《電信管理論壇：2024混合云經濟學管理報告：實現平衡優化運營 （英文版）（32頁）.pdf》由會員分享，可在線閱讀，更多相關《電信管理論壇：2024混合云經濟學管理報告：實現平衡優化運營 （英文版）（32頁）.pdf（33頁珍藏版）》請在三個皮匠報告上搜索。

1、TM Forum|March 2024striking abalance tooptimiseoperationsAuthor:Charlotte Patrick,Contributing Analyst Editor:Dawn Bushaus,Contributing Editor Sponsored by:Managing theeconomics ofhybrid cloud:contents2inform.tmforum.orgsection 2:CSPs current cloud deploymentssection 1:CSPs embrace hybrid cloudsthe

2、big pictureadditional features&resources 07050326section 4:the economics of capex vs.opexsection 3:public vs.private cloud130916section 5:the cost of deploying private cloudssection 6:navigating the road to public cloud18section 8:make it happen hybrid cloud strategies to optimize costssection 7:CSP

3、s prefer private clouds for network functions2421the big pictureInitially,CSPs implemented public cloud strategies to reduce costs by transitioning from expensive on-premises infrastructure while enabling rapid capacity scaling and access to new developer ecosystems and tools like AI.However,it beca

4、me evident that the cloud can be costly if not utilized effectively.Over the past two years,a typical technology investment cycle has been characterized by periods of significant spending on new cloud solutions,followed by retrenchment and repatriation of a percentage of the workloads.“I talk to a l

5、ot of companies that fail,and it is because they have not understood that the cloud is not a virtual data center,”says Stephane Demartis,VP of Infrastructure at Orange Innovation.“Workloads have to be optimized for the public cloud.”Then,operators typically adjust their cloud strategies to create a

6、more balanced approach that harnesses the benefits of both public and private cloud(see infographic on p.4).Our discussion in this report is framed within the topic of hybrid cloud on the assumption that it is the preferred long-term architecture for most CSPs.Conversations surrounding communication

7、s service providers(CSPs)cloud deployments often take on an existential quality because of their importance.Such deployments are integral to shaping future network architecture,driving digital strategy,enhancing agility and redefining the operators role within new ecosystems.However,achieving this t

8、ransformation involves balancing cost considerations against performance and strategic decisions,including the complex issue of simultaneously collaborating and competing with hyperscale cloud providers.3inform.tmforum.orgI talk to a lot of companies that fail,and it is because they have not underst

9、ood that the cloud is not a virtual data center.”Stephane Demartis,Orange Innovation4the big pictureinform.tmforum.orgCloud economics is often tightly defined as the financial analysis and optimization of cloud resources within the context of decision-making about deployment of a specific workload.H

10、owever,making longer-term decisions about cloud deployment requires economic analysis plus multiple architectural,operational and strategic choices.Therefore,this research looks at how cloud economics will be impacted by cost and a range of other factors in the mid-to long-term and how good decision

11、ing can satisfy not only the CFO,but also the CIO,CTO and CEO.As part of our research,we surveyed 79 people working within CSPs worldwide about their use of public and private cloud hosting.The respondents are a subset of participants in a broader survey for our recent report BSS for B2B:operators d

12、iverge on the path to cloud.Most respondents are directly responsible for operational and business support systems(OSS/BSS)or enterprise IT architecture,giving them a clear view into their organizations plans for support system architecture and choice of supplier.In addition to the survey,we conduct

13、ed interviews with CSP executives and suppliers.Read this report to understand:Why hybrid cloud deployment is key The relative benefits of public versus private cloud The future economics of public cloud usage The movement of networks into the cloud and management of private cloud costs How to imple

14、ment a hybrid cloud strategy.TM Forum,2024Cloud optionsPublic Pool of IT resources automatically provisioned and allocated among multiple clients through a self-service interface Designed to scale out workloads based on demand Deployed as part of a heterogeneous mix of environments,leading to higher

15、 security and performance and wider availability of services and applicationsMulticloud Similar to hybrid with combinations of private and public cloud but using multiple cloud providers;so,a CSP could have AWS,Google and Microsoft clouds and move IT assets around in themHybrid Incorporates some deg

16、ree of workload portability,orchestration and management across 2 or more environments Includes at least 1 private cloud,1 public cloud,and a bare-metal or virtual environment connected to at least one of the cloudsPrivate Dedicated solely to the user and is usually within a companys firewall Runs o

17、n premises or in rented,vendor-owned data centers located of premises Underlying IT infrastructure is dedicated with completely isolated access5section 1:CSPs embrace hybrid cloudsinform.tmforum.orgThe graphic below illustrates the mid-to long-term factors pushing CSPs towards a hybrid cloud strateg

18、y.A brief explanation of the drivers follows in this section,and well look more closely at those drivers and the costs of public versus private cloud throughout the report.Future demand-side factorsPublic where useful private where necessary:public cloud services are scalable,enabling CSPs to expand

19、 infrastructure in response to demand in both the IT and network domains,and they underpin new business models and service innovation.Private clouds,on the other hand,better address the privacy,security and regulatory requirements of workloads supporting latency-sensitive network services and high-v

20、olume services that would be expensive to deploy in the public cloud.This balancing act requires the orchestration capabilities of a central platform to manage cloud costs while ensuring that performance and Decisions about next-generation cloud infrastructure must meet the economic expectations of

21、the telco CFO while simultaneously addressing requirements of other top executives for an agile and usable cloud.In the hybrid cloud environment,this means managing workloads that traverse diverse deployment locations public,private and on-premises to achieve the optimal balance between cost efficie

22、ncy,performance,security and regulatory compliance,while meeting the needs of developers and other users.Successfully managing hybrid cloud economicsTM Forum,2024Future demand-side factors afecting CSPs cloud deploymentsPublic where useful private where necessaryPublic-clouds are highly scalable and

23、 support new business models and innovation.Private clouds are more secure and support low-latency servicesPublic cloud by choice or necessityCost is key,alongside the need to influence hyperscalers roadmaps;small or greenfield operators are likely to adopt a public cloud-first strategy Managing clo

24、ud economics is not straightforwardCSPs must understand hyperscalers changing products and fluid pricing/discount structures,plus the cost of an expanding data center footprintNew strategy to manage increasing volumes of dataAI and cloud-native networks may push operators to adopt private cloudsFutu

25、re supply-side factors afecting CSPs cloud deploymentsMid-to long-term margin managementHyperscalers are under pressure to demonstrate ROI for vast cloud outlays afecting margins,alongside other factors like increasing energy prices and supply chain issuesMarket maturity and regulatory pressureEcono

26、mies of scale among hyperscalers are expected to be of interest to competitive authorities,although the market looks set to stabilize at a mix of hyperscalers and smaller niche providersRelationship between CSPs and hyperscalersCSPs are likely to continue collaborating and competing with hyperscaler

27、sHybrid cloud capabilities requiredManagement tools that have a singlepane of glass and provide a 360-view of workload deployment across the cloud footprintAbility to orchestrate workloads,enabling close management of cost vs.performanceSupport for FinOps for short-termdecisions and longer-term mode

28、llingAbility to trial new AI servicesthat have unknown ROIChoice in managing competitivesituation with other cloud players6section 1:CSPs embrace hybrid cloudsinform.tmforum.orgagility are maintained.For example,one large European operator we interviewed elected to move customer-facing workloads to

29、the public cloud about eight years ago,not to reduce costs but to take advantage of agility benefits.Public cloud by choice or necessity:small or greenfield operators that are looking to reduce time to market and avoid large capital outlays are likely to opt for public cloud,as are some large operat

30、ors that have taken a top-level,strategic decision to work closely with the hyperscalers:Amazon Web Services(AWS),Google Cloud Platform and Microsoft Azure.Vendor lock-in is a potential downside to this approach as a highly complex migration strategy is required.AT&T is a good example of a large ope

31、rator taking a strategic public cloud approach.In June 2021,the company sold its private telco cloud platform to Microsoft Azure,with development teams from AT&Ts Network Cloud organization moving to Azure for Operators.According to a blog post from Microsoft,this marked the first time a tier-one op

32、erator had deployed a hybrid cloud focused on the mobile network core rather than more typical go-to-market collaborations at the edge.Given AT&Ts previous use of the do-it-yourself private cloud approach,the partnership with Microsoft has been widely viewed as supportive of the public cloud model.M

33、anaging cloud economics is not straightforward:from a unit-cost perspective,a well-managed cloud deployment requires a team that understands hyperscalers products,pricing and discount structures,which are all fluid(see box on p.14),as well as the economics of the data center footprint.Additionally,t

34、hey need to be able to exert control over the users choice to upload new workloads into the cloud and which cloud they select.Our survey data suggests that CSPs are becoming more confident in this area after the early days of big bills and subsequent workload repatriations.New strategy to manage inc

