《NGMN:解耦网络白皮书v2.0(英文版)(101页).pdf》由会员分享,可在线阅读,更多相关《NGMN:解耦网络白皮书v2.0(英文版)(101页).pdf(101页珍藏版)》请在三个皮匠报告上搜索。
1、 Groer Hasenpfad 30 60598 Frankfurt Germany Phone+49 69/9 07 49 98-0 Fax+49 69/9 07 49 98-41 by NGMN Alliance Version:2.0 Date:20.09.2022 Document Type:Final Deliverable(approved)Confidentiality Class:P-Public Project:ODiN Operating Disaggregated Networks Leadership:Carlos Fernandes(Deutsche Telekom
2、)Javan Erfanian(Bell Canada)Lennart Olaivar(Smart Communications)Editor/Submitter:Carlos Fernandes/Javan Erfanian/Lennart Olaivar Contributors:Deutsche Telekom,Bell Canada,PLDT Smart,BT,China Mobile,Orange,TIM,TELUS,1&1,US Cellular,Chunghwa Telecom,Turkcell,Keysight technologies,InterDigital,Hewlett
3、 Packard Enterprise,Juniper Networks,Fraunhofer FOKUS Programme Office:Chris Hogg(NGMN)Approved by/Date:NGMN Board,20th September 2022 Version 2.0,20th September 2022 Page 3(101)Contributors:Osman Akkaya(Turkcell)Afrim Berisa(Turkcell)Jason Budloo(BT)Chiung-Jang Chen(Chunghwa Telecom)Vincent Danno(O
4、range)Lingli Deng(China Mobile)Javan Erfanian(Bell Canada)Carlos Fernandes(Deutsche Telekom)Varun Gowtham(Fraunhofer FOKUS)Erdal Harput(Turkcell)Kevin Holley(BT)Jinri Huang(China Mobile)Han-Peng Jiang(Chunghwa Telecom)Wei Jiang(China Mobile)Chien-Hua Lee(Chunghwa Telecom)Jian Li(China Mobile)Ting Li
5、(China Mobile)Weiyuan Li(China Mobile)Fabrizio Moggio(TIM)Weichen Ni(China Mobile)Joseph Lennart Olaivar(Smart Communications)Weisen Pan(China Mobile)Frank Qing(Telus)Roy Reyes(Smart Communications)Cheng Choon Si(Singtel)Arvin Siena(Smart Communications)Stephan von Malottki(1&1)Tse-Han Wang(Chunghwa
6、 Telecom)Ming-Yen Wu(Chunghwa Telecom)Pin-Hua Wu(Chunghwa Telecom)Han Yan(China Mobile)Zhiqiang Yu(China Mobile)Herve Oudin(Keysight technologies)Sebastian Robitzsch(InterDigital)Andreas Krichel,(Hewlett Packard Enterprise)Sridar Gopalaswamy(Hewlett Packard Enterprise)Graziano Catucci(Hewlett Packar
7、d Enterprise)Andreas Volk(Hewlett Packard Enterprise)Pavan Kurapati(Juniper Networks)Andreas Meisinger(Juniper Networks)Yuhan Zhang(China Mobile)Version 2.0,20th September 2022 Page 4(101)2022 Next Generation Mobile Networks Alliance e.V.All rights reserved.No part of this document may be reproduced
8、 or transmitted in any form or by any means without prior written permission from NGMN Alliance e.V.The information contained in this document represents the current view held by NGMN Alliance e.V.on the issues discussed as of the date of publication.This document is provided“as is”with no warrantie
9、s whatsoever including any warranty of merchantability,non-infringement,or fitness for any particular purpose.All liability(including liability for infringement of any property rights)relating to the use of information in this document is disclaimed.No license,express or implied,to any intellectual
10、property rights are granted herein.This document is distributed for informational purposes only and is subject to change without notice.Readers should not design products based on this document.Version 2.0,20th September 2022 Page 5(101)Network Disaggregation is one of the mobile telecommunication i
11、ndustrys biggest opportunities while also being a major challenge.The opportunities coming with Network Disaggregation are appealing:a healthier and more resilient ecosystem and supply chain,lower barriers to market entry for new players enabling increased competition whilst also fostering increased
12、 innovation with potentially faster time to market for new products and services.However,disaggregation presents several new challenges,which operators,along with their suppliers,need to address.To ensure disaggregation can be achieved whilst maintaining service levels it demands,a new way of workin
13、g is needed,most likely significant additional integration efforts,changes in the operational model to embrace new processes,as well as the adoption of new skills and new tools.There is also a need to validate whether the benefit of lowering the TCO(Total Cost of Ownership)can be achieved for operat
14、ors and if it will outweigh the complexities involved.Each operator will need to eventually make its choices,depending on its strategy,its starting point(e.g.greenfield or brownfield),its geolocation,competition,market,etc.However,there are many topics which need to be analysed and addressed jointly
15、 to support global standards,economies of scale and to enable competition.Mastering the Route to Disaggregation is a key strategic focus topic of NGMN.This deliverable-the second white paper in the series-has been developed by NGMN partners-operators,vendors,and system integrators.Whilst the first w
16、hite paper 1 outlined the opportunities and challenges of network disaggregation,this paper provides a detailed breakdown of how network disaggregation impacts the network,the organisation and the processes that support the planning,deployment,service providing,optimisation and maintenance of the di
17、saggregated network.RAN,Core and Transport disaggregation is covered as well as complementing topics and activities to the network such as Cloudification,Network Automation,DevSecOps and interoperability and performance testing.In covering-in a methodical way-the impacts to both the network and to t
18、he operators organisation,it is hoped that the paper can act as a reference for operators that covers both network disaggregation technology and process issues.NGMN plans to provide a third white paper to further build on this work and extend it to provide further and more detailed guidance by outli
19、ning network disaggregation architectural options and related operating models matched to specific deployment scenarios and operator needs.It is anticipated that Operating Disaggregated Networks White Paper 3 would be released in 2023.Version 2.0,20th September 2022 Page 6(101).9 1.1 A New Operating
20、 Model.9 1.2 Network Disaggregation-the Transformation Catalyst.10 .12 2.1 Adoption Flexibility.12 2.1.1 More Solution Choices and Flexibility.12 2.1.2 Supply Chain Benefits.13 2.2 Innovation Acceleration.13 2.2.1 Better Functionality,Features and Solutions.13 2.2.2 Better Performance,Improved User
21、Experience.13 2.2.3 Speed of Change.14 2.3 Cost efficiency.14 2.3.1 Lower Cost Attributed to Improved Competition.14 2.3.2 Assumed cost gains attributed to improved Resource Efficiency.14 2.4 Openness for Further Innovation in Automation and AI Platform.15 .16 3.1 Impacts to the Network.16 3.1.1 Int
22、eroperability and compatibility.16 3.1.2 Security.17 3.2 Impacts to the Organization and Processes.18 3.2.1 Impact to Procurement Processes.18 3.2.2 Newly Added System Integration Processes.18 3.2.3 Potential Additional Cost.20 3.2.4 Need for training and competency development.20 3.2.5 Shift of Sco
23、pe and Responsibilities.21 .24 4.1 Operations Layers.24 4.2 Operational Processes.24 4.3 Disaggregation activities and impacts to operating a network.26 4.3.1 RAN disaggregation.26 4.3.2 Core Disaggregation.38 4.3.3 Transport disaggregation.42 4.3.4 Non domain specific Operational Activities.47 4.3.
