cloud deployment Prescriptive suggestions to field services deploying cell-site assetsOptimization of cost,

35、performance and CXAnomaly detectionSupport for predictive modelsMulti-agent systemsGenerate vendor-specific instructionsAssurance,trouble management and self-healingNetwork managementNetwork deployment and maintenanceService orchestrationNetwork planning/capacity managementIntelligence to support op

36、erational teamsScheduling for field services Optimization of travel or supply chainNext-best action suggestionsAI modelsComputer visionVR or AR used on site visitsOptical character recognitionGen AIDesign suggestionsDigital assistantsCreation and maintenance of documentationDeclarative statementsCod

37、e creationData management for monitoring and reportingVisualization of device-level energy consumption on the network Diagnostics in network testingPrediction of best service to be deployedNext-best-action from simulation toolCreation/updates of topology and architecture 9inform.tmforum.org Speciali

38、st mathematical tasks such as those used in the RAN to calculate tilt on an antenna Simple predictive problems where results are fed to humans for decisioning;for example,predictive maintenance.Virtualization of network functions and service orchestration are pushing forward the current wave of inte

39、lligence rollouts.Cloud deployments come from the IT world with some intelligence in their lifecycle automation processes,and with 5G Advanced on the horizon,intelligence will be needed within the service orchestration process to support 5G network slicing.The next few years will see the broad deplo

40、yment of intelligence across NLA areas.Some use cases have solid ROI,such as the potential to save network capex and opex using models that optimize capacity or improve network design and planning.Other deployments underpin a move up the levels of the autonomous network as defined by TM Forum(see in

41、fographic left)and will take longer to roll out due to the complexity of deploying cross-domain,multi-vendor automation.Fully autonomous network:The system has closed-loop automation capabilities across multiple services,multiple domains(including partners domains)and the entire lifecycle via cognit

42、ive self-adaptation.Highly autonomous network:In a more complicated cross-domain environment,the system enables decision-making based on predictive analysis or active closed-loop management of service-driven and customer experience-driven networks via AI modeling and continuous learning.Conditional

43、autonomous network:The system senses real-time environmental changes and in certain network domains will optimize and adjust itself to the external environment to enable,closed-loop management via dynamically programmable policies.Partial autonomous network:The system enables closed-loop operations

44、and maintenance for specific units under certain external environments via statically configured rules.Assisted operations and maintenance:The system executes a specific,repetitive subtask based on pre-configuration,which can be recorded online and traced,in order to increase execution efficiency.Ma

45、nual operations and maintenance:The system delivers assisted monitoring capabilities,but all dynamic tasks must be executed manually.014325Autonomous network levelsTM Forum,2024The next few years will see the broad deployment of intelligence across NLA areas.10inform.tmforum.orgFocus on GenAIGenAI i

46、s currently mainly used in improving human interactions.A first wave of automation using non-intelligent technologies such as robotic process automation(RPA)and rules-based analytics has significantly reduced the time it takes for humans to complete repetitive tasks.This has helped CSPs make improve

47、ments in areas such as scheduling of field services.The introduction of GenAI offers new types of help including more advanced digital assistants,code creation and anomaly detection in areas such as testing.Most APAC respondents to our AI Benchmark survey said they expect the technology to make a si

48、gnificant difference to their businesses in the next few years (see pie chart opposite).The next section analyzes current deployments of AI and machine learning in Asia-Pacific.When will GenAI/LLMs have a significant impact on your business?(APAC responses)TM Forum,2024n Over the next 1-2 yearsn Bet

49、ween 2 and 5 yearsn More than 5 years58%11%32%how much AI and machine learning are CSPs implementing today?section 211inform.tmforum.orgTM Forums AI Benchmark survey finds that 65%of CSP respondents in the Asia-Pacific region agree with the statement:“We have made good progress in using AI and machi

50、ne learning in different parts of our business.”However,the reality on the ground is that a relatively small number of machine-learning models are actually in production.The graphic on the next page shows the results of our analysis into APAC CSPs use of intelligence in the six main NLA processes ou

51、tlined in the previous section.This analysis comes from primary and secondary research material from TM Forum,recent(and past)discussions with vendors and CSPs,plus publicly available vendor case studies.To conduct the analysis we split each of these six processes into multiple steps:1.Automated ser

52、vice orchestration including service provisioning,resource management and services design2.Automated and intelligent network planning and capacity management including automated planning tools3.Intent-based networks and automation in network management including domain orchestration power management

53、,RAN management and service orchestration4.Automated assurance,trouble management and self-healing including automated resource and fault management,proactive fault management,and assurance of customer experience12inform.tmforum.org5.Network deployment and maintenance including predictive maintenanc