35、reasing volumes of data:increasing use of AI and efforts to move network functions into the cloud can change the cloud economics for CSPs.Companies with large volumes of data such as Dropbox and Snapchat are moving to private cloud architecture,and this trend is likely to continue,particularly with

36、the deployment of high-volume generative AI(GenAI).Future supply-side factorsMid-to long-term margin management:hyperscalers are under pressure to demonstrate returns on their huge cloud investments,resulting in cost-cutting measures,including significant layoffs and pressure to improve margins.Mark

37、et maturity and regulatory pressure:AWS and Microsoft have won the largest proportions of the public cloud market,followed by Google,but they arent the only choices.Companies such as Oracle and IBM also provide public cloud services that often cost less but dont offer the depth of ecosystem provided

38、 by the hyperscalers.Regulators are considering market competitiveness.For example,Ofcom in the UK has released a cloud services market study and called on the UK Competition and Markets Authority to investigate the cloud services market.Its hard for challengers to replicate the economies of scale g

39、enerated by the hyperscalers,although it appears there is a theoretical maximum of the market that hyperscalers can take because the public cloud is not suitable for all workloads.Therefore,the future market will include hyperscalers and competitors potentially each with a specific niche competency

40、plus a significant amount of private cloud deployment.Relationship between CSPs and hyperscalers:CSPs find themselves in a position of“co-opetition”with hyperscalers wanting to offer cloud and edge cloud services to their customers,either by going to market jointly with the hyperscalers or in direct

41、 or indirect competition.In some cases hyperscalers procure dark fiber services from CSPs and offer them as private network solutions to enterprise clients.This illustrates significant convergence in both groups competitive strategies,potentially leading to conflicts of interest.Embracing a private

42、cloud approach via a hyperscaler competitor allows CSPs to carve out strategic space for product development.Hybrid cloud capabilitiesTo successfully manage hybrid clouds,CSPs need new skills and capabilities including tools to provide“a single source of truth”around spending and utilization,and to

43、support long-term,strategic decision-making(taking into account scenarios where there are changes to infrastructure costs,pricing and the competitive landscape).They need the ability to orchestrate services across the cloud footprint,allowing for real-time changes while adhering to governance,compli

44、ance and other policies.Finally,they need suitable KPIs that span a broad range of requirements including budget,strategic needs,technical criteria and sustainability.In the next sections well look more closely at these requirements,exploring the advantages and disadvantages of public and private cl

45、oud and CSPs plans for both.7section 2:CSPs current cloud deploymentsinform.tmforum.orgCSPs adoption of public cloud services began nearly a decade ago.Today about half of survey participants have a public cloud-first policy in place instructing IT and technology teams to prioritize procurement of s

46、olutions that are hosted in public clouds.Workloads moving to the public cloud are taking advantage of its scalability,hyperscalers analytics and AI services,and new software-as-a-service(SaaS)capabilities in areas such as the contact center and marketing.However,maturing public cloud usage over the

47、 last couple of years has uncovered two major issues leading operators to move some workloads back to the private cloud or on-premises infrastructure.One is a challenge associated with scaling down workloads,as it has become evident that doing so is not nearly as easy as scaling up,given the complex

48、ity of resizing resources and potential interdependencies between workloads.Operators must closely monitor and adjust resources in complex environments where resources are often used inefficiently.This has resulted in escalating costs for more mature workloads.Another issue is that CSPs lack clarity

49、 around the costs of public cloud.A supermarket shopping cart analogy is often used to describe hyperscalers pricing strategies meaning,the cloud provider provides a block of credits encouraging CSPs to load up their trolley with services.According to a supplier we interviewed,one large CSP using th

50、is approach exceeded its opex budget four times in six months,while another smaller operator in a developing market repatriated its workloads after only two weeks due to cost concerns.However,despite these challenges the prevailing sentiment toward the public cloud remains positive.The CSPs we inter

51、viewed expressed much satisfaction with the cloud services offered by hyperscalers and said they see the cloud-native expertise of hyperscalers as a valuable source of knowledge that their teams can leverage.We asked a survey question about the continuing appeal of hyperscalers developer ecosystems

52、and the use of the public cloud to access new and innovative services such as AI.The graphic below shows that nearly all respondents see these capabilities as important,illustrating the ongoing appeal of working with hyperscalers.Importance of developer ecosystems andAI in deciding whether to use pu

53、blic cloudTM Forum,2024Very importantModerately importantNot very importantAdvanced technologies such as AIAccess to developer ecosystems56%35%9%55%36%9%8section 2:CSPs current cloud deploymentsinform.tmforum.orgTypes of public cloudCSPs are purchasing cloud services using a mix of models.The graphi

54、c above shows the three main flavors of public cloud,as defined by the US National Institute of Standards and Technology(NIST):infrastructure-as-a-service(IaaS),platform-as-a-service(PaaS)and software-as-a-service(SaaS).A more detailed explanation of each type follows.Infrastructure-as-a-service Iaa

55、S focuses on providing infrastructure including scalable solutions for computing,storage and networking(as shown in the pink blocks):Compute provision of the central processing units(CPUs),graphical processing units(GPUs)and internal memory(RAM)required to perform compute tasks;IaaS users request co

56、mpute resources through virtual machines(VMs)or cloud instances;hyperscalers offerings include high-performance solutions for AI and machine learning Storage provision of data storage resources(typically either block,file or object storage types);cloud provider offerings include data lakes and long-

57、term backup Networking logically isolated and customizable virtual network to define and launch IaaS resources;includes monitoring and security Supporting tools IaaS providers offer a range of management tools such as integration,billing and service monitoring.Platform-as-a-servicePaaS provides CSPs

58、 with an environment to develop and deploy their apps as well as the infrastructure to support it,enabling quicker development of new services and platforms(for example,new IoT platforms).Software supports an apps lifecycle from development through testing,deployment,updates and maintenance,and it s

59、upports various languages and frameworks.The range of tools provided by hyperscalers and other cloud providers are such that although IaaS and PaaS remain fundamentally different types of cloud offerings,many IaaS providers now offer services that extend their platforms into PaaS-like products.As pe

60、r the white blocks in the graphic,CSPs use both IaaS and PaaS services for their workloads.Software-as-a-serviceCSPs purchase a range of business applications as SaaS(shown in the gray box).In the white box are applications that can be purchased from hyperscalers and other cloud providers,with new a

61、nd innovative AI and machine learning tools as popular choices in the last few years.However,cloud providers also offer their own or partners products delivering other horizontal capabilities(e.g.,data management)and vertical solutions(e.g.,network management).The next section looks at advantages an

62、d disadvantages of public versus private cloud.Options for purchase of cloud servicesTM Forum,2024Platform-as-a-service(PaaS)Software-as-a-service(SaaS)Infrastructure-as-a-service(IaaS)1Allows CSPs to use cloud-based applications without managing the underlying infrastructure applications made acces

63、sible via web browsers or program interfaceCSPs can deploy their applications on the cloud using supported tools,with control over their applications and hosting environment settings but not the infrastructureCSPs provision computing resources for running software,including OS and apps,with control

64、over these resources but not the underlying infrastructureApplicationsDataRuntimeMiddlewareOperating SystemVirtualizationServersStorageNetworkingManagementApplicationsDataRuntimeMiddlewareOperating SystemVirtualizationServersStorageNetworkingApplicationsDataRuntime2Middleware3Operating System4Virtua

65、lizationServersStorageNetworkingAnalytics-Business intelligence-Real-time video-Data warehousingAI and machine learning toolsDeveloper toolsRange of use case specific tools(e.g.,fraud management)Development frameworkAnalytics or business intelligence Additional services:-Workflow-Directory-Security-

66、SchedulingLift-and-shift migration5Test and developmentStorage,backup and recoveryWeb appsHigh-performance computingBusiness applications:-Contact center-Sales-Marketing-etc.1 Terms as defined by NIST2 Runtime or software that runs the application3 Middleware or software that monitors the applicatio

67、n4 Operating systems on which the application runs5 White boxes highlight the main types of workload typically running on diferent models9section 3:public vs.private cloudinform.tmforum.orgThe graphic below summarizes the primary benefits of public and private clouds to a typical CSP.Detailed analys

68、is of the pros and cons follows.Benefits of public cloud Supports transformation:as noted,collaboration with hyperscalers enables CSPs to take advantage of the cloud providers advanced technology and partner ecosystems.In addition,CSPs regard the hyperscalers services as high quality and believe the

69、y can apply the hyperscalers cloud-first knowledge to their own cloud journeys.Offers scalability and resilience:public cloud services provide scalable solutions,enabling CSPs to expand infrastructure in response to increased demand in both IT and network domains,though most CSPs are just beginning