24、5 Support for Green Technology.50 4.3.6 Integration to End-to-End Service Orchestration and Common Management and Service Assurance System.51 Version 2.0,20th September 2022 Page 7(101).53 5.1 The overall Blueprint.53 5.2 Cloudification domain impacts to essential activities in operating a network.5
25、5 .58 6.1 The Importance of Test for Disaggregated Network from Design,Integration to Operation.59 6.1.1 Front haul Open RAN Disaggregation testing.60 6.1.2 xHaul Transport Disaggregation Testing.61 6.1.3 Disaggregated Core Network Test.63 6.1.4 O-Cloud O-DU/O-CU/RIC Validation at Scale.69 6.1.5 Dis
26、aggregated Network Visibility.72 6.2 The integrated test workflow CI/CD/CT.74 6.3 The Test Taxonomy and Requirements from Test Assets,Process Automation to TaaS(Test as a Service).78 6.4 The Transformation on test process,tools,competencies and organization.79 .80 7.1 Abstract.80 7.2 Expectation of
27、benefit-saving potential.80 7.3 DevOps.80 7.3.1 Recommendations for Joint DevOps functionality.80 7.3.2 Recommendations for Joint DevOps implementation.81 7.3.3 Recommendations to the Standard.83 7.4 DevSecOps.84 7.4.1 What does a DevSecOps approach means to network disaggregation?.84 .85 8.1 Measur
28、able goals of operations.85 8.2 Operation processes evolution.87 8.2.1 Operation processes blueprint.87 8.2.2 Operations Use Cases.88 8.3 Intent based Service Management.90 8.3.1 Goals of intent.90 8.3.2 Execution of Intent.91 8.3.3 Service orientation to its full extent.91 8.3.4 Intent based Servic
29、e Orchestration.93 8.3.5 Intents as knowledge plane.94 .97 .98 .101 Version 2.0,20th September 2022 Page 8(101)Version 2.0,20th September 2022 Page 9(101)1.1 A New Operating Model Mobile network operators have optimised their way of work,leveraging multiple key criteria when planning,deploying,and r
30、unning their networks.Criteria such as ease of integration,performance,capacity,security,and resilience are important to ensure,a high-quality customer experience can be delivered using proven processes and procedures that can cope with the multiple technology vendors and technology generations invo
31、lved.Digital Transformation observed in several domains and areas(e.g.automated industries,market and societal needs,environment,etc.)has required technologies such as 5G to provide solutions to cope with a growing number of use cases with diverse needs such as requiring increased levels of agility,
32、flexibility,scalability,as well as being responsive and cost/energy efficient.This not only applies to the technology but will also impact the teams,processes,and partnerships needed to bring new solutions and services to market.Ultimately,this digitization points to a great deal of prospects but no
33、t without complexities and risks in the absence of sufficient insights,tools,operating models,and end to end alignment.In parallel,the IT and mobile networking technologies which operators rely on to provide their services are in moving to cloud-based solutions.This has resulted in:Separation of fun
34、ctions from underlying hardware;cloudification,and orchestration of containerized functions Service-based architecture,stateless functions,exposure,discovery,and consumption of capabilities Separation of monolithic services into granular micro-services with open APIs Flexible and agile teams combini
35、ng both software development and IT operations(DevOps),providing continuous integration and delivery(CI/CD)of new features and software;use of open-source automation and orchestration platforms Open,interoperable and multi-vendor interfaces,and granular components in a broad ecosystem Network operat
36、ors are following these trends while having high expectations of reliability,resiliency,speed,and low latency,among others,essential for telecommunications networks.Version 2.0,20th September 2022 Page 10(101)As a response to these factors and challenges,the industry continues to drive network disag
37、gregated solutions.As these solutions mature and become increasingly competitive versus the established monolithic integrated ones,operators start to incorporate them in their portfolio.However,this adoption(one can even say transition)of new network capabilities presents several challenges to the e
38、stablished ways of working.Are operators able to use those capabilities without jeopardising their quality of service and operational excellence while at the same time remaining or becoming even more cost effective?Is there a need to adapt the current operating model and how can they decide what,how
39、 and when to do it?1.2 Network Disaggregation-the Transformation Catalyst Network Disaggregation can be seen as both a consequence of the Digital Transformation as well as an accelerator of that journey,and it can be observed broadly from two perspectives:vertical disaggregation,where network functi
40、ons decouple software from hardware,allowing multiple combinations to be used horizontal disaggregation,where established network functions are decomposed into more granular elements and new interfaces are designed and specified Ultimately,this creates more players,able to develop specific component
41、s of the overall architecture,broadening the ecosystem,and leading to an acceleration of innovation.As a consequence,networks are expected to become increasingly agile,flexible,and responsive.All of these factors provide the means to deliver new communication services tailored to the user needs.This
42、 leads not only to new business opportunities but also to many different services which needs to be managed and operated.Considering that those services are based on a multi-vendor ecosystem and on new self-caring technologies,it is evident that there is huge impact on operations.This,in a broad sen
43、se,involves people,processes,technologies and the ecosystem.Disaggregation enables this end-to-end,particularly through openness,cloudification and softwarization,providing network features such as:Separation of control and user plane,programmability and software-defined networking,including SD-RAN
44、Flexibility of the user-plane function,hybrid cloud and edge Version 2.0,20th September 2022 Page 11(101)System flexibility,composable core,granular QoS architecture,multiple-access Network slicing As mentioned above,this leads to potentially significant benefits to performance,user experience,and b
45、usiness opportunities,which will be further detailed.These also leads to the necessity to define and adopt a new operating model,as the deployment and leveraging of such capabilities is deeply intertwined with the way operators are able to control and exploit them.Since 2021,NGMN has identified Mast
46、ering the route to Disaggregation as one of its strategic imperatives creating a new programme named“ODiN”(short for Operating Disaggregated Networks)to address the issue of how to successfully plan,deploy and manage disaggregated networks.The ODiN programme will provide a solid and meaningful guide
47、 to operators,industry partners and telecommunication ecosystem players in general on how to successfully execute this journey.This is the second document from the ODiN programme.Version 2.0,20th September 2022 Page 12(101)The agility and flexibility of disaggregated networks has the potential to pr
48、ovide many benefits,as identified below.It is also expected that a parallel and equal improvement in how operators can operate those networks will depend on the native tools and best practices that come within a disaggregated and cloudified ecosystem.Many of these aspects have an impact on the opera
49、ting model in terms of for example new technologies to master or new processes to set up to take full advantage of the provided benefits.2.1 Adoption Flexibility 2.1.1 More Solution Choices and Flexibility Disaggregation further enables a multi-vendor environment.Vendors can focus on a subset of the
50、 whole pack of solutions which once was expected to be provided by a single vendor.This in turn will allow vendors to specialize on specific products and allocate their resources on those.It is expected this will provide more focus and more competitive products for similar functionality.When this is
51、 replicated by more companies,there will be more competing brands and products in the telecommunication market.This has not happened in the past to such extent.Now,with lowered entrance barriers and,consequently,more suppliers in the market,we expect there will be more choices for the operators to s
52、elect from.This will also enable them to mix and match based on their needs(best of breed approach).Operators for example can now source a Radio Unit(RU)from a different vendor than those of a Distributed Unit(DU)and Centralized Unit(CU).They could choose the best RU,DU or CU and they could do more
53、combinations for each area or cluster type.The same way with Core,operators can now source the hardware from a vendor different than their software or functions vendor.Some operators who have internal Research and Development Teams and are engaged in industry testing and development initiatives may
54、be able to develop solutions faster and customize solutions based on their needs.Bottom line is,it is expected that users get more value for their subscription as disaggregation allows more opportunities to make operator networks more efficient with improved performance.Version 2.0,20th September 20
55、22 Page 13(101)One of the many advantages of disaggregation is the separation of software from hardware.This is actually the key factor that allows more flexibility because hardware and software are now developed separately.This allows more innovations on both.Aside from allowing more developments i
56、n software,operators will also have more choices or options in terms of hardware.Since COTS(Commercial Off-The-Shelf)can now be used,IT branded hardware could also be used for telecommunication applications.This should enable operators to acquire the best,latest and most technologically advanced and
57、 efficient hardware.The latest technologies and functionalities could then easily be deployed using software upgrades.It is expected this will also lead to a faster time to market.2.1.2 Supply Chain Benefits On one hand,disaggregation adds the possibility and capacity to ensure operators can access
58、a more diverse supply chain,sourcing components from multiple vendors and multiple geographies and therefore allowing for more resilient networks and processes.On the other hand,global operators could have more opportunities to localize the supply,giving opportunities to competent local companies.Th
59、is may encourage more local suppliers to develop solutions and join the telecommunication business even if they are from other industries.2.2 Innovation Acceleration 2.2.1 Better Functionality,Features and Solutions Disaggregating or breaking the network components into more parts and opening the in
60、terfaces will allow more companies,including disruptive and emerging ones,to develop solutions as well as hardware and software products.This will allow faster development of technology as more minds,teams and companies are working towards one goal improving solutions,customer experience,and making
61、networks more efficient.Disaggregation will also allow for more and better customization of products based on the specific needs of each operator.Operators are able to buy only the features and functionalities that they need.It is expected that this will translate to a more efficient solution.2.2.2