54、e and self-configuration of the network6.Intelligence to support humans including digital assistantsThese steps were then divided further into the primary tasks to be completed in each step.For example,service provisioning requires tasks such as“automatic creation of technical work orders”and“automa

55、ted testing of service”.Each task was then considered against two questions:1.Does intelligence need to be added?If the task needed AI or machine learning,it was added to others within the process that also needed intelligence.Based on this,we calculated the percentage of tasks within each process t

56、hat need intelligence(see the primary axis shown in the graphic).2.What is the progress to date?Are the most advanced tier-1 CSPs that are claiming they have deployed the functionality really deploying it,or is it still only on vendor PowerPoint?A relatively small number of machine-learning models a

57、re in production in Asia-Pacific.13inform.tmforum.orgOur measurements of progress were as follows:1=nothing much developed,but discussions seen in industry groups and academic settings 2=some intelligence available from vendors for a single use case but not as sophisticated as it could be(meaning,th

58、ere are some diagnostic capabilities available,but a full closed-loop,prescriptive version is not yet available)3=sophisticated intelligence is available from vendors 4=deployed in a production network for certain domains or use cases but not everywhere;or deployed by a few tier-1 CSPs only 5=widely

59、 deployed.The results shown in the graphic reveal that a significant proportion of processes underpinning NLA(about 60%to 70%)could benefit from the addition of intelligence.The noticeable dip for service orchestration is due to the availability of simpler automation techniques that only require rul

60、es-based decisioning.The black line on the chart,which shows progress to date,gives a score of 3.84 to assurance for the quite sophisticated use of AIOps in many assurance deployments.This drops for the other processes,with some having pockets of deployment but many proving challenging to deploy.The

61、 final area of intelligence used to support humans has suddenly been brought into focus with the availability of GenAI but is still relatively immature in deployments so far.The next section expands on this analysis.TM Forum 2024(source:Charlotte Patrick Research)Progress of intelligence deployment

62、in NLA100%90%0%10%20%30%40%50%60%70%80%of processes requiring intelligence Progress seen to date1.Assurance,trouble management and self-healing2.Network management3.Network deployment and maintenance4.Service orchestration5.Network planning/capacity management6.Intelligence to support humans5.004.50

63、2.004.003.503.002.50examples of intelligence in network lifecycle automationsection 314inform.tmforum.orgThis section of the report builds on our assessment of the use of intelligence in network lifecycle automation,providing examples from CSPs and suppliers operating in the Asia-Pacific region.15in

64、form.tmforum.org1.Assurance,trouble management and self-healingSeveral trends and requirements are impacting the deployment of intelligence in assurance:Observability across customers and domains.The last few years have seen continuing improvement in network resilience mainly from the adoption of vi

65、rtualization and automation.This means less focus is required on individual network issues in the NOC.Along with preparations for new services that take advantage of 5G standalones cloud-native core,these improvements have shifted the focus to service assurance.Data from the network and many other s

66、ources(test data,weather patterns and customer sentiment,for example)is needed in service assurance,alongside new machine-learning models to deal with this increasing volume of data and to undertake anomaly detection,prediction and optimization.In addition,virtualized networks require assurance to s

67、pan both enterprise IT and traditional networks,and 5G will increasingly blur the line between the two as enterprise workloads run at the networks edge.Machine learning models are needed to deal with the volume of data from multiple domains.Self-healing networks.Self-healing networks require a range

68、 of monitoring solutions that observe the network against a set of key performance indicators(KPIs)and benchmarks.Data is collected when anomalies are detected,and intelligence is used to predict and prescribe remediation.Self-healing capabilities require active assurance using time-series databases

69、,with storage and monitoring happening in a distributed architecture,to support massive amounts of data from multiple domains.There is also a need for machine learning to predict and prescribe the best way to resolve issues in near-real time.Getting assurance data to the right destination.To meet en

70、terprise customers requirements for new services and to ensure satisfaction for all customers,CSPs must integrate their data and operations across diverse and complex ecosystems.Delivering the right data(a mix of assurance,inventory and external third-party data)to the right people and processes at

71、the right time will provide new visibility and automation to support these new services.The last few years have seen improvements in network resilience from the adoption of virtualization and automation.16inform.tmforum.orgAn example of AI-driven service assurance from Thailands AIS is featured in T

72、M Forums Autonomous Networks Journey Guide.AIS designed a proactive approach that consolidates near-real-time customer experience data,network events and process insights using AI models to enable proactive intervention and optimization.The strategic initiative aims to promptly detect and address po