70、to move network functions to the cloud.Well discuss this more in section 6.Underpins new business models:public cloud expedites software releases,freeing staff to focus on strategic goals like faster service creation and deployment.It also Over the past three to five years,telcos have initiated clou

71、d transformations with goals of improving customer experience and increasing customer lifetime value via cross-selling and upselling of value-added services.CSPs new relationships with hyperscalers provide access to the cloud providers cutting-edge technologies such as machine learning and AI and th

72、eir broad ecosystems of vendors,partners,collaborators and customers,fostering opportunities for co-creation and value exchange.But as operators have moved workloads to the public cloud,theyve discovered that some are better suited to private clouds.Weighing the benefits of public vs.private cloudsT

73、M Forum,2024Short-term benefits Supports CSP transformation Ofers scalability and resilienceAdditional,longer-term benefits Underpins new business models Delivers cost savings Agility and cost savings for smaller CSPsShort-term benefits Allows CSPs to make use ofexisting assets Gives them control ov

74、er datasovereignty and securityAdditional,longer-term benefits Supports low-latency products Cost-efective for certain workloads(e.g.,streaming on the network)Provides long-term competitive benefits(e.g.,preserves vendor neutrality)Public cloudPrivate cloud10section 3:public vs.private cloudinform.t

75、mforum.orgenhances connectivity and collaboration with B2B customers and helps CSPs extend the geographic reach of cloud services.For example,a telco with operating companies in multiple European countries could extend their cloud services to other European countries where they do not have physical

76、presence.Delivers cost savings:while public cloud has often proved expensive when not properly managed,it can deliver cost savings in specific scenarios,such as where local private cloud deployment is not cost effective because the telco does not have a point of presence or when network capabilities

77、 are only required for a short period of time.Agility and costs savings for smaller CSPs:smaller or greenfield CSPs benefit from streamlined access to infrastructure and operational services,speeding up time to market and helping them navigate supply chain challenges and reduce capital expenditures.

78、Reliance Jio,for example,is working with Google Cloud.This includes moving Jios core retail businesses to the hyperscalers infrastructure in order to use AI and machine learning,e-commerce,and demand forecasting tools.Google delivers fully automated lifecycle management of Jios 5G network and servic

79、es,and the companies are partnering to create 5G edge computing solutions.Benefits of private cloudAllows CSPs to make use of existing assets:private cloud allows CSPs to improve the profitability of their existing data center infrastructure by adding new workloads into existing facilities and creat

80、ing additional economies of scale.Ensures data sovereignty and security:private cloud addresses the privacy,security and regulatory requirements of certain workloads such as those that require high security or rely on use of personally identifiable data.Supports low latency products:private cloud su

81、pports certain types of network functions which are not suited to the public cloud due to requirements for specific SLAs around latency.These might include control applications in manufacturing or telepresence services,for example.Cost-effective for specific workloads:for certain high-volume,predict

82、able workloads like streaming traffic on the network,private clouds can offer better value for money because they would incur high data transit charges when moved out of the public cloud,which is important as data volumes increase with the adoption of 5G and AI and machine learning applications.Prov

83、ides long-term competitive benefits:Preserves vendor neutrality for CSPs building a hybrid cloud that includes private cloud capabilities for end-to-end orchestration offers CSPs a crucial control point to avoid vendor lock-in by allowing a choice of vendors and the ability to move workloads between

84、 them easily.Maintains separation between CSPs services and those of competing hyperscalers deploying private clouds with hyperscalers or other platform providers allows CSPs to strategically develop products that are not dependent on the hyperscalers roadmaps.Offers opportunities in cloud service p

85、rovisioning CSPs can explore specialized offerings such as a“compliance cloud”with their Where to run workloads?TM Forum,2024Public cloud IT workloads/SaaS applications(e.g.marketing and customer-facing apps and cloud-native apps like Salesforce)Data,analytics and AI benefit from the proximity of in

86、novative tools and developer ecosystemsApplication development and testing tasks benefit from the ability to scaleB2B workloads(e.g.where enterprise prefers public cloud or resold hyperscalers oferings)OSS/BSS workloadsNetwork workloads such as disaster response and backup,transitory and test and in

87、novation environments.Private cloudCSP or enterprise workloadsrequired to remain in a localprivate cloud for data sovereigntyand regulatory reasons(e.g.,GDPR or vertical regulations)Workloads for high-throughputservices such as streamingWorkloads requiring low latency,where public cloud cant guarant

88、ee QoS!On-premisesWorkloads with high-levelsecurity requirementsLegacy workloads with no business model for moving to the cloud or very old applications that would be hard to enable in a cloud11section 3:public vs.private cloudinform.tmforum.orgprivate cloud capacity.A compliance cloud is designed a

89、nd configured to meet specific regulatory compliance requirements such as GDPR(General Data Protection Regulation)in Europe or HIPAA(Health Insurance Portability and Accountability Act)in the US.It also provides security,data governance and auditing/reporting capabilities.Allocating workloadsGiven t

90、he strengths and weaknesses of public and private clouds,it is possible to pinpoint the types of workloads that are best suited for public versus private or on-premises implementations.The graphic on the previous page shows the types of workloads that are generally best for each scenario.But a long-

91、term balance of all three will be crucial for CSPs to optimize cloud costs.Public cloud OSS/BSSIn our survey,we asked specifically about moving OSS/BSS workloads to the public cloud.The graphic above shows the percentage of respondents whose companies have already moved or plan to move functions to

92、the cloud.CSPs need to make decisions about public versus private cloud OSS/BSS deployments because the applications have a mix of requirements.Some like mediation need high-intensity and guaranteed performance,for example,while others like digital channels benefit more from scalability.Indeed,surve

93、y respondents are most enthusiastic about moving digital services such as chatbots and IoT capabilities to the public cloud.A more mixed picture emerges,however,for workloads that might have privacy implications such as analytics and AI.CSPs desire to innovate and create developer ecosystems should

94、mean they favor moving such applications into the public cloud,but fewer than half of respondents are doing so.Public or private for AI?Deciding where to deploy AI is not a clear-cut choice for CSPs.Public cloud provides substantial computational resources,energy and specialized hardware required fo

95、r processing complex algorithms and large amounts of data.However,hyperscalers must address non-technical issues related to cost,economics,sovereignty and data privacy.And transit costs may become prohibitive if a CSP solely utilizes a public cloud to process large amounts of data.We spoke to one op

96、erator that has a strict“no public”policy for its GenAI trials due to uncertainty around data privacy and other AI regulations.In some circumstances,a private cloud could be cost efficient,but deploying one requires significant initial capital expenditure not only on data center capacity,but also hi

97、ghly specialized hardware.This dilemma is then compounded by the lack of return on investment seen from GenAI trials so far.TM Forum members are working together in collaboration projects and Catalyst proofs of concept to solve common AI challenges for the telecoms industry.Central to that is a new

98、Innovation Hub,hosted by Jio in Mumbai,which opened in December 2023.Founding members are Accenture,Deutsche Telekom,Google Cloud,Orange,Reliance Jio,Telenor and Vodafone Group.Building a hybrid cloud that includes private cloud capabilities for end-to-end orchestration offers CSPs a crucial control

99、 point to avoid vendor lock-in.OSS/BSS moving to the public cloudTM Forum,2024Digital channels/chatbotsNew lines of business(e.g.IoT)Enterprise resource planningAnalytics/AIProduct catalogCRMBillingOrder managementOSSMediationRevenue assurance52%50%46%43%43%38%36%36%29%23%22%12section 3:public vs.pr

100、ivate cloudinform.tmforum.orgAmong the early projects will be the development of TMF Guru,a GenAI/LLM conversational search capability to simplify and improve how to find relevant material digitally.The project members are setting out to create an architecture blueprint that outlines how GenAI techn

101、ology can be seamlessly integrated into the telco landscape,addressing challenges related to security,privacy,accuracy,performance and scalability.Learn more about GenAI in telecoms from our in-depth report:REPORTTM Forum|December 2023operators take their first stepsAuthor:Mark Newman,Chief Analyst

102、Editor:Ian Kemp,Managing Editor Sponsored by:generativeAI:We cover cloud BSS in depth in the report linked below.The next section looks at the economics of moving from a capital expenditure(capex)model to an operating expenditure(opex)model for cloud deployment.Learn more about cloud BSS:REPORTTM Fo

103、rum|December 2023Author:Ed Finegold,Contributing Analyst Editor:Dawn Bushaus,Contributing Editor Sponsored by:BSSfor B2Boperators diverge on the path to cloudSurvey respondents are most enthusiastic about moving digital services such as chatbots and IoT capabilities to the public cloud.13section 4:t

104、he economics of capex vs.opexinform.tmforum.orgIn existing data centers CSPs costs are split,with about 50%to 60%going toward capitalized infrastructure(purchase of servers and storage,for example)and the rest stemming from opex costs to power,cool and staff data centers.In the traditional data cent