62、Better Performance,Improved User Experience Increased innovation in each part of the network is expected to cause better KPIs and improve performance in mobile networks.These developments will allow each network Version 2.0,20th September 2022 Page 14(101)component to contribute to a better performi
63、ng system and will ultimately lead to better end-user experience and improved services.2.2.3 Speed of Change Increased competition due to lowered barriers to market entry will also provide incentives to vendors who will develop better and more efficient products that in effect develops the market as
64、 a whole.This will benefit the operators even more as it is expected that technology will keep on getting better in a shorter span of time.The level of flexibility and agility,which affects the speed of change,will be dependent upon the development of each vendor.It is assumed that some will be more
65、 flexible,and some will be less flexible.With this,some will be capable of being faster than others.At any rate,changes to the network will be faster compared to today.This will allow operators to scale their networks better based on their needs.Due to more granularity and flexibility of solutions,o
66、perators will be able to do expansions,upgrades or any changes with less effort and faster.By disaggregating the network,operators can better manage network demand by scaling up network functions as needed.Upgrades and maintenance can also be better managed as operators can upgrade each part singula
67、rly as opposed to upgrading the entire network.This greatly improves the life cycle of network services and time-to-market when providing new services to new customers.2.3 Cost efficiency 2.3.1 Lower Cost Attributed to Improved Competition One of the promises of disaggregation and open interfaces is
68、 lowering cost,with expectations-based on the experiences of one greenfield operator-for specific network domains such as RAN to reach 30%to 40%lower CAPEX and OPEX 2.As networks continue to expand,develop,and transform,operators need to invest significantly.Operators are simultaneously trying to en
69、sure cost efficiency and disaggregation is the best candidate solution to fulfil that.2.3.2 Assumed cost gains attributed to improved Resource Efficiency Disaggregation allows operators to centralize functions and control.This enables better efficiency by leveraging on pooling gains which could also
70、 translate to lower CAPEX and maintenance cost.Since software and functionalities are disaggregated,they could be installed or housed on shared hardware.With this,operators could implement a common hardware or infrastructure from RAN,Edge to Core in order to simplify engineering,Version 2.0,20th Sep
71、tember 2022 Page 15(101)implementation and operations.It is assumed this leads to reduced costs by leveraging on economies of scale.2.4 Openness for Further Innovation in Automation and AI Platform By disaggregating the business capabilities and control capabilities in the network,common capabilitie
72、s are achieved and provided in a“platform”way.A unified AI platform is built,where specific services could be called through network elements.The unified AI platform can provide intelligent application R&D with infrastructure services including centralized computing power,algorithm frameworks and ge
73、neral AI capabilities,realizing one-stop management of network intelligent application R&D,operation and maintenance.After disaggregation of the networks business capabilities and intelligent capabilities,independent R&D can be carried out based on the AI platform,promoting R&D efficiency improvemen
74、t and cost reduction.Version 2.0,20th September 2022 Page 16(101)Benefits outlined in the previous section will not be achieved unless the industry manages to overcome a number of challenges.3.1 Impacts to the Network Making solutions more flexible and scalable has an inherent challenge which is com
75、plexity.As hardware and software are separated,as well as their individual development,the overall solution becomes more complex because there are more and more solutions available that need to be able to work together.As each company developing different parts has different roadmaps,the complexity
76、that it will bring will add to the challenge of ensuring compatibility and interoperability between different vendors solutions.This will have significant impacts on the options operators have for managing this complexity.This compares to the present situation where the operator manages this complex
77、ity themselves or(more often)appoints a single or small number of system integrators or vendors to have oversight of the solution.The system integrator(s)/vendor(s)then work closely with the other partners in the project to manage the complexity.3.1.1 Interoperability and compatibility The move to a
78、 disaggregated network solution needs to go hand in hand with the assurance of interoperability and compatibility.Operators must not find themselves in a situation where vendors are pointing at each other on how to integrate or problem solve when the customer experience and brand value is on the lin
79、e.This is the part where operators need to be assured as this will impact network quality,customer experience,time to implement,optimization and maintenance.In having multiple and many suppliers,one of the major concerns is interoperability.Interoperability issues can only increase in the near term
80、with the longer-term aim to reduce these issues.Currently,most if not all,traditional partnership model vendors who are complying to 3GPP and other telecommunication standards are also performing interoperability tests with each other.Yet,operators are still encountering interoperability issues espe
81、cially on inter domain connections or interfaces.This is for instance caused by each vendor applying their own interpretation of the standards.It is anticipated that this will escalate further,rather than Version 2.0,20th September 2022 Page 17(101)improve,when disaggregation is introduced into the
82、networks since there will be more types of solutions and vendors to connect.Each of these solutions may have different and independent developers,roadmaps and interpretations of the standards leading to different implementations.They also have different timelines in terms of development,e.g.,typical
83、ly software has a faster development cycle than hardware.Therefore,it has to be expected that compatibility and interoperability will be a huge challenge.3.1.2 Security Operators must continue to ensure that their networks and services are secure.This is even more important when mobile networks are
84、increasingly becoming critical national infrastructure providing services to a wide range of industries.Although operators and vendors have a long-established history of managing the security of mobile networks,the systems and processes used are likely to come under additional strain as the number o
85、f vendors,functions,and interfaces in a typical mobile network increase.For example,as more components and functionalities are introduced,the network potentially becomes vulnerable as there are more integration points.This could arise in the open interfaces in any network domain(RAN,Core,Transport),
86、open-source software and off the shelf solutions.Functional splits and Edge computing could also contribute to wider physical attacks.Disaggregation,if hosting software to the public cloud,could also introduce more vulnerabilities and attacks as the network tends to be more exposed to the public dom
87、ain.Different software and different hardware might cause new vulnerabilities as they are developed separately.Each of these vendors or companies have different experiences towards attacks.This is where consolidated monitoring becomes essential.Each of the software and hardware needs to be fully mon
88、itored to detect any possible intrusion.Multiple patches and updates could also introduce incompatibility and security risks across different versions.The need to maintain backwards compatibility and have a strict regression testing regime will be key to maintain security.Version 2.0,20th September
89、2022 Page 18(101)3.2 Impacts to the Organization and Processes 3.2.1 Impact to Procurement Processes Disaggregation will have a huge impact in the supply chain and procurement strategies and processes.The number of suppliers will increase,and this may mean that the procurement team needs to expand t
90、o be able to handle more suppliers.Though disaggregation brings a huge benefit by expanding the telecommunication ecosystem,it will also cause complexity to the system and processes,as more vendors would need to be managed.Different vendors/suppliers have different SLAs,hence the variables in purcha
91、se and delivery will tend to broaden.More bricks lead to bigger challenges-both technical and legal/contractual.Responsibility delineation is also expected to be a huge challenge to the supply chain.Procurement teams may encounter difficulties in identifying who should be responsible for a specific
92、purchase,warranty,operations,etc.More and smaller components mean more parties(either new players,or existing players that were masked by integrators/vendors in the past)to deal with.There will be new software components which again means more parties to cooperate with.New players(including possibly
93、 start-ups)will impact our current processes/habits to interact with the industry players.3.2.2 Newly Added System Integration Processes 3.2.2.1 Changes in the Organization and Processes The huge impact and changes brought about by disaggregation will cause an impact to the organisation of each of t
94、he operators.System integration is one of the biggest challenges in adopting disaggregation,since this was not typically part of the organisation during the traditional or legacy days.So,the operators would need to either build their own team of system integrators or tap an external entity or compan
95、y that will do system integration for them.Both options will entail huge effort,adjustments,changes in the organisation,and potentially additional cost.Today,almost all the system integration work is being done by the vendors of the respective equipment.Typically,operators biggest responsibility is
96、interfacing and understanding the needs of the business,translating them to technical solutions,planning the implementation,decision making and governance.The rest is mostly passed on to the vendor for execution Version 2.0,20th September 2022 Page 19(101)from detailed design,to implementation,optim
97、isation,and maintenance.Now that with disaggregation networks are broken into smaller parts,the responsibility of bringing everything together cannot be passed on anymore to a single vendor because of this new multi-vendor environment.There will no longer be a single vendor taking care of the overal
98、l solution and its management.The responsibility of successful integration will now be on the shoulders of the operator.If the operator decides to build its internal system integration team,it will have to make either major rearrangements to re-purpose manpower or hire significant resources to fill
99、the gaps.There will also be a massive change and adjustments in the processes of the organisation.Coming from a set-up where much is done by the vendor,to the operator playing a bigger role in terms of putting everything together,from design to operations and management,will entail change at all org