73、or customer experience,utilizing intelligent and automated integration to identify problem areas or potential root causes and integrate this capability into the complaint process to resolve customers network problems and improve satisfaction.2.Network managementIntelligence in network management is

74、needed across domains such as:RAN management.This has been an area of progress in intelligence deployment in the last few years,particularly with the addition of new machine learning to solve complex problems in near-real time.It includes RAN-specific activities such as beamforming,channel estimatio

75、n,interference management,handover optimization,dynamic spectrum sharing,scheduling,random access channel(RACH)and physical cell ID optimization.The addition of new machine learning and automation offers an opportunity to reduce opex and improve customer experience.Published case studies from networ

76、k equipment providers offer examples,although the capabilities are most likely not operational outside the largest global CSPs.An example of a CSP using intelligence in RAN management again comes from AIS,which must deal with frequent power interruptions during Thailands rainy season.Storms trigger

77、thousands of cell outages monthly,causing degradation of customer experience and revenue loss.AIS has introduced the Cell Outage Detection and Compensation(CODC)solution to ensure swift detection of cell outages,precise coordination of compensation and optimized network performance(see graphic).1.Ou

78、tage detection2.Cell recovery(D-SON)4.Roll back3.Compensation(MAE-OPT)Action Cell recovery Move UE to other cellSUCCEEDENGINEER REPAIRFAILTypically recover outage cell 60%areaFast RF Adjust(15 min)Remote repair in cell out-of-service scenariosTM Forum,2024(source:AIS)17inform.tmforum.orgThe solution

79、 constitutes a remote repair mechanism,which orchestrates a near-real-time response to cell outages after triggering an alarm when a cell outage is identified.The system determines the exact point of traffic impact and selects the most suitable neighboring cell for load balancing.Mechanisms for opti

80、mization of radio frequency coverage are put in place,along with ongoing monitoring of KPIs.Power management.This is a popular area for intelligence deployment,given its ability to decrease opex in an era of rising energy costs.A proof of concept from Globe Telecom in the Philippines showed how to a

81、utomatically switch off idle equipment when not in use,using AI and machine learning to understand the energy utilization pattern of network elements.Using energy efficiency software-as-a-service,Globe Telecom was able to achieve the equivalent of annual power savings ranging from 3%to 6%.The soluti

82、on“uses AI/ML to learn the energy utilization pattern of each network element,like the base station”,Gerard Ortines,Head of Network Solutions and Capex Management at Globe Telecom,explained to TM Forum.“In turn,it reduces energy cost without compromising the network performance and is easy to implem

83、ent The solution enables dynamic energy consumption through advanced algorithms based on real-time network traffic while maintaining premium mobile user experience,”he added.In addition to device-level power management,CSPs are looking at network power balancing,which uses traffic steering to provid

84、e the optimum network footprint for energy-saving,as well as management of power-related assets such as batteries or air conditioning and management of power backup to ensure optimum battery backup for emergencies.Intent-based networks.The process of deploying intent-based networks(see sidebar on p.

85、18)has started with services like network slicing.Still,automation capabilities,new intelligence and orchestrators are required across the network to increase the kinds of intents that can be fulfilled.Intelligence will be required in intent-based networking to define the behavior expected from the

86、network,creating end-to-end service goals and KPIs for each domain.It is also needed to dynamically optimize and remediate services;to deal with multiple,potentially conflicting intents;to calculate the benefit to customer experience or cost in taking a particular action;and to control risk by predi

87、cting the outcome of decisions.Digital Nasional Berhad(DNB),a special-purpose vehicle company and wholesale mobile operator established in 2021 to serve the six national operators of Malaysia,is implementing intent-based network operations,with a goal of ensuring that 5G connectivity is affordable f

88、or Malaysians.DNB is adding automation and intelligence to network management,and the deployment is delivering impressive results.Learn more about network automation use cases in TM Forums Autonomous Networks Journey Guide:1 Back to ContentsObjective:to share the latest updates of business requireme

89、nts,architecture,capabilities and collaborative activities,best practices,and reference implementations on Autonomous Networks to enable value-driven transformation.AN JOURNEY GUIDEEmpowering digital transformation evolving from Level 2/3 towards Level 4 SEPTEMBER 2023NetworksAutonomous18inform.tmfo

90、rum.orgFor example,auto-analysis and correlation of alarms and actuation in the network led to a 500%reduction in alarm count six months after its introduction,and AI-based predictive 5G network management using network data resulted in network uptime greater than 99.8%.DNB implemented a trial of a