105、er procurement model,CSPs can take advantage of tax deductions on depreciation.Telcos and investors developed this capex-focused model long ago,but cloud is forcing them to re-evaluate it.Transitioning towards an opex model offers several advantages.Departing from on-premises infrastructure reduces

106、capital outlay,and public cloud solutions deliver increased economies of scale and better scalability.In addition,while traditional capex investments presuppose a certain level of predictability in business decisions,the cloud model empowers companies to adapt and pivot in response to evolving busin

107、ess needs and market dynamics.Moreover,the capex model incurs significant expenses,including for complex setup and maintenance and protracted cycles of technology acquisition and refresh.However,the adoption of opex models presents internal challenges for CSPs.The increased accessibility of cloud re

108、sources often results in decentralized decision-making and enables internally unregulated cloud deployments potentially leading to an uncontrolled escalation of cloud costs.It is difficult for operators to effectively attribute these cloud costs to their respective owners.It is also hard to enforce

109、a policy of migrating only those workloads capable of deriving substantial benefits from the cloud environment.Despite these difficulties,telco CFOs can make an argument to investors that the move to an opex model offers agility that will lead to new revenue,even if it is at the expense of short-ter

110、m margins.The use of public cloud is shifting the traditional data center model,which relies on capex,to an opex model.An opex model offers important advantages when it comes to agility,but it also has drawbacks such as unpredictable costs.Telco CFOs can make an argument to investors that the move t

111、o an opex model offers agility that will lead to new revenue.14section 4:the economics of capex vs.opexinform.tmforum.orgManaging opex costsFinOps,short for financial operations,is a discipline that focuses on optimizing the financial management of cloud resources by aligning spending with business

112、goals and improving cost efficiency in cloud computing environments.It involves monitoring,analyzing and controlling cloud-related expenses to ensure cost-effectiveness and resource optimization.We asked survey respondents about their confidence in being able to manage the costs of public cloud usin

113、g approaches like FinOps,negotiation of volume discounts,raw cost monitoring and increasing their cloud skills.The graphic above shows that respondents are confident in their ability to increase cloud skill sets to handle the placement and day-to-day management of public cloud workloads most likely

114、due to increased spending on training and recruitment but they are less sure about using FinOps.Currently,CSPs must commit substantial volumes to secure the lowest price points from hyperscalers.As CSPs use of data analytics and AI,software-defined networking,and cloud-native OSS/BSS continue to gro

115、w,they will migrate more workloads to the public cloud.This could lead to potentially higher discounts from multiple hyperscalers in the coming years,improving cost efficiency.But this trend is only one part of the picture when a CSP calculates the relative cost of public cloud.Anecdotal evidence fr

116、om the executives we interviewed points to more important short-term issues particularly,that billing across the contract term does not reflect the original prices in the contract due to the number of product changes and upgrades over the contract period.Hyperscalers offer various pricing models for

117、 public cloud services,the most flexible of which are on-demand instances where users are billed by the hour or second for the compute capacity they use.For more predictable workloads,reserved instances offer cost savings and more predictable billing for a term commitment,with users receiving a redu

118、ced hourly rate for their chosen instance type.Spot instances allow users to bid for spare cloud capacity,resulting in potentially significant savings.However,the hyperscaler reserves the right to terminate these instances if the capacity is needed elsewhere,making this option more suitable for comp

119、ute-intensive tasks that tolerate downtime and interruption such as batch jobs.Other“specialist”products are also available including dedicated hosts,where compliance or licensing restrictions require dedicated physical servers,object storage(highly scalable storage),block storage(offering low-level

120、 access and management)and storage for workloads that are not accessed often.Lastly,another significant cost item for CSPs is the charge for data transfer in and out of public cloud data centers,with pricing dependent on the volume of data and the specific region in which the transfer occurs.How do

121、hyperscalers price public cloud services?CSPs confidence in managing public cloud costsTM Forum,2024Very confidentModerately confidentNot confidentRaw cost monitoringUsing FinOpsIncreasing cloud skillsNegotiation of volume discounts22%55%24%27%46%27%42%46%12%28%49%23%15section 4:the economics of cap

122、ex vs.opexinform.tmforum.orgMore or less?To explore CSPs attitudes about public cloud costs further,we conducted an informal poll on LinkedIn.We asked whether a telco workload that is best suited to public cloud(meaning,it is characterized by variable demand)will cost less or more in a public cloud

123、than a private one or on-premises.The results are shown in the pie chart opposite.The informal poll uncovered some interesting trends.Two-thirds of respondents found themselves spending more on cloud than they would expect to pay or had paid in the past for private cloud/on-premises solutions.Some f

124、ound this to be acceptable,presumably because they thought they received other non-financial benefits,but more than 40%were unable to manage their costs satisfactorily.Conversely,35%of respondents were managing costs to their satisfaction,with about one in five finding that their choice of workloads

125、 on the public cloud allowed them to always save on costs.The diversity of responses shows the complexity of cloud economics:While some operators are less cost sensitive,most perceive the additional cost as a problem to be solved.As CSPs build their FinOps capabilities,they will likely be able to ba

126、lance their spending better and manage contracts at a more granular level,which could bring costs down.But the cost of public cloud services will likely rise over time unless the economics of building extra capacity to handle increased volumes of data created by GenAI and other new capabilities are

127、such that they allow hyperscalers to reduce their own costs.The next section covers the costs associated with private clouds.CSPs expectations about public cloud costsTM Forum 202421%43%21%14%They will always cost less on publiccloud than private/on premisesThey will sometimes cost less onpublic clo

128、udThey will cost more,but thats OK They will cost more,and thatsa problem 16section 5:the cost of deploying private cloudsinform.tmforum.orgThese considerations need to be balanced against a true reflection of public clouds ability to scale larger data centers,high-performance hardware and so on acc

129、ess to hyperscalers mature management tools,and the additional cost of purchasing services that include hyperscalers retail margins.In addition,operators must consider factors beyond economics that force them to use private clouds for some workloads for example,regulation of where network functions

130、can be deployed.CSPs must implement cost containment strategies for workloads that are unsuitable for the public cloud or would not strategically benefit the company if deployed in a public cloud.Such strategies include retirement of legacy infrastructure and facilities,use of cloud platforms mature

131、 automation and management tools,retaining staff to manage the increasing complexity of cloud deployments,and use of managed services to bring in additional expertise as needed.The graphic on the next page shows the major costs associated with deploying a private cloud data center.The current trends

132、 that are accelerating costs are shown in pink.Major actions CSPs can take to reduce costs are depicted in blue.A more detailed explanation of the graphic follows.Infrastructure:purchasing servers,storage and compute infrastructure is highly capital intensive.Other infrastructure costs include firew

133、alls,load balancers and racking.Digital infrastructure information provider Dgtl Infra calculates that it generally costs between$600 and$1,100 per gross square foot,or$7 million to$12 million per megawatt of commissioned IT load,to build a data center.The amount operators spend depends on the sophi

134、stication of capabilities required with an ongoing need for improvements in silicon to support requirements in areas such as the rollout of 5G.Increased utilization reduces capital costs,with legacy on-premises storage often being under 50%utilized to allow for peak time spikes.Greater automation wi

135、ll help CSPs increase efficiency and pack network functions more efficiently into hardware.When considering a long-term strategy of deploying private clouds,CSPs must calculate the total cost of ownership and balance it against the benefits of a public cloud approach.The main costs associated with p

136、rivate clouds are the ongoing capex needed to build and upgrade data centers and the continuing opex needed to operate and staff them.In addition,telcos need to figure out how to optimize the capital investments theyve already made in private clouds.17section 5:the cost of deploying private cloudsin

137、form.tmforum.orgPower and cooling:power costs have increased more than any other data center expenses recently due to high gas prices,the transition to renewable energy and the war in Ukraine.Energy efficiency measures,including greater automation to manage power consumption,can help to reduce bills

138、.Personnel and maintenance:operators are being forced to spend more to retain staff,but hyperscalers typically offer higher salaries for the same jobs making it difficult for CSPs to keep staff and acquire talent.On the other hand,telcos can turn to hyperscalers and other cloud platform providers fo

139、r private cloud deployments and managed services to fill skills gaps.Any improvements in infrastructure usage have a positive knock-on effect,increasing efficiency in the cost of both onsite personnel and vendor maintenance charges.Efficiency can also be increased by using more modern management sys

140、tems to reduce or automate manual tasks such as configuration,provisioning and fault management.Facilities:the cost of building facilities is increasing because of rising construction and labor costs and the increasing cost of capital due to inflation.This has left smaller data center operators stru