100、anisational levels of the operator.Integration is said to be a huge challenge for the operators.On the other hand,it will also be challenging to the vendors.They would need to be more conscious of what the other vendors are doing and developing as they need to interoperate.This fact would need to be
101、 considered also by procurement to ensure nothing is missed in the process or purchase.3.2.2.2 Many Components and Companies to Deal With At the moment,operators are talking to two to four vendors per domain.With disaggregation,operators could be talking to more than five for RAN alone.That could ev
102、en go higher if they choose to be more flexible and choose more vendors.That will add to the complexity,not only to the solution,but also to the organisation and processes of the company.Firstly,it is assumed that operators need to add more manpower to handle such vast number of vendors.Secondly,ope
103、rators need to adjust their procurement processes and strategies to adopt to more vendors supplying the requirements.The separation of software and hardware alone will instantly add to the complexity as operators are historically used to buying both software and hardware from the same vendor.Having
104、them separate would mean there will be separate services for each.This may not only add to the cost but make things more difficult as operators will talk to more people and have more interfaces.The challenge will truly come during implementation,troubleshooting and problem resolution.Operators would
105、 need to talk to at least two entities for a single node.Version 2.0,20th September 2022 Page 20(101)3.2.3 Potential Additional Cost 3.2.3.1 Integration Cost System integration of different parts of each domain is normally not part of what operators spend on at the moment,as the same vendor is suppl
106、ying both the hardware and the software,and most solutions are integrated.There is no integration needed because software is already integrated in the hardware.The beauty of the current set-up is its simplicity.With disaggregation,since software is separated from hardware and functions are separated
107、,operators would need to have an integration team that will combine or install the software to the hardware making sure of its compatibility,facilitate proper operation of the equipment and proper interoperation of all network functions as well as ensure functionalities work together as a whole.This
108、 will either need a separate vendor or a formation of a new team within the company.This will translate to additional effort and potentially additional cost.With all the additional efforts and costs,it is imperative that the benefits outweigh the challenges.The expected cost effectiveness and the ne
109、w business possibilities enabled by the new ecosystem,attributed to the separation of hardware and software,should offset the operational changes that need to be made.Cost reduction needs to cover for the needed offset and still maintain a worthwhile net decrease in TCO.All the efforts should be com
110、pensated in the long term.Otherwise,the decision making will be very hard as everything needs to be justified given that the current set-up and system is working well.There is a saying that goes“Do not fix something that is not broken”.The current system and solutions are not broken but are not flex
111、ible enough to sustain new possibilities and services of the upcoming years.Disaggregation is there because of the industrys desire to make things better and more efficient.3.2.4 Need for training and competency development Training and competency development are efforts that are needed in disaggreg
112、ation.Most of the engineers are more familiar with integrated solutions as they worked on them for many years.Disaggregation is new to most of the engineers in the network.This is why there needs to be proper education within the organization and industry in order for the manpower to be better equip
113、ped to operate the new network.If operators decide to build a new integration team within their company,operators would need to spend time and budget to build the competency and expertise of that team.On the other hand,if operators decide to outsource system integration,they will still need to devel
114、op the competency of manpower who will govern external SI.Version 2.0,20th September 2022 Page 21(101)3.2.5 Shift of Scope and Responsibilities 3.2.5.1 Who is responsible?The breaking of a whole system into parts raises the question“Who is responsible?”.This will be highlighted especially when there
115、 are network issues or collective customer complaints where there is a problem in the network that hasnt been identified yet.In the traditional process,a single vendor would conduct tracing of the whole network and identification of problem and isolation is easier because that vendor has all the cou
116、nters and measurements for each part of the network,therefore it is easier to identify which part of the network is causing problems.Since that vendor is providing all the parts and solutions to the whole system,only one entity or company is responsible.Therefore,escalation and troubleshooting is si
117、mpler.With disaggregation,even in one node,several vendors might be involved.For example in a gNodeB,one vendor might be supplying hardware for the CU,another for the DU and another for the software of both or one of these functions.Then another vendor is supplying the RU and another for the antenna
118、.If say that gNode B or site is having performance issues,it is not straight forward to tell which component is problematic.Is it hardware or is it software?Or is it the combination of the two that causes the trouble?Even when the problem is identified,the escalation and troubleshooting wont be as e
119、asy as several teams from different parties may need to cooperate to solve the issue.Lets take for example an extreme case in traditional solution having only one vendor for all the RAN components in the network.Since that single vendor handles and is responsible for the RAN of the whole network,it
120、will have a large support system in terms of resources,tools,manpower,technical support,ticketing system that escalates all the way to R&D and the main office.In that case,any issue in the network will be identified,troubleshooted and resolved by that single vendor who might have significant resourc
121、es that could respond to an issue anywhere in the network,quickly.On the other hand,when network is disaggregated and the components of the network are handled by different vendors,each of those vendors would most likely have less or fewer resource supporting the networks.This might cause them to re
122、spond slower.This goes back to the bottom line and question again of“Who is responsible?”.To resolve this,there should be a central team that oversees and monitors all components.Then again,building and equipping this team wont be as easy as there are many different components in the network with di
123、fferent brands.That would mean additional effort as the Version 2.0,20th September 2022 Page 22(101)team should be capable of monitoring and understanding all the components in the network.This again would imply effort on training and competency development and translate to OPEX.One of the common qu
124、estions is when there are major issues in the RAN network,who would identify if the issue is caused by the software or the hardware?There should be a way or a tool that could easily pinpoint that.Now the question arise on who will develop that tool and if that tool supports any software or hardware
125、vendor,given the vastness of the ecosystem that is continuously growing until now?Troubleshooting will eventually be more complex when networks get disaggregated.With that,it can be expected that vendors can differentiate themselves through support offerings.3.2.5.2 Software Asset Management(SAM)Dea
126、ling with software components will require proper licensing and asset management:The choice of licensing model should be adjusted to the use cases.Concerning the duration of the rights of use,the rental model(i.e.,subscription)could be considered with caution to avoid explosion of OPEX and perpetual
127、 rights should be preferred to optimise investments.On the other hand,as cloud services are generally monetised based on their usage,it is logical to reflect this model on the licenses of network functions,which will have to be based on Pay-as-you-Use models.This means that usage rights must be quan
128、tified based on usage metrics related to the value generated by the network function.The potential impact of a significant increase in operational costs related to the SAM process throughout the life cycle of Virtualised Network Function(VNF)/Cloud-Native Network Function(CNF),from procurement to de
129、commissioning,could be mitigated by an adapted tooling approach,to guarantee usage compliance while controlling operational costs.Automation of SAM processes is therefore essential,just like other business processes of the operator.This automation is only economically feasible if this process can be
130、 applied in a unified manner to all network software and regardless of the suppliers.This is best possible if this is based on standards.The purchase of licenses adapted to the usage implies being able to estimate this usage at a certain time in the future.This can be done by observing the evolution
131、 of current usage,but this is not sufficient.Network functions can be viewed as being“organized/deployed”as“stacks”(in a“client/server”or“vertical”type of association)and“service chains”(i.e.,in a“horizontal”type of association).To simplify,it can be said that it is useless to buy usage rights for a
132、 function if the usage rights of the one(s)on which that function relies would Version 2.0,20th September 2022 Page 23(101)not allow to exploit them.On the other hand,it is useless to buy usage rights for a function that is part of a chain of functions(Network Service),if the usage rights of one of
133、the functions in the chain would not allow to exploit them.It is therefore considered necessary to be able to consider the management of usage rights for network functions in a comprehensive way.Version 2.0,20th September 2022 Page 24(101)This section introduces the overall organisation of the netwo
134、rk operation,and further analyses the potential impact to the existing operation and organisation from the disaggregated activities in various domains.4.1 Operations Layers The overall network operation includes 3 layers:Business Operation Layer Business operation is about CSPs Product Portfolio pla
135、nning,development,operations and other roles,information or activities toward market and customer requirements.Service Operation Layer Service operation layer represents roles,information and activities that are involved in the strategic planning,definition,development,and operational aspects of ser
136、vices that are used to realise product offerings to the market.Resource Operation Layer Resource operation layer is about the activities related to the enterprise infrastructure,e.g.,computing,networking,and storage resource capabilities to support the operation of the services.In the context of our
137、 focus on operational models in this document,the following analysis is limited to resource and service operation layers only.4.2 Operational Processes Each operation layer includes several vertical operational processes including:planning,deployment,maintenance,optimisation,and service providing.Th