91、new intent-based model to handle RAN partitions for each RAN cell site,while also managing the networks resources.This resulted in:8%to 10%better throughput of premium services with no intent violations Avoidance of 12 SLA breaches per cell per day 15 to 17 hours of partition overuse prediction shar

92、ed with the customer per week Network slices with individual guaranteed SLAs.3.Network deployment and maintenanceOpportunities to deploy AI and machine learning to assist with network deployment and maintenance can be divided into three main groups of use cases:Planning and deployment management too

93、ls.These tools enable all processes around site deployment from site selection and surveying to supply chain management.Vision recognition tools,for example,gather images and information from drones about cell site configurations.This data can then be input into a 3D model or digital twin of the sit

94、e,enabling engineers to make engineering decisions remotely based on the model.In TM Forums Autonomous Networks Project,CSPs and their suppliers are working on how to implement intent-based automation and integration between autonomous domains.They are building on and extending previous standards wo

95、rk on intent wherever possible.For example,the teams definition of intent-based networking is compatible with that of the Internet Engineering Task Force(IETF).The systems in autonomous networks are governed by intent,which sets expectations as requirements,goals and constraints.These are abstracted

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

97、achine learning and AI,while enabling supplier innovation that can be easily integrated and adopted.Intent can be expressed at various levels(business intent,service intent and resource intent,for example).In the TM Forum AN Reference Architecture intents are used with closed-loop management to auto

98、mate the full lifecycle of services.Closing the loop means collecting and analyzing data to figure out how networks can be optimized,and then implementing those changes in an automated way.To learn more about TM Forums work on intent and autonomous networks,please contact Alan Pope.Defining intent a

99、nd showing how to implement it19inform.tmforum.orgAnalytics and machine learning can also optimize tools for the allocation of design and field resources needed for making changes and emergency repairs to the network.Coverage optimization tools allow antenna attributes such as height,tilt and azimut

100、h to be calculated from a three-dimensional photograph,and tools for capacity auditing and acceptance compare sites current performance with baseline goals,capturing possible issues such as installation problems or equipment defects.Ongoing site management tools.Intelligence solutions in this area i

101、nclude“humans in the loop”with use cases such as technicians running power into a cell site,installing equipment,connecting cables and troubleshooting connectivity.Testing.CSPs are beginning to include testing as part of closed-loop service and network orchestration,allowing end-to-end service orche

102、stration and rapid deployment of network functions and devices.Intelligence provides anomaly detection for diagnostics and a range of predictions,including future performance and potential failure scenarios.4.Service orchestration Automation has been available for some time within certain parts of t

103、he service lifecycle such as service assurance,but the end-to-end provisioning and deployment of the service on the network depends on the availability of suitable orchestration and the ability to modify multiple network elements.With the introduction of service orchestration solutions,vendors are a

104、rticulating intelligence use cases for:Prescriptive and predictive capabilities to enable deployment of the complex set of new network technologies,devices and features Detection of unusual patterns or outliers in inventory data Understanding patterns of change in the inventory database to provide i

105、nformation on how the network is evolving The ability to define new services using GenAI humans would use declarative statements to provision new services,such as“provide a connectivity solution for secure communication from point A to point B”.Telstra has been using a service orchestration solution

106、 and TM Forum Open APIs to integrate new domains such as cloud and SD-WAN with established wireless and transport domains,with an average seven day shorter order fulfilment time.Service composition is first defined in a declarative template,with the solution automatically designing the service based

107、 on specified parameters such as latency before instantiating it.It can modify or terminate the service using workflows dynamically generated from the service template and provide end-to-end management and visibility of services.20inform.tmforum.orgClosed-loop operation then enables the composite do

108、main to maintain the required service level throughout the lifecycle of the service as conditions and capabilities change in the underlying domains.5.Network planning/capacity managementPlanning and capacity management are areas where intelligence and automation can result in significant capex and o

109、pex savings.The past four to five years have seen the emergence of the first tools adding intelligence to the planning teams activities.The infographic opposite describes the typical use cases for intelligence in network planning and capacity management.Sanders notes that Telstras team tended to be

110、very experienced but was still using multiple manual planning and design documents,which created a large labor cost to support.6.Intelligence to support operations teamsSupport for operations teams historically has included RPA to complete simple tasks at the right moment for a human.These capabilit

111、ies have expanded over time to take on many repetitive tasks,particularly in large teams where the time savings add up to a significant cost benefit.The appearance of GenAI offers new functionality via simple,digital assistants.Principally the use of natural language understanding and generation bri