141、ggling to remain competitive.In addition to the cost of buildings,computer room infrastructure has been subject to a shortage of equipment,and a tightening in the lithium supply has increased the cost of batteries.Also,new regulations calling for high-efficiency energy design require spending on eng

142、ineering,structural changes and upgraded cooling systems.Like infrastructure costs,facility costs can be somewhat managed by improving data center usage.Software:CSPs are also spending to implement new cloud platforms,automation and management tools.There may be some decrease from retiring legacy so

143、ftware and efficiencies gained by deploying more modern software packages.Learn more about telcos energy consumption:REPORTAuthor:Joanne Taafe,Editor in Chief,InformEditor:Ian Kemp,Managing Editor TM Forum|December 2023the sustainable telco:navigating the maze of scope 3 emissionsThe next section lo

144、oks at the journey to public cloud,featuring one unnamed CSPs experience.How to reduce the cost of deploying workloads in private cloudsTM Forum,2024Notes:The proportions of each box were calculated by comparing information gathered from multiple internet sources.In real-life scenarios,the proportio

145、ns difer by size of data center,its location,the energy efciency measures that are in place and the specific services ofered.The rising costs were also compiled from multiple internet sources.The cost reductions were taken from publicly available studies completed by various analyst firms and publis

146、hed by diferent vendors.Each put forward a case for reducing costs via diferent activities such as moving from on-premises to public cloud or automation of the data center and were not directly comparable.But the diagram shows general conclusions about opportunities for CSPs to reduce costs.Addition

147、alvirtualization automation Management tools Software modernizationFacilitiesSoftware Inflation Efciencies in infrastructure usageMaintenance Cost of capital Construction and labor costs Supply chain shortages Improved utilization via new automationPersonnel Salary increases due to market pressure I

148、mproved management systems Automation of tasks such as configuration Efciencies in infrastructure usageInfrastructure Supply chain issues Shortages in areas such as processors Increasing demand for high-spec silicon Improved utilization via new automationInfrastructure Higher gas and electricity pri

149、ces due to instability in global energy supply Move to renewable energy Power management initiatives 18section 6:navigating the road to public cloudinform.tmforum.orgThe graphic below shows the journey of an unnamed CSP whose experience is typical.The blue line describes the corporate target for clo

150、ud deployment,starting with a goal of 100%of potentially suitable workloads going into the public cloud.The pink line demonstrates the CSPs actual cloud deployments.On day one of the CSPs cloud migration journey the CEO expressed a strong ambition to transition the organization to a 100%cloud-native

151、 environment.However,as the company began considering which applications to move first it rapidly became evident that achieving 100%cloud-native operations was not feasible,with certain legacy workloads better kept on-premises.The pink line shows that cloud deployments actually ran at approximately

152、10%to 15%,with deployments particularly in areas such as marketing services,which were naturally suited to cloud-based solutions.Executing a corporate public cloud-first strategy is not an easy journey for CSPs.Indeed,it is beset with commercial,technical and practical challenges.Deployment of workl

153、oads into the public cloud:target vs.actual deploymentTM Forum 2024(source:with thanks to Anthony Behan,Cloudera)Day 16 months12 months18 months100%90%80%70%60%50%40%30%20%10%0%100%90%80%70%60%50%40%30%20%10%0%Target cloud deploymentCloud deployment achieved19section 6:navigating the road to public

154、cloudinform.tmforum.orgSix months later senior executives expectation levels had dipped to 70%due to the increased appreciation of challenges around scale,cost,security and privacy.The company then began identifying services that could sensibly benefit from public cloud deployment,particularly those

155、 with no external dependencies and low data volumes,alongside any newly purchased cloud-native services.However,the pink line demonstrates the reality that comprehensive migration is a long-term endeavor and not all workloads can be seamlessly transitioned despite best efforts.As the CSPs journey re

156、ached the one-year mark the organization continued to appreciate the infrastructure advantages of the public cloud and cost savings compared to traditional data centers.Yet the challenges of migrating larger workloads became more evident,and existing contractual commitments to on-premises solutions

157、further limited the scope of what the operator could move to the cloud.Additionally,the company discovered that around 20%of the workloads initially placed into the public cloud were not suitable,as the organization(plus its customers and partners)identified security and privacy concerns which requi

158、red the workload to be repatriated for example,workloads that contained personally Identifiable Information.This led to movement of workloads back and forth between cloud and on-premises environments,as shown by the pink line.After about a year and a half the CSPs experience of operating in the publ

159、ic cloud deepened.While initial public cloud onboarding discounts were beneficial,it became clear that it would cost less to keep some workloads on-premises,especially those with limited scalability requirements.In general,a lack of close financial management of cloud resources also contributed to h

160、igher costs.The operations team began to feel less pressured by senior management to default to the public cloud,and there was a noticeable trend of workloads being repatriated to existing data center infrastructure to mitigate public cloud-related expenses.In the final stage of maturation cloud dep

161、loyment expectations versus actual deployment began to converge as the organization built good FinOps capabilities including cost tracking,optimization and governance and orchestration of capacity while setting the expectations of senior management.Weighing the costsThe CSPs public cloud journey hig

162、hlights the complexity of balancing costs,performance,security and privacy.Indeed,as discussed in the previous section,the cost of public cloud services will impact every CSPs decisions about using it in the future.Most operators will take one of the following approaches:Public by choice some operat

163、ors want to shut as many of their own data centers as possible to achieve maximum economies of scale in the public cloud.Hyperscaler lock-in is not necessarily negative as it brings the opportunity to drive the cloud providers roadmap towards delivering new functionality and features with the promis

164、e of a closer relationship.Cloud infrastructure ownership is seen as a commodity and not a major source of competitive differentiation,at least in the IT and macro-network domains.Public by necessity smaller CSPs that dont have the ability to create economies of scale in their data centers but need

165、to go to market quickly see the hyperscalers as long-term partners.But issues may arise with public cloud prices over time if they are locked into a single cloud platform with a high cost to move workloads away.Public where useful and private where necessary(the partner approach)most CSPs are moving

166、 towards a hybrid approach where a mix of public and private cloud strengths are exploited,and workloads are managed at a top level to balance cost versus performance.Mindsets will be to use public cloud where useful and private cloud where necessary,where regulatory or other restrictions rule out t

167、he public cloud or make it much more costly,The cost of public cloud services will impact every CSPs decisions about using it in the future.20section 6:navigating the road to public cloudinform.tmforum.orgfor example.In the partnership approach,cloud software infrastructure is purchased from a hyper

168、scaler or other cloud platform vendor and deployed either on-premises or in the cloud while hardware infrastructure is procured directly from the original equipment manufacturer.Public where useful and private where necessary(the DIY approach)this requires deep pockets,a commitment to building cloud

169、 engineering capabilities and a desire to build competitive differentiation with cloud services.Typically,the CSP will purchase cloud software infrastructure from several providers and create a solution specific to their needs.Cost optimization requires a good deal of sophistication across all areas

170、 of data center and cloud management,making this a strategy for only the most committed CSPs.In the next section,we look at the drivers for putting network workloads in public clouds and why CSPs still prefer private clouds.Most CSPs are moving towards a hybrid cloud approach to balance cost versus

171、performance.21section 7:CSPs prefer private clouds for network functionsinform.tmforum.orgOrange Innovations Demartis explains that“there are tough regulations in France.We are not currently allowed to move network functions outside of our network in France.”In theory,however,the financial advantage

172、s of placing network functions in a public or private cloud present a compelling business case.A recent Capgemini report cites potential annual opex benefits ranging from$200 million to$400 million for a typical CSP,along with revenue uplift from new service opportunities in the range of$100 million

173、 to$200 million.Even so,the transition to the cloud for network functions has proven expensive,with CSPs first deploying virtual machines(VMs)and static applications rather than containerized applications,which reduces efficiency and increases costs.In addition,CSPs that embarked on this virtualizat

174、ion journey found additional challenges with their private cloud deployments due to vendor immaturity and an overall lack of cost effectiveness when deploying their own assets.However,over time they have successfully built their skill sets(addressing one of the primary hurdles in private cloud devel

175、opment)and now typically have a significant amount of legacy deployment,which will be one factor affecting their longer-term decisions to move to public cloud.Early adopters of network in the public cloud,such as AT&T and Dish Network,have detailed that they encountered challenges related to network

176、 latency and performance,along with a shortage of telco-specific skills within hyperscaler companies.We conducted another informal poll on LinkedIn to gauge industry sentiment about the future of network functions in the public cloud,asking simply whether people believe that CSPs network functions w

177、ill go into the public cloud in the future.Nearly two thirds of respondents expressed skepticism about a public cloud future for networks.Backing up this attitude,the Capgemini study suggests that CSPs overwhelmingly favor private cloud for the core,RAN and edge(see graphic).The CSP executives we in