138、e following technical-agnostic descriptions are applicable:Planning Based on market and product portfolio strategy and forecasts,research&analysis is performed to determine service and resource targets as well as strategies.This includes expansions of the existing service and resource capabilities a
139、nd the identification of new service and resource capabilities,service and resource support levels and approaches Version 2.0,20th September 2022 Page 25(101)required,service and resource design elements to be developed,as well as service and resource cost parameters and targets and defining the way
140、 that new or enhanced infrastructure may be deployed.These processes also define the policies relating to technical services or resource and their implementation.Deployment Based on demand,plan and deliver the total capabilities required to deliver changes to service,as necessary.This may involve in
141、tegration of capabilities delivered from within the MNO,and capabilities delivered from an external party.This also involves the use of capability definition or requirements to deploy new and/or enhanced technologies and associated resources,ensuring that network,application and computing resources
142、are deployed according to the plans set.It also entails delivering the physical resource capabilities necessary for the ongoing operations and ensuring the basis on which all resources and services will be built.Maintenance Managing Service and Resource infrastructure,ensuring that the appropriate s
143、ervice capacity,application,computing,and network resources are available and ready to support Fulfilment,Assurance and Billing processes in instantiating and managing service and resource instances,and for monitoring and reporting on the capabilities and costs.Optimisation Managing,tracking,monitor
144、ing,analysing,improving and reporting on the performance of specific services and resources.Collect and/or distribute management information and data records between resource and service instances and other operator IT processes.Service providing Manage problems associated with specific services.The
145、 objective of these processes is to respond immediately to reported service problems or failures to minimize their effects on customers,and to invoke the restoration of the service,or provide an alternate service as soon as possible.Version 2.0,20th September 2022 Page 26(101)4.3 Disaggregation acti
146、vities and impacts to operating a network 4.3.1 RAN disaggregation 4.3.1.1 RAN Disaggregation activities RAN is a challenging and complex domain to disaggregate due to the legacy and the number of moving parts within the system.RAN disaggregation will involve a broad and diverse ecosystem providing
147、flexibility and choice.Disaggregation considerations in RAN There are a number of considerations with respect to disaggregation in RAN such as:a)Operators need to maximize their investment on Distributed RAN(DRAN)since legacy technologies(4G,3G,2G),and even 5G,were deployed using DRAN.Transitioning
148、to virtualised RAN(vRAN)or Open RAN(O-RAN)would entail a total change in architecture and cause major adjustments in operators organisation,processes,planning,dimensioning and more.There are still many unknowns that translate to challenges,particularly for brownfield operators.There may also be fact
149、ors that are unseen or not visible at the moment that may cause major impact in the future once disaggregation in the RAN starts and matures.b)Transitioning to Centralised Unit Distributed Unit Radio Unit(CU-DU-RU)in the Radio Access Network and(in the transport network)to Front Haul-Mid Haul-Back H
150、aul(FH-MH-BH)will greatly impact transport configurations and topology as well as requirements on bandwidth and latency.This is one of the most important considerations during disaggregation.Over the years,most operators have simplified their architecture so as to become more efficient and simpler i
151、n terms of operations,yet disaggregation impacts this trend.Disaggregation may require transport reconfigurations with potentially significant and costly effort.In the past,1G was enough for back-haul.Now,the minimum has become 10G and the ideal is 100G.This is because of the huge bandwidth requirem
152、ent of 5G especially in the millimetre wave band and even in sub 6 GHz band.Along with that,strict latency requirements will now need to be considered for front haul.This will again require huge cost and reconfiguration of transport.So,there is a risk the savings from RAN may be challenged by the tr
153、ansport cost.c)The dimensioning principle will become totally different.As complex as the technologies from 3G to 5G have been,operators did manage to simplify the dimensioning in order to be flexible especially in budgeting and allocating resources.With vRAN/O-RAN,operators Version 2.0,20th Septemb
154、er 2022 Page 27(101)will need to dimension even the smaller parts of the hardware such as the processors/compute and storage.Despite these many challenges,if RAN disaggregation proves to be beneficial in terms of performance,customer experience,better services,energy efficiency,and lower costs,it wi
155、ll have a huge positive impact to the network and to operators businesses.RAN Function Disaggregation(Horizontal Disaggregation)In RAN function disaggregation or Horizontal disaggregation,the functions of the BBU,where the processing of baseband signals and intelligence are done,is split into CU and
156、 DU.This divided the RAN architecture from BBU-RU to CU-DU-RU.The split of functions allows the operators deployment flexibility in a way that each function could be positioned in different parts of the network.This means that,aside from separating the functions,the hardware can also be separated.Fo
157、r example,DU and CU can now be placed either on site or further away,e.g.at a Data Centre.Another option is to place CU and DU separately closer or farther from the site.The positioning of these functions will impact or potentially make the deployment cost become lower.If it is positioned in a way t
158、hat the DU or CU functions are centralized,cost will potentially decrease since the resources are pooled and more sites are sharing the same hardware and software.This is called pooling gains for which its notable gains need to be validated.Different positions require different transport requirement
159、s in terms of bandwidth and latency.This would require a conscious decision by the operator on which topology or configuration to use as one would be less expensive in the RAN but might cause higher cost in the transport.A balance of this would be required to achieve the most cost effective and effi
160、cient solution.A distinct benefit of horizontal disaggregation is flexibility,choice and diversity of providers to maximize performance,capabilities and efficiencies.Operators now have the liberty to choose different vendors for CU,DU and RU thereby giving flexibility in terms of solution and costin
161、g.Doing such may come with challenges to be addressed,in terms of visibility,management,and integration,among others.Introduction of RAN Intelligence Another advantage of disaggregation is that it enables operators to easily inject intelligence into the RAN network.ORAN developed RAN Intelligent Con
162、troller(RIC)which incorporates Artificial Intelligence(AI)and Machine Learning(ML)for dynamic,intelligent,and predictive allocation,policy,management,optimization,and operation.Version 2.0,20th September 2022 Page 28(101)Figure 4-1:RAN Intelligence As a software-defined platform,the RIC brings intel
163、ligence,programmability,and extensibility to radio access networks.RIC uses AI and machine learning(AI/ML)applications that automate RAN operations and support innovative use cases.With the RIC,network operators have a platform to deliver new functions and user experiences with greater agility and e
164、ase.RIC comes in 2 forms-Non-real time RIC and Near-real time RIC.Non-real time RIC integrates intelligence into RAN system design in performing network management tasks and work for control loops over 1 second.Near-real time RIC controls CUs and DUs and performs network improvement and optimization
165、 decisions that happens between 10 milli seconds to 1 second.In addition,there are specialized applications called rApps and xApps.The Non-RT RIC,expected to run in a cloud,enables greater-than-one-second control and policy guidance over the RAN elements and their resources through rApps.It also ena
166、bles AI/ML capabilities for the RAN.The Near-RT RIC is responsible for fast loop control of the RAN Version 2.0,20th September 2022 Page 29(101)network functions.It provides less-than-one-second control over the RAN nodes and resources which are driven by the non-real-time RIC.It can host and deploy
167、 specialized xApps.The Non-RT RIC communicates with the Near-RT RIC over the A1 interface to provide policy-based guidance to the xApps running on the Near-RT RIC to optimize RAN behaviour,such as for capacity,customer-specific service levels,or energy efficiency.The Non-RT RIC uses long-term networ
168、k data,such as performance metrics as well as enrichment data from external applications to train and generate AI/ML-driven applications.The RIC platform is based on a cloud-native microservices architecture and needs to be fully compliant with the O-RAN specifications and interfaces.It needs to sup
169、port both an open API and a software development kit(SDK)for integration with any third-party O-RAN-compliant xApps or rApps,giving network operators greater flexibility and choice of suppliers.rApps and xApps are the foundation for innovation and agility in the RAN.These specialized,AI-driven appli
170、cations,allow operators to enable new business models,personalize the service experience,and optimize CapEx and OpEx.Key use cases include RAN slice SLA assurance,tenant-and slice-aware admission control,traffic steering,energy efficiency,M-MIMO optimization,and quality of experience(QoE)optimizatio
171、n.Figure 4-2:RIC Framework As mentioned above,rApps are running on the Non-RT RIC since they are less low latency critical.Examples for rAPP use are:Network deployment use-cases Network automation use-cases Network optimization use-case Version 2.0,20th September 2022 Page 30(101)Network healing use
172、-cases xApps are addressing use cases which are more time critical,not necessarily but very often they work in a combination between xApps on Near-RT RIC with the support of rApps from the non-RT RIC.Some example use-cases,which are using a combination of xApps and rApps are:Network SlicingNetwork s
173、licing is a key advancement in 5G networks,with end-to-end connectivity and data processing tailored to specific customer requirements or workloads.The service levels are guaranteed and must be continuously assured across the delivery chain.rApps/xApps can continuously monitor each slice and collect
174、 slice-specific performance metrics.If the application detects a SLA violation,it can immediately initiate corrective action by making the appropriate configuration changes to the centralized and distributed units(CUs and DUs)and updating the policy accordingly.The changes are monitored and confirme
175、d as meeting the specified service levels.Tenant-and Slice-Aware Admission ControlThis application allows for real-time tracking and enforcement of radio resources such as packed data units(PDUs)per slice,and user equipment per slice.This use case is required to provide priority services for hospita
176、ls,schools,public safety,and other high priority users to ensure that communications are delivered with efficiency and predictability.SteeringTraffic steering allows operators to meet capacity demands while avoiding additional capital investments.The RIC and the associated apps can monitor the dynam