112、ngs a new sophistication in human-machine communications.There are no publicly available examples yet of CSPs using GenAI or other LLMs in network operations.Indeed,TM Forums research for the AI Benchmark found that network operators are largely skeptical about the technologys usefulness in the day-

113、to-day running of telecoms networks.Intelligence used in network planning and capacity managementVisualization&DesignMachine learning to provide an understanding of the network and optimum connectivity of network components based on factors like cost or customer experienceImproved network dimensioni

114、ng (determining how to meet minimum capacity requirements at peak hour)Simulation tools including what-if analysis and future use of digital twin Analysis&DiagnosticsDetailed traffic analysis to find high-value areas and bottlenecksNetwork diagnostics providing real-time fault analysis and understan

115、ding poorly performing network elementsLoad balancing predictions used in demand forecasting and long-term predictions of network behaviorOptimizationSophisticated models for optimization(e.g.,a desired base station location is unavailable due to a lack of electricity or the landowner refusing acces

116、s;this means that the optimal network may not be possible,but machine learning can be used to revise network designs based on limitations)RAN PlanningReinforcement learning(e.g.,the autonomous configuration of tilt and power of antennas in a business district to achieve the best possible coverage or

117、 quality per the mobility patterns at that time)TM Forum 2024(source:Charlotte Patrick Research)21inform.tmforum.org“Ive met a lot of network experts in this industry that are trying to find some real-time data monitoring or root cause analysis use cases for GenAI.But I dont think its possible at th

118、is moment,”Chung Suk-guen,Chief Global AI Officer at SK Telecom,said during an interview for the report.“GenAI is not fast enough to handle tons of data in real time.And also,the data is not human readable.”That said,there is a range of(mostly)AI-related functionality that could be added in the futu

119、re to support operations teams,such as:Support for more complex problems.Increasingly,intelligence will be built as digital assistants to support more complex problem solving.These will use data from inside and outside the organization,as needed.More intelligent interactions.Here,machines start to i

120、nteract on a more ad-hoc basis,offering suggested corrections(or making automatic corrections)where human error has crept in.A machine may also proactively undertake tasks as it perceives a need based on human activity in the recent past.Support for specialist tasks.GenAI is used to support declarat

121、ive statements(allowing humans to express their high-level objectives in natural language).It can also be used for code creation(for example,real-time creation of API script),or it may be involved in the creation and maintenance of documentation.Onsite technology.There are a range of newer AI techno

122、logies for field technicians undertaking tasks such as running power,installing equipment connecting cables and troubleshooting connectivity,which require intelligence.For example,video-based tower inspections can be conducted using a drone to stream video to the ground operators mobile device.This

123、uses computer vision to detect issues with the radio tower installation in real time.Computer vision can also identify potential problems with cable weatherproofing or color coding(to identify crossed cables).This AI can be augmented with a convolutional neural network(CNN)to identify“bad”versus“goo

124、d”cabling installations.And optical character recognition can locate the sticker or tag area in images to bring back information about the details of installed equipment.Finally,deep neural networks could also check for installation issues:Are cables neatly bound?Is equipment that was expected to be

125、 deployed evident in photos of the base station taken as the engineer leaves?The next section offers CSPs advice for increasing deployment of AI and machine learning in operations.Increasingly,intelligence will be built as digital assistants to support more complex problem solving.make it happen str

126、ategies to take advantage of new intelligencesection 422inform.tmforum.org23inform.tmforum.orgCSPs operating in the Asia-Pacific region are shifting the focus of intelligence deployment towards the task of industrializing AI practices.In doing so,they are moving from an early phase of experimenting

127、in multiple AI projects to developing and implementing an AI and machine learning strategy that enables them to make good decisions commercially and technically.Here are four actions that require immediate attention.Determine the real benefits of implementationCreating intelligence roadmaps is diffi

128、cult.A large CSP may have developed hundreds of machine learning models,of which only a small handful are currently in useful deployment.One executive we spoke with commented on the disjointed approach to AI that is often seen today and the tendency to focus on the next new and exciting technology.I

129、ntelligence projects should be selected based on a mixed set of measurements,including:Financial value to the organization,both individually and as part of the total vision for network lifecycle automation Immediate feasibility given the maturity of models and experience within the organization Imme

130、diate usefulness for building knowledge and experience for example,small domain-specific automation that includes machine learning may not offer a huge financial upside but could provide an opportunity to build intelligence skill sets without needing complex cross-domain data gathering Place within

131、a particular value stream again,the project may not bring much immediate financial value but may be an integral part of a wider NLA goal that will fail without it Board-level enthusiasm many early intelligence projects have failed to develop the necessary models,so the highest levels of management m