178、terviewed agree that certain workloads,particularly network workloads,are better suited to private clouds.Often this is driven by regulatory requirements.Cloud architecture choices for the networkTM Forum,2024(source:Capgemini Research Institute)82%Public cloudPrivate cloud ofered by hyperscalers an

179、d large cloud solution providers with end-to-end capabilitiesPrivate cloud ofered by network equipment providersPrivate cloud ofered by our own organizationCoreRAN/open RANOSS andorchestrationMulti-accessedge16%2%59%44%15%22%52%10%45%11%35%3%3%22section 7:CSPs prefer private clouds for network funct

180、ionsinform.tmforum.orgChoosing a pathCSPs are likely to fall into one of two camps when it comes to deploying network functions in the public cloud:those that are willing to primarily embrace the public cloud and those that will choose a hybrid strategy.Survey respondents believe that once pioneerin

181、g companies demonstrate that networks can operate in public clouds,other CSPs will follow suit.These opinions were partly based on the successful deployment of public cloud networks in industries such as manufacturing although the scale of a public telco network versus a private network in a factory

182、 is markedly different.More compelling is the argument that many CSPs will want to avoid deploying the necessary private cloud capacity to run a network because of issues around developing specialist talent,procurement of infrastructure,and the need to manage the efficiency and performance of these

183、data centers.The success of CSPs that choose to deploy their networks primarily in public clouds will depend on the desire of hyperscalers to invest in suitable infrastructure,particularly improved silicon,and their continued rollout of geographic capacity in areas where CSPs do not have operations.

184、In the other camp,a hybrid cloud approach to the network will see workloads spread among public,private and on-premises deployments.The graphic above demonstrates the potential top-level decision-making behind the deployment of different types of network workloads.Operators are likely to deploy data

185、-light network functions that are less critical to customer experience(such as those in the control plane)in the public cloud,along with some other functions that need scalability and therefore can be deployed more economically in public clouds.Beyond architectureFactors besides technical suitabilit

186、y impact CSPs decisions about where to place network functions.Other drivers for private cloud networks include:Security,privacy and regulation regulations often prohibit operators from migrating network functions to the public cloud as is the case in France,for example.Data-heavy network functionsP

187、acket gateway and other user-plane functions sensitive to latency and managing high volumes of dataLatency-sensitive RAN functionsNetwork connection may be lost if latency is increased beyond a certain limitLatency-sensitive servicesVoice or video services where customer experience issues may occur

188、because of latency High-throughput servicesStable volumes such as fiber-to-the-X user-plane functions are better supported by on-premises hardwareGeographically dispersed workloadsRAN trafc in geographic locations not served by hyperscalers public cloudsEdge and private cloudCompetitive diferentiati

189、on by deploying hybrid cloud rather than public-cloud onlyLikely division of network functions across public and private cloudsTM Forum,2024Suited to public cloudSuited to private cloudControl-plane network functionsAccess and mobility management function(AMF)and policy management(these are less lat

190、ency sensitive than user-plane functions,and they manage limited volumes of data)Temporary networksScalable infrastructure to create temporary networks(e.g.,during a sporting event)Disaster recovery and backupScalable infrastructure to provide as-needed network capacity Test and innovationScalable i

191、nfrastructure to allow quick network deployment during service developmentEdge and private cloudDeployments where customers require use of public cloud or it is cost beneficial for architectural reasonsOnce pioneering companies demonstrate that networks can operate in public clouds,other CSPs are li

192、kely to follow suit.23section 7:CSPs prefer private clouds for network functionsinform.tmforum.org Cost as previously discussed,the public cloud is less suitable for network functions with high throughput.This directly impacts public cloud costs,where egress costs would be high.Strategic benefits th

193、ere is much discussion about private cloud as a point of differentiation,allowing CSPs to provide cloud services with an emphasis on sovereignty,trust and(edge)geographical coverage not available in the public cloud.Differentiation of edge and private network services edge and private network opport

194、unities have been slow to take off,but CSPs have an opportunity to build their own capabilities without being dependent on a hyperscalers roadmap and technical decisioning.Indeed,it is becoming increasingly important for telcos to demonstrate their relevance in a cloud-native world.“If telecom is no

195、 longer the industry that knows how to best run highly distributed,highly performant infrastructure(edge-cloud),and telecom is no longer the industry that is delivering the best end user experiences above connectivity on such infrastructure,then what IS telecom?”Rakutens CMO Geoff Hollingworth wrote

196、 recently on LinkedIn.In the final section of the report,we offer some recommendations for hybrid cloud strategies that can help CSPs optimize costs.TM Forum members are collaborating on best practices and standards to help CSPs with cloud migration of OSS/BSS applications.The TM Forum Open Digital

197、Architecture,part of the Open Digital Framework(see page 30),is a component-based approach that enables operators to evolve to fully automated,cloud-native operations that rely on analytics and AI to deliver zero-touch services.The ODA defines standardized,interoperable software components organized

198、 into loosely coupled domains.These components expose business services through Open APIs,which are built on a common data model.Importantly,the ODA provides machine-readable assets and software code,including a reference implementation and test environment.The graphic opposite illustrates an ODA co

199、mponent,which is an independently deployable piece of software,typically built out of one or more microservices.Components have an“envelope”that provides metadata to describe its core function and specify which Open APIs it exposes or depends upon.To learn more about the ODA and Open APIs,please con

200、tact Ian Turkington.Read this e-book to learn more about the ODA:Helping CSPs evolve to a cloud-native architectureCore functionNotification/ReportingSecurityManagement/OperationEnvironment Dependence/RequirementsTM Forum,2021RequiredOpen APIsProvidedOpen APIsAdopt FinOpsFinOps tools used to manage

201、a CSPs hybrid cloud are designed to balance the performance and requirements of users at the best cost(see this list from Gartner for some examples of cloud management tools).The rate of change will be fast,and the ability to view cloud operations through a single pane of glass is necessary.Adopting

202、 a FinOps approach will provide a single source of truth to answer key questions:What is our spending across all clouds?Which cloud services are individuals within the organization using?It will also provide planning and forecasting to manage discounts across all hyperscalers and provide insight bac

203、k to operating units and developers about their own usage and costs.In addition,long-term planning models will be needed to provide input into strategic decisions around cloud usage.For example,will AI deployments be best placed in public or private clouds?Modelling will require projects looking at

204、the cost and availability of infrastructure and considering likely changes in hyperscalers public cloud pricing,products and overall strategies.This type of modelling will require finance teams and cloud architects to work together closely,and it will need to be revised periodically,considering past

205、 projections and actual outcomes.Telcos IT and network teams will need to manage all kinds of applications and workloads across public and private clouds and other on-premises locations with a goal of striking a balance between cost efficiency,performance,security and regulatory compliance.They must

206、 do so in a way that meets the financial and strategic goals of the telco C-suite and supports the diverse and changing needs of customers and developers.The practice of effectively managing a complex hybrid cloud and data ecosystem will emerge as a critical competency for CSPs,irrespective of their

207、 strategic objectives regarding cloud utilization.They should consider their governance and control over these hybrid environments as a key strategic asset.Here are four key steps to take.section 8:make it happen hybrid cloud strategies to optimize costs24inform.tmforum.orgFocus on orchestrationThe

208、hybrid cloud will require management tooling to provide governance,compliance,lifecycle management,brokering and automation across the various clouds.Such tools need to be able to cope with a multi-vendor environment,performing orchestration with extreme efficiency to ensure that costs are managed w

209、hile performance requirements are always met.Today,CSPs typically use different tools provided by individual hyperscalers.However,deployment of hybrid clouds will require a management system that is agnostic and provides unified observability.Such a system increasingly will need to automate critical

210、 actions in real time without human oversight.25section 8:make it happen hybrid cloud strategies to optimize costsinform.tmforum.org Set KPIsThe hybrid cloud will need to be managed against a range of KPIs,dependent on an operators corporate and cloud strategy.Areas of measurement include:Cost is th

211、e deployment remaining within budget and maintaining good efficiency levels(for example,monitoring of compute and storage utilization or how well the infrastructure responds to changes in workload)?Revenue growth and innovation is the deployment allowing the speedy development of new services and su

212、pporting customers needs?Customer satisfaction are performance and other customer-facing metrics like net promoter score(NPS)being met?Sustainability are metrics for carbon reduction being met?Compliance,data privacy,security and sovereignty is cloud orchestration meeting defined security policies?F

213、uture investment plans is the hybrid cloud and its underlying usage of cloud locations supporting long-term decisions on where the CSP would like to invest?Competitive strategy will the hybrid cloud support the operators competitive strategy and allow changes to this strategy as needed?Build strong

214、teamsThe hybrid cloud will require additional training on cloud-native software skills as the cloud deployment teams within CSP organizations expand.New tools across the cloud footprint will also be needed to support the team in decision-making,alongside work on the development of new automations th