177、ically changing network load,using AI/ML-based steering algorithms to distribute the load to different frequencies within the same base station,to neighbouring base stations,or even to different radio access technologies,resulting in efficient utilization of operator resources.Energy EfficiencyAI-dr
178、iven predictions and controls can be used to optimize energy efficiency of the RAN,switching off antennas as needed to increase energy efficiency.Insight into traffic,coverage,interference,and other factors can also be factored in to identify long-term trends and enable strategic planning.Massive MI
179、MO CoverageA key advantage of 5G,massive multiple input and multiple output(M-MIMO)provides greater capacity and minimizes interference.By applying AI/ML and decision-making in real time in conjunction with M-MIMO and beam forming,the RIC can proactively and continuously improve the subscribers expe
180、rience even in dense areas or at times where demand is surging,such as in crowded cities or entertainment venues.This could potentially Version 2.0,20th September 2022 Page 31(101)contribute to energy efficiency as the beams are efficiently utilized to capture or cover more users and traffic while u
181、sing the same or even lower power.Quality of Experience(QoE)Intelligent,real-time controls allow a better user experience for latency-sensitive or bandwidth-intensive applications like cloud virtual reality,drones,or autonomous vehicles.The RIC and associated applications can use analytics to take p
182、olicy-based actions,ensuring that priority users maintain a satisfactory QoE and experience even during peak loads.Software Disaggregation from hardware(Vertical disaggregation)In vertical disaggregation,software becomes independent of hardware.This means that operators could choose different vendor
183、 for hardware and for software.This allows operators to choose the best of breed solution.They could choose the best and most economical hardware based on the needs of their network and subscribers.They could also choose the most flexible and cost-effective software that will allow them to be agile
184、in terms of developing and releasing new services.Just like Horizontal disaggregation,the flexibility of vertical disaggregation comes with a price,which is complexity.Since software is developed independently of hardware,the possibility of interoperability issues is higher.This could be countered t
185、hough through testing and certification.4.3.1.2 Disaggregation impacts on RAN operational process Impact on RAN Planning The overall architecture will change from RU-BBU to CU-DU-RU to implementing the different functional split options.This will significantly impact the planning,dimensioning,and en
186、gineering of sites.From simply putting RU-BBU to each of the sites and just dimensioning the number of bands and carriers based on the expected and future traffic that needs to be supported and carried,to dimensioning the DUs depending on how many RUs and how much traffic will be homed/connected or
187、supported.CUs will also have to be dimensioned separately depending on how many sites or DUs will be homed to it.These will all depend on what functional split options operators choose to implement.Additionally,when planning the new RAN in addition to normal engineering based on traffic profiles and
188、 available sites,frequency bands and carriers,operators will now also need to consider the impact on the Resource Operation Layer to ensure that sufficient computing power,networking and storage is available to support the Services that comprise the Version 2.0,20th September 2022 Page 32(101)disagg
189、regated RAN.The amount of traffic that a specific Service instance(e.g.DU or RU)can support will depend on the number of cores/processors in the instance.Planning and Engineering teams need to learn different styles of dimensioning that is similar to IT and apply it to network.Impact on RAN deployme
190、nt:RAN deployment in disaggregated network becomes flexible and complex at the same time.In the traditional networks all RAN components(BBU and RU)are installed to each site,which is costly because each site needs all the supply and services for all RAN components.In disaggregated RAN,operators can
191、choose to either replicate the traditional,which is deemed costly,or centralize DU and CU.In centralised architectures,the site could consist only of RU,antenna and ancillaries which will be faster in terms of implementation.BBU functions,which now becomes the CU and DU could be placed in a central
192、location or Data Centre which could handle more sites.This is like how the RNC is positioned in 3G and the BSC in 2G.The difference is that this new architecture could centralise more sites as it allows stacking up of servers.In this case,the deployment of massive number of sites could potentially b
193、e faster and cheaper.The challenge here is the resiliency of the solution.Since the network functions are centralized,it is prone to more down time during disasters.To counter this,a good resiliency or multi-homing architecture would be required.In addition,further approaches to resiliency such as h
194、ot-standby should be considered.Impact on RAN maintenance Since RAN is broken down to more parts horizontally and vertically,RAN maintenance will become more tedious as more expertise and tools are needed for the operations or field teams.Disaggregation will cause operators to use more brands theref
195、ore requiring operations engineers and staff to have more knowledge and skills both in hardware and software of each brand.This impact would have to be managed properly to ensure cost and organisational effects will be at an acceptable level.The current set-up that could manage 3 to 5 vendors wont b
196、e sufficient anymore.The team has to be capable of handling multiple hardware and software vendors.This would require more training and even more people.With this,huge and proper preparations are needed by operators in order to ensure down times and service interruptions are avoided.The whole organi
197、zation has to be ready when disaggregation is adopted.Version 2.0,20th September 2022 Page 33(101)Impact on RAN Optimisation process Since hardware and software is separated during disaggregation,optimisation will be done separately.Optimisation will also become more complex as there is more hardwar
198、e and software to measure,monitor and troubleshoot.The optimization team would need capabilities to build expertise on hardware and software separately.That would at least be twice the effort and might be twice the cost as well.The optimisation engineer and teams would need to fully understand the t
199、opology in the areas he/she handles for him/her to determine where the problem,bottleneck or failures are.This would be simpler if the monitoring and service assurance would be connected to just one platform where the CU,DU and RU hardware and software can be viewed.Otherwise,it will be more difficu
200、lt for the optimisation engineer and team because he/they would have to look at many monitoring tools.This also needs understanding and expertise in each of the software and hardware brand for each functional element so intense training and competency development is required.Separation of function i
201、n disaggregation will also impact the operations team as they need to build expertise in all the new brands or vendors that will be integrated in the network to ensure all issues will be addressed or resolved.Impact on RAN Service providing Service providing impact needs to be determined to ensure d
202、isaggregation will not affect problem management and troubleshooting of services,functions,or features in a negative way.Disaggregation should help expedite the process and make it more efficient.The operations team should be able to respond quickly to any service problems or failures and ensure the
203、re is no,to minimal,effect to subscribers whenever there are service interruptions.There is a hope that complexity of disaggregation will not make matters worse.Need for Identification of applicable and best topology/configuration There were originally several functional splits/options to choose fro
204、m that were identified by 3GPP(see figure 4-3)in RAN which allows horizontal flexibility in a way that operators have the option to place RAN functions anywhere in the network depending on what is best in terms of flexibility and efficiency.The industry has adopted an upper and a lower functional sp
205、lit approach.In particular,Option 2 is standardized by 3GPP and O-RAN has adopted option 2(upper split)or option 7.2x(lower split).Version 2.0,20th September 2022 Page 34(101)PDCPLow-RLCHigh-MACLow-MACHigh-PHYLow-PHYPDCPLow-RLCHigh-MACLow-MACHigh-PHYLow-PHYOption 5Option 4Option 6Option 7Option 2Opt
206、ion 1RRCRRCRFRFOption 8DataDataHigh-RLCHigh-RLCOption 3 Figure 4-3 RAN Functional split options source:3GPP TR38.801 Operators could formulate several topology and configurations based on the upper and lower splits and be able to come up with models that will be applicable to different situations in
207、 the network that considers flexible use of resources while maintaining high performance and user experience.Need for Selection of vendors Vendor choice for each part Operators need to formulate several options and analyse its Pros and Cons so they could decide which best will fit their network requ
208、irements,future plans and the current capabilities of their organization similar to the following extremes:Option 1:Use same vendor for hardware and software Figure 4-4:Use of same vendor for Hardware and Software Version 2.0,20th September 2022 Page 35(101)The advantage of this option is that the c
209、oordination,planning and engineering is simpler since we coordinate and work with fewer vendors and type of hardware and software.The disadvantage is that it somewhat defeats the purpose of disaggregating which should allow us to choose more vendors/suppliers and be able to mix and match them.Option
210、 2:Use Totally Different Vendors Figure 4-5:Use of totally different vendors The advantage of this option is that we are able to take advantage of the disaggregation in a way that we are able to choose the best vendor for each part.The disadvantage is that we are talking,coordinating,planning,engine
211、ering and implementing with too many vendors.This is very complex and would require more manpower and effort,not only to planning,engineering,build and operations team but also to procurement and other parts of our organizations.ODiN project will be developing such models in the next Phase.The vendo
212、r selection process for functional RAN disaggregation should strongly focus on standard based compliance.New disaggregated RAN components such as RIC(Near-and Non-RT RIC)are being specified in industry forums such as the O-RAN Alliance(https:/www.o-ran.org/).The O-RAN Alliance also runs,interoperabi
213、lity tests and PlugFests which are being performed in open Testing and Integration Centres(OTIC).Version 2.0,20th September 2022 Page 36(101)Support wide adoptions of O-RAN specifications Organize and run PlugFest and proof of concepts Test RAN equipment based on O-RAN specifications Run test betwee
214、n different RAN vendors to verify interoperability Provide feedback to O-RAN community according to test results and potential interoperability gaps.For more details please refer to the corresponding O-RAN Alliance description(i.e.https:/www.o-ran.org/testing-integration)Here are a few important poi