132、ay need to evangelize AI more widely across the company.Take a mature approach to road map developmentBuilding out a new technology without an overarching vision risks creating a whole new layer of technical complexity and debt.CSPs should take a step back to understand any weaknesses in their curre

133、nt road maps for using intelligence.Questions to ask include:24inform.tmforum.org Have we selected the simplest possible solution to each problem,given that this may also be the most cost effective and have the highest chance of success?Are we moving swiftly down the path of creating reusable data a

134、ssets?Without these there is a risk of creating silos of data or duplicating requests from the network or other equipment for the same data.Do we need a plan(in conjunction with others in the industry)to introduce a more open ecosystem around intelligence that limits the reduction in the value of ou

135、r data by allowing third parties,such as vendors,to benefit from ingesting it into their models?Is there an evaluation process that removes in-progress projects where the intelligence is failing?This could be because the problem requires more mature models than used or because the models are proving

136、 problematic(for example,they are prone to creating subtle and difficult-to-spot errors).Create a robust intelligence disciplineOne of CSPs key strengths is their background in developing sophisticated capabilities in network engineering.This kind of thinking will also be needed when deploying intel

137、ligence into NLA.The creation of a small,centralized team supporting intelligence and automation efforts should be considered.The team might include technical experts and business-focused employees responsible for ongoing development of vision and strategy,business cases,and risk management.An examp

138、le of this type of team has been running in Telstra for the last five months.The team begins by reaching out to areas where intelligence and automation could be deployed to answer questions and demonstrate potential value to skeptical team members.It brings technical expertise(as needed)as well as c

139、ommercial expertise,drafting business cases and analyzing the operational benefits of projects.One of CSPs key strengths is their background in developing sophisticated capabilities in network engineering.25inform.tmforum.orgStaff issues,including insufficient headcount or expertise,remain a signifi

140、cant headache for APAC CSPs.Build an intelligence-native teamAs the dust settles around the hype of GenAI,CSPs are beginning to understand the technologys uses and limitations(some of which may be permanent or temporary until more sophisticated or different models become available).This knowledge,ba

141、cked up with practical experience from deploying AI models into NLA will need to be instantiated in the organization and spread across departments to reuse it.Actions might include encouraging individuals with an early interest in testing intelligence technologies to get together in informal groups

142、to explore topics and exchange ideas.New job titles should be considered for addition to the organization.For example,the“knowledge engineer”will obtain understanding from the process expert in the operating unit and incorporate this human understanding into the intelligence within automation or dig

143、ital assistants.Staff issues,including insufficient headcount or expertise,remain a significant headache for APAC CSPs,particularly smaller organizations.Our research finds that there is not one agreed-upon path to overcoming these problems,with CSPs expecting to build up new skill sets through recr

144、uitment,reskilling,and acquiring or partnering with specialist AI firms.Our AI Benchmark report discusses some of the new AI skills needed,and an upcoming report focusing on how telco-to-techco transformation impacts people will look at where CSPs can find the talent they need.tm forum open digital

145、framework26inform.tmforum.orgThe TM Forum Open Digital Framework provides a migration path from legacy IT systems and processes to modular,cloud native software orchestrated using AI.The framework comprises tools,code,knowledge and standards(machine-readable assets,not just documents).It is deliveri

146、ng business value for TM Forum members today,accelerating concept-to-cash,eliminating IT and network costs,and enhancing digital customer experience.Developed by TM Forum members through our Collaboration Community and Catalyst proofs of concept and building on TM Forums established standards,the Op

147、en Digital Framework is being used by leading service providers and software companies worldwide.Core elements of the Open Digital FrameworkThe framework comprises TM Forums Open Digital Architecture(ODA),together with tools,models and data that guide the transformation to ODA from legacy IT systems

148、 and operations.Open Digital Architecture Architecture framework,common language and design principles Open APIs exposing business services Standardized software components Reference implementation and test environmentTransformation tools Guides to navigate digital transformation Tools to support th

149、e migration from legacy architecture to ODAMaturity tools&data Maturity models and readiness checks to baseline digital capabilities Data for benchmarking progress and training AIGoals of the Open Digital FrameworkThe Open Digital Framework aims to transform business agility(accelerating concept-to-

150、cash from 18 months to 18 days),enable simpler IT solutions that are easier and cheaper to deploy,integrate and upgrade,and to establish a standardized software model and market which benefits all parties(service providers,vendors and systems integrators).A blueprint for intelligent operations fit f