215、at better integrate the various cloud platforms.When looking to improve the economics of the cloud,training for all users should be extended to include cost awareness and improve their ability to self-manage costs and usage.The cloud team should also be encouraged to work closely with the variety of

216、 users that will require support,considering additional development of the teams hard and soft skills if necessary.Cloud strategies must balance costs,performance,security and regulatory compliance.27|Full stack FinOps optimises costs for private and public clouds and delivers radical operational im

217、provements30|TM Forum Open Digital Framework31|TM Forum research reports32|Meet the Research&Media team26inform.tmforum.orgadditional features&resourcesFull stack FinOps optimises costs for private and public clouds and delivers radical operational improvements27inform.tmforum.orgSPONSORED FEATURETh

218、e ChallengeCommunication Service Providers(CSPs)are competing in a market thats being transformed by rapid technology and social change.They have a set of common challenges:Continued investment is needed to deliver data and video services.Yet customers are not prepared to pay more for these services

219、.Meanwhile,as CSPs move workloads to public cloud,using the scalability and elasticity inherently provided,they find it harder to track costs-which can grow unexpectedly with demand.With rising expectations,higher costs and stagnant revenues,cost cutting alone is not enough:a new approach with capab

220、ilities for sustainable,operational growth is essential.Current FinOps practices are insufficientAs Communication Service Providers(CSPs)face growing complexity in dealing with multiple public and private cloud environments,it can be challenging to shift mindsets and capability towards new cloud tec

221、hnology approaches such as cloud-native and modern working practices(e.g.DevSecOps).Consuming cloud as if it is a data centre fails to drive cloud benefits and cost efficiencies that are otherwise achievable.Traditional private cloud implementations when CSPs transition to on-demand public cloud con

222、sumption are moved from CAPEX to OPEX.Previously,this has been a barrier to large scale public cloud adoption,despite the operational benefits introduced:elastic autoscaling,cloud-native supporting infrastructure and rapid provisioning via APIs.Such a barrier may still exist amongst some CSPs.Howeve

223、r,the asset value contribution of public cloud infrastructure is now more widely accepted.FinOps is a financial and cost management approach and framework,originally designed to help organisations understand and manage costs in the public cloud.To fully govern and optimise costs,CSPs must approach c

224、loud cost management holisticallyThats why Full Stack FinOps is so important,as it covers all cloud hosting environments:Private cloud on-premise Public cloud Hybrid combining both private an d public clouds.CSPs will continue to manage hybrid cloud environments whilst balancing costs for investment

225、(CAPEX)and operations(OPEX).We also expect them to keep developing and committing to new cloud-native services,including private cloud-hosted,cloud-native network functions(CNFs),Edge compute and Open RAN.In this complex and increasingly dynamic technology landscape,CSPs need complete visibility and

226、 the ability to manage complexity through integrated,multi-disciplinary FinOps teams and a near real-time single Source of Truth(SoT).Thats what Full Stack FinOps is designed to deliver.FinOps is a broad framework that helps CSPs to understand,manage,and forecast cost at an application or service le

227、vel with capabilities delivered from 6 key FinOps domains with particular outcomes.FinOps builds a consistent and regularly maintained view of how applications consume cloud infrastructure,drives collaboration and culture change across the organisation and can deliver a business level cost and profi

228、tability view of applications and services through unit economics.A key aspect of FinOps is to mature capabilities in knowledge,metrics(KPIs),adoption,automation and processes using continuous improvement,driving ever increasing operational efficiency.There is a need to change direction from separat

229、e public cloud and on-premises perspectives,delivering a common approach across hybrid cloud,the Full Stack,where value comes not from individual solutions but a holistic approach across all environments regardless the cloud platform or environment.Taking a FinOps perspective,many environments hoste

230、d in private cloud have a lack of comparative visibility in comparison to public cloud.Most public cloud environments have the ability to both consume a cost-based view easily and to allocate it to particular workloads and applications dynamically.Full Stack FinOps helps to address this,delivering a

231、 holistic view of cost and efficiency across all platforms.Taking FinOps to the Next LevelLeading telecom players are working towards cloud-native status for applications and processes by 2030,while intelligent networks and the use of connectivity to deliver innovative B2B services across multiple i

232、ndustries will transform their growth potential.CSPs end customers are not purchasing underlying capabilities such as network functions and workloads,in fact they do not have visibility of these.It is important for CSPs to take a service-based view and use an approach such as Full Stack FinOps to dr

233、ive a more consistent and agnostic view of environments and workload placement,a disassociated viewpoint from the service to deliver efficiency-based outcomes.This proves to be a valuable insight for workload hosting selection and bundling multiple workloads as a service to drive even greater effici

234、ency.Adopting this approach also allows data driven decision-making removed from services,considering a range of factors such as:Cost Where is the lowest cost hosting solution?This perspective may change as public cloud and data centre costs vary.Value Do I get more value by hosting in a particular

235、environment?e.g.where do I need rapid scalability or have specific technical requirements?Sustainability What are the comparative sustainability outcomes from hosting a workload in different environments?As ESG strategies and net-zero targets progress there will be a cost for delivering sustainable

236、outcomes.It is important to start building an operational framework that delivers sustainability versus cost viewpoints to ensure decisions made are informed and deliver the most benefit.28inform.tmforum.orgSPONSORED FEATURERadical improvements through xOpsEnabling this approach requires more than a

237、 single discipline.Management of a hybrid cloud approach using Full stack FinOps is a key part of the solution.However,an integrated management approach is needed to truly deliver radical improvements in how CSPs handle cloud operations.We call this xOps,combining FinOps,DevSecOps and AIOps,to deliv

238、er a large-scale operational transformation.Full Stack FinOps brings accountability and cost management together.However,to be really effective,end-to-end Network Observability and Lifecycle Management(LCM)Automation are also needed.By combining different data types,Network Observability provides ho

239、listic awareness of the health posture and resource utilisation of hybrid cloud infrastructure and associated hosted components.By adopting a DevSecOps mindset,Lifecycle Management(LCM)Automation streamlines Day 0,Day 1 and Day 2 Operations of hybrid cloud infrastructure,hosted applications and serv

240、ices.NTT DATA has built its own set of accelerators to effectively address those areas where additional work is needed for private and hybrid environments.We have identified:7 areas in which changes are needed to extend our existing public cloud FinOps solutions 2 areas where entirely new solutions

241、are required to fulfil fundamental FinOps capabilities for private and hybrid clouds.In an increasingly complex technology landscape,NTT DATA guides CSPs towards a path of agility and transformation with a comprehensive set of solutions which deliver lean hybrid cloud infrastructure management.Take

242、FinOps to the Full StackRegister your interest for a complimentary 45-minute Full Stack FinOps fast-start assessment.About NTT DATANTT DATA is a global leader in telecoms,media and IT services.We deliver the expertise and strategic partnerships that enable customer-centric digital transformation.Thi

243、s enhances sustainable,operational efficiency,customer engagement and brand loyalty.29inform.tmforum.orgSPONSORED FEATURE30A blueprint for intelligent operations fit for the 5G eraThe TM Forum Open Digital Framework(ODF)provides a migration path from legacy IT systems and processes to modular,cloud

244、native software orchestrated using AI.The framework comprises tools,code,knowledge and standards(machine-readable assets,not just documents).It is delivering business value for TM Forum members today,accelerating concept-to-cash,eliminating IT&network costs,and enhancing digital customer experience.