215、nts to consider for operators when selecting a RIC:Strong standards compliance with O-RAN Openness(Open&standard APIs,Ability to work with any O-RAN compliant 3rd party systems(RAN NFs,x/rApps,SMOs,etc.)RAN vendor independence&continuous support/contribution to O-RAN toward fully interoperable RAN(e
216、.g.continuous extension of O-RAN E2SMs)RIC as a platform to enable RAN innovation Flexibility and support for x/rApps developers to develop new use-cases(both SDK and API based)Portfolio of x/rApps,and ability to develop more Support for in-house and 3rd party x/rApps RIC as a platform for AI/ML-dri
217、ven RAN Proven interop in plugfests,PoCs,etc.Engagement in the O-RAN community It is also worth to mention that a simple DIY(Do it yourself)or non-standard RIC platform,may come with the promise of an easy and non-complicated start,which may even promise a quick and easy win.However,since the main g
218、oal of the RIC platform is to leverage the innovation power of an entire industry eco-system,it will be important to consider Version 2.0,20th September 2022 Page 37(101)standardization and world market acceptance over implementation speed to leverage the innovation power of multiple industry partne
219、rs.The vendor selection process for hardware and software disaggregated IP&Optical Transport Networks should follow a pragmatic approach,which systems offers the best solution for the operators preferences.HW&SW disaggregated solutions are functional comparable with non-disaggregated solutions,there
220、fore,they can be easily compared on a functional and commercial basis.However,if new aspects such as cloudification,the introduction of containerized networking functions,are being introduced,this will also influence other operational aspects which will require more complex commercial TCO evaluation
221、s.Lets have a look on the example at a typical Cell-Site router:One option is a non-disaggregated legacy Cell Site Router which is provided by a single vendor with a specific set of required Network functions(i.e.IP/MPLS,SR/MPLS,SRv6,etc.).The same functions could be also provided by a disaggregated
222、 Cell-Site Router,where the dNOS is provided by vendor A,the required hardware by Vendor B and the integration of HW&SW by a specific System Integrator(S.I.).Since the required networking functions are the same,a simple commercial and operational comparison of the disaggregated and non-disaggregated
223、 solution would be sufficient.The picture is changing in case the Cell-Site Router function is being performed by containerized Software running on COTS based Hardware platform,which is at the same time being used to host other functions,such as CU or DU.In this case,the comparison becomes more comp
224、lex and requires the consideration of multiple factors such as:Reduced Set of hardware Reduced Maintenance Cost Reduced Truck Roll costs during Roll-out and Hardware repair.Leveraging Cloud native tools such as CI/CD pipelines,which are already being used for other cloud based solution components.HW
225、&SW disaggregation in the IP&Optical Transport domain,is not the primary goal by itself,the various pros and cons of disaggregation need to be evaluated against operator specific preferences.Version 2.0,20th September 2022 Page 38(101)4.3.2 Core Disaggregation Similar to the RAN domain,5G has brough
226、t a tremendous change to the Core Network domain too:the move towards a Service-Based Architecture(SBA)and leveraging numerous cloud principles.This actually came ahead of RAN disaggregation and now RAN is following a similar journey.SBA essentially breaks up the static one-to-one relationship betwe
227、en two functions,allowing any service to be consumed by any other function.This is achieved through the unification of the communication,i.e.utilising HTTP/2 with JSON-encoded payloads.Furthermore,compared to 4Gs Evolved Packet Core(EPC),3GPP has defined a much larger set of Network Functions for 5G
228、 which have a much smaller scope of functionality.This follows the idea of a standardised methodology to disintegrate the 4G EPC with each 5G Core Network Function offering a Service-Based Interface(SBI)which 3GPP defines.The reason for such effort is driven by:a.The desire to enable multi-vendor de
229、ployments where a 5GC is not necessarily composed of a single vendor solution but offered by more than one vendor based on the operators needs.Ultimately,it should be the operators(or verticals)choice of required functionality that dictates which Network Function is acquired from which vendor b.Adop
230、tion of cloud principles(cloud-native procedures,DevOps workflows for implementing 5GCs,and utilising microservice-targeted software design patterns)is the second key driver behind the shift towards SBA allowing 5GC vendors to scale their software solutions similar to cloud solution providers(i.e.a
231、service can scale on demand).However,not all 5GC Network Functions utilise SBA principles(in particular SBIs)and there is further standardisation work required to fully arrive at a disintegrated 5G Core Network.c.The softwarisation of Network Functions combined with cloud-native workflows also elimi
232、nates the necessity for customised hardware and a tightly interlinked software development process.With virtualisation technologies heavily adopted in the cloud domain,e.g.Linux Containers,Docker or Kernel Virtual Machines,the separation of hardware and software is key towards fully(vertically)disin
233、tegrated 5G Cores.4.3.2.1 Core Disaggregation Activities To comprehend the missing pieces in the Core Network to enable a fully SBA-driven 5G Core,Figure 4-6 illustrates the current(Release 17)5G system architecture with blue interface lines indicating the availability of a Service-Based Interface a
234、nd green lines the existence of a Non-SBI.Note,it is only N4 that is the last remaining Non-SBI that interconnects 5GC Networking Functions,i.e.the SMF and UPF.Version 2.0,20th September 2022 Page 39(101)Figure 4-6:5G Core Network System Architecture with Emphasis on Service-Based(blue)and Non-Servi
235、ce-Based(green)Interfaces 4 Approved Study Item in 3GPP 4 focuses on enhancing N4 to support event exposures and real-time service flow information(for Network Data Analytics Function(NWDAF)purposes)marking a small step towards the N4-to-Nupf transition.Equally important,but with much larger impacts
236、,is the N1 interface allowing UEs to communicate to the 5G Core via Non-Access Stratum(NAS)procedures.This part of the control plane remains untouched so far and poses a challenge to the Access and Mobility Management Function(AMF)as the single point of entry into an SBI-enabled 5G Core.The AMF esse
237、ntially operates in two separate worlds(SBA and non-SBA)imposing challenges to 5GC vendors who have adopted microservice software architectures for their products(more information on that in Section 5).And the N2 interface between the Access Network(AN)and the AMF(as the underlying protocol stack fo
238、r N1 communications)plays a significant role in the complexities AMF vendors face when implementing the AMF as a Cloud-Native Network Function(CNF).4.3.2.2 Disaggregation Impacts on Core Operational Process This section describes the impact of disintegration on 5GCs with regards to planning,deployme
239、nt,maintenance,optimisation,and provisioning.Impact on Core Planning Process The planning of resources(compute,storage,networking)has changed in a way that the system can be scaled up and down based on demand and a 5GC NF can exist as a set of instances across different locations.While the input to
240、such planning is still the number of users that are expected to connect in a given amount of time,potential fail over scenarios or Version 2.0,20th September 2022 Page 40(101)system upgrades through DevOps procedures impacts the calculation and differs from pre-5G Core planning procedures.The flexib
241、ility of disintegrated 5GCs also allows operators to plan for a multi-vendor 5G Core where NF1(e.g.AMF)comes from Vendor 1 and NF2(e.g.NWDAF)from Vendor 2.This decision is driven by the NF functionalities offered by a specific vendor.Impact on Core Deployment Process Disaggregation has allowed flexi
242、ble deployment of Core Networks.Since Core Network software and functionalities are now independent of hardware and is now getting more cloud native,core functions can now be installed or hosted either in a private cloud/premise or in a public cloud that is not in the premise of the operators.Hostin
243、g the Core functions in the cloud potentially reduce CAPEX and convert spend to OPEX.The adoption of cloud principles and the realisation of 5GC NFs using a microservice-based software architecture allows the deployment process to be automated using cloud-native procedures.Using orchestration framew
244、orks that focus on container-based service provisioning,the deployment process is a defined through a workflow identical to the one of cloud service providers.In that workflow a descriptor is defined which declares the required microservices(container names/packages)in use and their properties(CPU,R
245、AM,Storage)combined with policies how the orchestrator should react to a change in load.These deployment procedures also cover scenarios of hardware failures or system failures increasing the capability to react to system changes in a robust and automated fashion.Furthermore,this also enables the de
246、ployment of 5GC Network Functions implemented by multiple vendors or the choice to only deploy the 5GC Network Functions required for a specific network.For instance,in a Private Network setting only a handful of Network Functions could be required to provide the capabilities needed.5GC as a service
247、 is also now possible wherein the 5GC functions could be outsourced.Impact on Core Maintenance Process DevOps has become the norm to realise a continuous development and integration of software without the need to bring down an entire service.Tightly linked with cloud-native orchestration workflow,m
248、icroservice-based software realisations can be upgraded for a subset of service requests allowing to observe whether the upgrade causes unexpected service behaviour.Such maintenance workflow has been only possible due to the disintegration of the 5GC and the creation of SBIs.Furthermore,if vendors c
249、hoose to adopt a microservice software architecture for the realisation of their NF,issues in maintenance are Version 2.0,20th September 2022 Page 41(101)always limited to the scope of the microservice instead of the entire NF.The same applies to each NF,as they are separated through standardised SB
250、Is.Hosting Core NFs in the public cloud will help reduce the efforts of operators in maintaining the network as many of the responsibilities,especially with the hardware,are now passed on to the cloud provider.This will help ease operations and maintenance.Impact on Core Optimisation Process The dis
251、integration of 5GCs enabled vendors to optimise their NF to a greater extent,allowing operators to pick a specific NF from a specific vendor,if desired.In particular for user plane specific QoS requirements around optimised latencies,local breakout or customised capacities,the disintegrated 5GC allo
252、ws fine-tuned UPF realisations.For instance,if a Private Network owner or operator of a Public Network aims to deploy 5GLAN with support for Time Sensitive Networking,only the SMF and UPF must support such feature,while preserving any other 5GC NF required to operate a fully-fledged 5GC.Another exam