151、or the 5G eraLearn more about collaborationIf you would like to learn more about the project or how to get involved in the TM Forum Collaboration Community,please contact George Glass.27inform.tmforum.orgtm forum research reports28inform.tmforum.org29inform.tmforum.orgOctober 2022|www.tmforum.orgAut

152、hor:Teresa Cottam,Contributing AnalystEditor:Dawn Bushaus,Contributing Editorsponsored by:from transformationDIGITAL OPERATIONS MATURITY:achieving business valueREPORTAuthors:Dean Ramsay,Principal Analyst Rahul Gupta,Senior AnalystEditor:Ian Kemp,Managing EditorSPONSORED BY:November 2022|inform.tmfo

153、rum.org5Gcore:exploring CSPapproachesREPORTAuthor:Mark Newman,Chief AnalystEditor:Ian Kemp,Managing EditorSponsored by:Supported by:evolvingfor future servicesbusiness support systems December 2022|inform.tmforum.orgREPORTAuthor:Joanne Taafe,Editor in Chief,InformEditor:Ian Kemp,Managing Editor,TM F

154、orumSponsored by:thesustainable telco:engineeringnetworks for net zeroDecember 2022|inform.tmforum.orgREPORTAuthors:Mark Newman,Chief Analyst,TM ForumDawn Bushaus,Contributing Analyst,TM ForumSponsored by:Editor:Ian Kemp,Managing Editor,TM Forumestablishing links:platform models in the Open API econ

155、omy March 2023REPORTAuthor:Sponsored by:Teresa Cottam,Contributing AnalystDawn Bushaus,Contributing Editorcounterusing AI to improve customer experienceintelligenceBENCHMARKAuthors:Mark Newman,Chief Analyst,TM ForumDawn Bushaus,Contributing Analyst,TM ForumJoanne Taafe,Editor in Chief,InformEditor:I

156、an Kemp,Managing Editor,TM ForumSponsored by:a roadmap fortelecomsgrowthTM Forum|March 2023Asia-PacificREPORTAuthor:Ed Finegold,Contributing AnalystDawn Bushaus,Contributing EditorTM Forum|May 2023who does what and can CSPs compete for more?partner ecosystems:REPORTEditor:Author:Sponsored by:Dean Ra

157、msay,Principal Analyst,TM ForumIan Kemp,Managing Editor,TM Forummakingwaves:the future for Open RAN technologyJune 2023|inform.tmforum.orgREPORTDigital Transformation Tracker 7 with automation and Author:Dawn Bushaus,Contributing EditorEditor:Ian Kemp,Managing EditorSponsored by:TM Forum|June 2023cu

158、tting complexity AICSPs take key steps to modernize network inventoryREPORTAuthor:Mark MortensenEditor:Dawn BushausISBN:Sponsored by:June 2023|www.tmforum.orgAuthor:Dr.Mark H.Mortensen,Contributing AnalystEditor:Dawn Bushaus,Contributing Editorsponsored by:Sponsored by:Author:Editor:ISBN:REPORTtrans

159、formingBSS:racing to a flexible,customer-focused futureJune 2023|www.tmforum.orgAuthor:Teresa Cottam,Contributing AnalystEditor:Ian Kemp,Managing Editorsponsored by:Sponsored by:Author:Editor:ISBN:REPORTstandout strategies:how telcos are innovating in a crowded marketJune 2023|www.tmforum.orgAuthor:

160、Mark Newman,Chief AnalystEditor:Ian Kemp,Managing Editorsponsored by:August 2023|www.tmforum.orgAuthor:Ed Finegold,Contributing AnalystEditor:Ian Kemp,Managing Editorsponsored by:how software-as-a-serviceis reshaping business support systemsReignitingtelecoms growthAugust 2023|www.tmforum.orgsponsor

161、ed by:Authors:Sangeet Paul Choudary,Platform Thinking LabsNik Willetts,CEO,TM ForumAnthony Rodrigo,CIO,AxiataDean Ramsay,Principal Analyst,TM Foruma Playbook for CEOsSeptember 2023|www.tmforum.orgAuthor:Mark Newman,Chief AnalystEditor:Dawn Bushaus,Contributing Editor sponsored by:wholesalechanges:re

162、thinking support systemsfor new fiber operatorsSponsored by:Author:Mark Newman Chief Analyst Editor:Dawn Bushaus,Contributing Editor REPORTrethinking support systems for new fiber operatorsSeptember 2023|URL TBCwholesalechanges:Sponsored by:Author:Mark Newman Chief Analyst Editor:Dawn Bushaus,Contri