245、Developed by TM Forum member organizations through our Collaboration Community and Catalyst proofs of concept,building on TM Forums established standards,the Open Digital Framework is being used by leading service providers and software companies worldwide.The framework comprises TM Forums Open Digi

246、tal Architecture(ODA),together with tools,models and data that guide the transformation to ODA from legacy IT systems and operations.Open Digital Architecture Architecture framework,common language and design principles Open APIs exposing business services Standardized software components Reference

247、implementation and test environmentTransformation Tools Guides to navigate digital transformation Tools to support the migration from legacy architecture to ODAMaturity Tools&Data Maturity models and readiness checks to baseline digital capabilities Data for benchmarking progress and training AIGoal

248、s of the Open Digital FrameworkThe aim is to transform business agility(accelerating concept-to-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

249、(service providers,their suppliers and systems integrators).Learn more about member collaborationIf you would like to learn more about the Open Digital Framework,or how to get involved in the TM Forum Collaboration Community,please contact George Glass.tm forum open digital frameworkinform.tmforum.o

250、rg31inform.tmforum.orgtm forum research reportsNovember 2021|www.tmforum.orgnext generationthe Telefnica wayAuthor:Mark Newman,Chief Analyst,TM ForumEditor:Annie Turner,Contributing Analystsponsored by:with the support of:March 2022|www.tmforum.orgAuthor:Ed Finegold,Contributing AnalystEditor:Dawn B

251、ushaus,Contributing EditorCALL TO ACTIONAuthors and editors:xxxxxxknowledgeREPORTSponsored by:for integrationsetting new standardsDIGITAL ECOSYSTEMSImage used on tileReport CoverKey WordsBridging boundaries with common standardsWorking together Handing over common standards to each otherLogistics ch

252、ainsCooperationJune 2022|www.tmforum.orgAuthor:Dr.Mark H.Mortensen,Contributing AnalystEditor:Dawn Bushaus,Contributing EditorDIGITAL ECOSYSTEMS setting new standards for integrationsponsored by:August 2022|www.tmforum.orgAuthor:Ed Finegold,Contributing AnalystEditor:Dawn Bushaus,Contributing Editor

253、sponsored by:REPORTSponsored by:Author:Annie TurnerEditor:Dawn BushausISBN:978-1-955998-27-7can telcos the into the future:August 2022|www.tmforum.orgAuthors:Annie Turner,Contributing AnalystDean Ramsay,Principal AnalystEditors:Ian Kemp,Managing Editor Dawn Bushaus,Contributing EditorEd Finegold,Con

254、tributing Analyst,TM ForumIan Kemp,Managing Editor,TM ForumAnnie Turner,Contributing Analyst,TM ForumEditors:Author:Sponsored by:September 2022|www.tmforum.orgAuthor:Ed Finegold,Contributing AnalystEditors:Ian Kemp,Managing EditorAnnie Turner,Contributing Editorsponsored by:Ed Finegold,Contributing

255、Analyst,TM ForumIan Kemp,Managing Editor,TM ForumAnnie Turner,Contributing Analyst,TM ForumEditors:Author:Sponsored by:REPORTAuthors:Sponsored by:Dean Ramsay(Principal Analyst)Editor:Ian KempISBN:000 from toautonomous networks:August 2022|www.tmforum.orgAuthor:Dean Ramsay,Principal AnalystEditor:Ian

256、 Kemp,Managing Editorsponsored by:October 2022|www.tmforum.orgAuthor:Rahul Gupta,Senior AnalystEditor:Ian Kemp,Managing Editorsponsored by:mainframemodernization:charting a course to cloud nativecharting a course to cloud nativeREPORTAuthor:Sponsored by:Rahul Gupta,Senior AnalystEditor:Ian Kemp,Mana

257、ging Editor,TM Forummainframemodernization:February 2022|www.tmforum.orgAuthor:Dean Ramsay,Principal AnalystEditor:Ian Kemp,Managing Editorsponsored by:next generationCEMnow!February 2022|www.tmforum.orgAuthor:Dr.Mark H.Mortensen,Contributing AnalystEditor:Dawn Bushaus,Contributing Editorsponsored b

258、y:October 2022|www.tmforum.orgAuthor:Teresa Cottam,Contributing AnalystEditor:Dawn Bushaus,Contributing Editorsponsored by:from transformationDIGITAL OPERATIONS MATURITY:achieving business valueREPORTSponsored by:Author:Editor:ISBN:at the edge:taking opportunities beyond connectivityIoTMay 2023|www.

259、tmforum.orgAuthor:Mark Newman,Chief 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:Mark Newman,Chief AnalystEditor:Ian Kemp,Managing Editorsponsored by:CSPs take

260、 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:REPORTtransformingBSS:racing to a flexi

261、ble,customer-focused futureJune 2023|www.tmforum.orgAuthor:Teresa Cottam,Contributing 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 suppor

262、t systemsReignitingtelecoms growthAugust 2023|www.tmforum.orgsponsored by:Authors:Sangeet Paul Choudary,Platform Thinking LabsNik Willetts,CEO,TM ForumAnthony Rodrigo,CIO,AxiataDean Ramsay,Principal Analyst,TM Foruma Playbook for CEOsREPORTSponsored by:leveling up:Author:Mark Mortensen,Contributing

263、AnalystEditor:Dawn Bushaus,Contributing Editor achieving Level 3 autonomous networks and beyondAugust 2023September 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 Mort

264、ensen,Contributing AnalystEditor:Dawn Bushaus,Contributing Editor achieving Level 3 autonomous networks and beyondAugust 2023September 2023|www.tmforum.orgAuthor:Mark Newman,Chief AnalystEditor:Dawn Bushaus,Contributing Editor sponsored by:wholesalechanges:rethinking support systemsfor new fiber ope

265、ratorsSponsored 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,Contributing Editor REPORTrethinking support sy

266、stems for new fiber operatorsSeptember 2023|URL TBCwholesalechanges:September 2023|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,Co

267、ntributing Editor TM Forum REPORTrisk management moves firmly into the telco spotlightcybersecurity strategies:September 2023|URL TBCrisk management moves firmly into the telco spotlightCybersecurity strategies:REPORTa numbers game:February 2024Sponsored by:Author:Richard Webb,Senior AnalystEditor:I

268、an Kemp,Managing Editor exploiting data to drive future strategiesREPORTclosing the loopFebruary 2024Sponsored by:Author:Dean Ramsay,Principal AnalystEditor:Ian Kemp,Managing Editor CSPs aim to automate service orchestration and assuranceBENCHMARKSPONSORED BY:SUPPORTED BY:Authors:Mark Newman,Chief A

269、nalyst Dean Ramsay,Principal AnalystEditor:Ian Kemp,Managing EditorMay 2022 inform.tmforum.orgAuthors:Mark Newman,Chief AnalystDean Ramsay,Principal AnalystDawn Bushaus,Contributing AnalystEd Finegold,Contributing AnalystEditor:Ian Kemp,Managing Editor January 2022 inform.tmforum.orgBENCHMARKAuthors

270、:Mark Newman,Chief Analyst,TM ForumDawn Bushaus,Contributing Analyst,TM ForumJoanne Taafe,Editor in Chief,InformEditor:Ian Kemp,Managing Editor,TM ForumSponsored by:a roadmap fortelecomsgrowthTM Forum|March 2023Asia-PacificREPORTAuthor:Sponsored by:Teresa Cottam,Contributing AnalystDawn Bushaus,Cont

271、ributing Editorcounterusing AI to improve customer experienceintelligenceREPORTAuthor: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 Ramsay,Principal Analyst,TM Foru

272、mIan 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 2023cutting complexity AIBENCHMARKAu

273、thors:Mark Newman,Chief AnalystDean Ramsay,Principal AnalystEditor:Ian Kemp,Managing EditorTM Forum|September 2023telcorevenuegrowth:time foroperators toplace new betsREPORTAuthor:Joanne Taafe,Editor in Chief,InformEditor:Ian Kemp,Managing Editor TM Forum|December 2023the sustainable telco:navigatin

274、g the maze of scope 3 emissionsREPORTAuthors: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 economy March 2023REPORTTM Forum|December 2023operators take their first st

275、epsAuthor: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 B2Boperators diverge on the path to cloudBENCHMARKTM Forum|March 2024building

276、anAI strategy:foundations in placetelcos put theAuthor:Mark Newman,Chief Analyst Editor:Ian Kemp,Managing EditorSponsored by:32meet the Research&Media teaminform.tmforum.orgReport Design:Intuitive Design UK Ltd infointuitive-design.co.ukPublished By:European OfficeTM Forum,Uncommon,34-37 Liverpool S

277、treet,London EC2M 7PP United KingdomPhone:+44 207 748 6615US Office181 New Road,Suite 304,Parsippany,NJ 07054 USAPhone:+1 862-227-1648ISBN:978-1-955998-75-8 2024.The entire contents of this publication are protected by copyright.All rights reserved.The Forum would like to thank the sponsors and adve

278、rtisers 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 personal capacities and are their sole responsibility.Their publication does not imply that t

279、hey 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 features in this publication does not in any way imply endorsement by TM Forum of products or servi

280、ces referred to therein.Report Editor:Dawn BushausContributing Editordbushaustmforum.orgManaging Editor:Ian Kempikemptmforum.orgGlobal Account Director:Carine Vandeveldecvandeveldetmforum.orgDigital Media Coordinator:Maureen Adongmadongtmforum.orgCommercial Manager:Tim Edwardstedwardstmforum.orgChie

281、f Analyst:Mark Newmanmnewmantmforum.orgReport Author:Charlotte PatrickContributing AnalystMarketing Manager:Ritika Bhatejarbhatejatmforum.orgHead of Operations:Ali Grovesagrovestmforum.orgPractice Lead:Dean Ramsay dramsaytmforum.orgEditor in Chief,Inform:Joanne Taaffejtaaffetmforum.orgSenior Analyst:Richard Webbrwebbtmforum.orgFor more information about TM Forums Open Digital Architecture,please contact George Glass