163、buting Editor REPORTrethinking support systems for new fiber operatorsSeptember 2023|URL TBCwholesalechanges:BENCHMARKAuthors:Mark Newman,Chief AnalystDean Ramsay,Principal AnalystEditor:Ian Kemp,Managing EditorTM Forum|September 2023telcorevenuegrowth:time foroperators toplace new betsSeptember 202

164、3|www.tmforum.orgAuthor:Patrick Donegan,Principal Analyst,HardenStanceEditor:Dawn Bushaus,Contributing Editor,TM Forumsponsored by:Sponsored by:Author:Patrick Donegan Principal Analyst HardenStanceEditor:Dawn Bushaus,Contributing Editor TM Forum REPORTrisk management moves firmly into the telco spot

165、lightcybersecurity strategies:September 2023|URL TBCrisk management moves firmly into the telco spotlightCybersecurity strategies:REPORTSponsored by:leveling up:Author:Mark Mortensen,Contributing AnalystEditor:Dawn Bushaus,Contributing Editor achieving Level 3 autonomous networks and beyondAugust 20

166、23September 2023|www.tmforum.orgAuthor:Dr.Mark H Mortensen,Contributing Analyst,TM ForumEditor:Dawn Bushaus,Contributing Editor,TM Forumsponsored by:REPORTSponsored by:leveling up:Author:Mark Mortensen,Contributing AnalystEditor:Dawn Bushaus,Contributing Editor achieving Level 3 autonomous networks

167、and beyondAugust 2023REPORTTM Forum|December 2023operators take their first stepsAuthor:Mark Newman,Chief Analyst Editor:Ian Kemp,Managing Editor Sponsored by:generativeAI:REPORTTM Forum|December 2023Author:Ed Finegold,Contributing Analyst Editor:Dawn Bushaus,Contributing Editor Sponsored by:BSSfor

168、B2Boperators diverge on the path to cloudREPORTTM Forum|December 2023the sustainable telco:navigating the maze of scope 3 emissionsREPORTa numbers game:February 2024Sponsored by:Author:Richard Webb,Senior AnalystEditor:Ian Kemp,Managing Editor exploiting data to drive future strategiesFebruary 2024|

169、www.tmforum.orgsponsored by:Author:Richard Webb,Senior AnalystEditor:Ian Kemp,Managing EditorREPORTclosing the loopFebruary 2024Sponsored by:Author:Dean Ramsay,Principal AnalystEditor:Ian Kemp,Managing Editor CSPs aim to automate service orchestration and assuranceFebruary 2024|www.tmforum.orgsponso

170、red by:Author:Dean Ramsay,Practice Lead,TM ForumEditor:Ian Kemp,Managing EditorREPORTclosing the loopFebruary 2024Sponsored by:Author:Dean Ramsay,Principal AnalystEditor:Ian Kemp,Managing Editor CSPs aim to automate service orchestration and assurancemeet the research&media team30inform.tmforum.orgP

171、ublished by:European Office TM Forum,Uncommon,34-37 Liverpool Street,London EC2M 7PP United Kingdom Phone:+44 207 748 6615 US Office 181 New Road,Suite 304,Parsippany,NJ 07054 USA Phone:+1 862-227-1648 ISBN:978-1-955998-81-9 Report Design:Paul Martin 2024.The entire contents of this publication are

172、protected by copyright.All rights reserved.The Forum would like to thank the sponsors and advertisers who have enabled the publication of this fully independently researched report.The views and opinions expressed by individual authors and contributors in this publication are provided in the writers

173、 personal capacities and are their sole responsibility.Their publication does not imply that they represent the views or opinions of TM Forum and must neither be regarded as constituting advice on any matter whatsoever,nor be interpreted as such.The reproduction of advertisements and sponsored featu

174、res in this publication does not in any way imply endorsement by TM Forum of products or services referred to therein.31inform.tmforum.orgMeet the Research&Media teamReport Author:Charlotte Patrick Contributing Analyst Managing Editor:Ian Kemp ikemptmforum.orgChief Analyst:Mark Newman mnewmantmforum

175、.orgPractice Lead:Dean Ramsay dramsaytmforum.orgHead of Operations:Ali Groves agrovestmforum.orgCommercial Manager:Tim Edwards tedwardstmforum.orgReport Editor:Dawn Bushaus Contributing Editor dbushaustmforum.org Editor in Chief,Inform:Joanne Taaffe jtaaffetmforum.orgGlobal Account Director:Carine Vandevelde cvandeveldetmforum.orgMarketing Manager:Ritika Bhateja rbhatejatmforum.orgDigital Media Coordinator:Maureen Adong madongtmforum.org To find out more about TM Forums AI and data work please contact Yvonne Kuimba