253、ple of how important the disintegration enabled optimised deployments could be is to consider a manufacturer utilising 5G for the communication technology of their robots:in such scenario billing and mobility is not required and can be removed as functionality from the 5GC without affecting the oper
254、ations of other NFs.Impact on Core Service Provisioning Process When considering the 5GC as a service,the disintegration of 5GCs enables never before seen provisioning possibilities due to the flexibility the system architecture permits.In particular for verticals,there is a range of questions that
255、must be evaluated related to the service provisioning and is commonly discussed under Public Network vs Non-Public Network service provisioning concepts:If coverage is not an issue,verticals may choose a network slice in an operators public network offering including a defined set of NF instances ex
256、clusively available to handle control plane communication of the verticals UEs Alternatively,verticals may deploy their own gNB on premise and connect it to a 5GC deployed in a cloud or to the 5GC of an operator(where the operator also provided the gNB).Version 2.0,20th September 2022 Page 42(101)To
257、 demonstrate even greater flexibility,the vertical could choose a sub-set of 5GC NFs to be owned and deployed locally(e.g.UPF and UDM for performance and data protection purposes),while utilising the remaining necessary 5GC NFs from a third party in a public cloud or from an operator.As the entire 5
258、GC is pure software without any hardware dependencies,the service provisioning may be compared to typical cloud offerings.4.3.3 Transport disaggregation In the fast-changing digital transformation,we see an agile network Evolution of the end-to-end Communication Service Provider.Weve seen how the tr
259、ansport network domain continues to evolve as it adapts to the new services and applications requiring low latency and massive bandwidth.As the transport network evolves,we need a separation of different functional components that is fulfilled by network disaggregation.We have seen giant hyper scale
260、rs such as Google,Facebook,Amazon,and Microsoft implement the first large-scale disaggregation of network and software in the data center and wide area network deployment that drives service innovation and market differentiation.The disaggregation of hardware and software is the critical enabler for
261、 deploying each functional component of the network device.Each element delivers specific roles and functions that separate the control and user-plane traffic.Transport disaggregation is in all forms of the transport networks such as Ethernet,Optical and IP.Its an open networking device consisting o
262、f IP routers,optical systems build on open APIs for software-defined networking(SDN).4.3.3.1 Transport disaggregation activities Telecommunication providers adapt to the emerging hardware and software separation ecosystem in their domains such as IP core,transport,and access networks.Standards Devel
263、opment Organizations(SDO)create standards for the Disaggregated Transport Networks.See different organizations below.Telecom Infra Project Internet Engineering Task Force(IETF)Broadband Forum(BBF)Version 2.0,20th September 2022 Page 43(101)Open Networking Foundation(ONF)Open Compute Project(OCP)Opti
264、cal Internetworking Forum(OIF)Disaggregation in Transport Network ONF ONFs target is to bring the approach of open networking to the optical network layer.The initial phase of Open Disaggregated Transport Network(ODTN)project will disaggregate transponders from open line systems to enable data centr
265、e interconnection to evolve at the speed of transponder improvement.The ODTN project is an ONF operator-led initiative to build data centre interconnects that will use disaggregated optical equipment,open and common standards,and open-source software.The objective is to drive innovation by disaggreg
266、ating the components of the network and provide open software to control a multi-vendor assembly of components.ODTN will enable a white-box optical peripherals ecosystem allowing multiple components to be combined and built into a complete solution.Vendors can then focus on building a specific compo
267、nent(for example,transponder)without the need to build a complete solution which leads to accelerated innovation and lower costs.This will allow operators to integrate the latest technologies once they become available rather than waiting for them through the previous siloed method.ODTN Phase 1 will
268、 focus on point-to-point data center interconnection.The open-source network controller controls the network infrastructure with well-defined open transport APIs,which allows a mix of paired transponders from different vendors running on the same physical links.We can see the role of SDN technology
269、in the Telecommunications ecosystem of Communication Service Providers,and it is the critical enabler for transport automation and disaggregation.The Transport domain is evolving into disaggregated hardware and software parallel to the Access and Core domain.Like the ODTN initiatives of ONF,The Tele
270、com Infra Project(TIP)has the Optical&Packet Transport Project(OOPT)with the common goal of disaggregation from software and hardware.Version 2.0,20th September 2022 Page 44(101)The open-source community is driving the innovations for Disaggregated network operating system.It complements our transpo
271、rt domain to adapt to the open networking wave that is fast evolving in the disaggregated transport network.Many hyperscalers and tech giants like Google,Amazon,and Facebook run their network operating systems using commodity networking switches which helps drive open networking innovations.Decoupli
272、ng hardware from the Network Operating System(NOS)to allow a more diverse eco-system is introducing new opportunities but also new operational challenges.In general,under the term Transport different device classes operating at the OSI-Layer 1-4 are being summarised to one solution domain.According
273、to operator preferences,Optical Transport systems as well as Ethernet Switching and IP-Routing devices are being used in this domain.Before we elaborate on the operational challenges it is also worth to mention,that meanwhile the next iteration of Transport disaggregation is taking place and is intr
274、oducing stronger cloudification aspects.One good example is the OCP(Open Compute Platform)industry standardisation group,which is hosting the Software for Open Networking in the Cloud(SONiC)Opensource initiative,which is not only disaggregating the NOS from the Hardware,here also the NOS itself is d
275、ivided into multiple functional parts in order allow Network functions to run as a containerized functions on top of a Linux based operating system.This will allow a more cloudified approach with the usage of standard Cloud technologies and their widely used tool chain.Another example,which goes int
276、o the similar direction,is to take those containerized network functions(i.e.dynamic Routing Stack)and to run them on standard X.86 COTS Server which are using as well a LINUX based operation system.With this,it now becomes possible to run the identical network functions on either networking special
277、ized white-boxes or on COTS standard servers.An example of these multiple disaggregation steps is illustrated in the following diagram,which shows the multiple steps from simple disaggregation towards cloudification on the example of a Cell-Site-Router.In the figure 4-7,the Cell-Site Router is takin
278、g the transport function to connect the Cell-Site with the aggregation network,while the first approach of disaggregation just focused on separating the underlying Hardware from the“monolithic NOS.This allows the flexibility to pick and choose Hardware and Software from different suppliers(dNOS).Thi
279、s flexibility comes with the cost of re-integrating it into a single working system.The next step of Version 2.0,20th September 2022 Page 45(101)cloudification shows that the transport function becomes now just a containerized Software,which can also run on the already existent Hardware to power CU/
280、DU functions on the corresponding sites.Figure 4-7:Disaggregated Network Operating System With this additional step of cloudification,new operational models which follow cloud principles,will be required.Details of cloudification in the Transport domain will be described in the following chapter of
281、cloudification.4.3.3.2 Disaggregation impacts on Transport operational processes Impact on Transport Planning process Disaggregated Transport is introducing a new architecture and will require different methods of planning and dimensioning.In the traditional architecture,Transport dimensioning mostl
282、y focuses on the capacity planning of the metro aggregation Network,which consist typically of optical Transport devices and/or Routing and Switching devices for Cell-Sites and aggregation locations(sites).Since disaggregation does introduce more potential variables in terms of which hardware and wh
283、ich software can be combined with each other and in general as a larger eco-system of suppliers will be able to provide their solutions,this will also impact the planning process.Potentially,also the role of a system integrator needs to be considered and responsibilities need to be clearly divided b
284、etween the systems integrator and MNO.Once the desired combination of hardware and software has been selected,the planning process Version 2.0,20th September 2022 Page 46(101)remains largely the same as with non-disaggregated solutions.This will change,if cloudification also comes into the picture.C
285、loudification is covered in the next chapter.Impact on Transport Deployment process It can safely be assumed that the deployment process of the disaggregated Transport,with its disaggregated hardware-&software,remains largely the same as with the non-disaggregated legacy approach.This is due to the
286、fact that the re-integration of the separated HW&SW packages takes place before the actual deployment is going to happen,therefore a plain disaggregated Transport solution will behave very much like the legacy Transport solutions.Again,the major change will come with the cloudification of the transp
287、ort.Impact on Transport Maintenance process In the disaggregated model,as the Transport is broken down to more parts horizontally and vertically,Transport maintenance will become more tedious as more expertise and tools are needed for the operations or field teams.Disaggregation will cause operators
288、 to use more vendors therefore requiring operations engineers and staff to have more knowledge and skills both in relation to the hardware and the software of each vendor.This impact would have to be managed properly to ensure cost and organisational effects will be at an acceptable level.However,he
289、re again in the model of cloudified disaggregated Transport the picture is changing.The maintenance process of hardware will be aligned with the RAN and cloud infrastructure.The maintenance of the Transport itself,will be reduced to primarily software aspects.This step will offer massive operational
290、 benefits in this domain Impact on Transport Optimisation process Identical to the previous chapter,since hardware and software is separated during disaggregation,optimisation will be done separately also with hardware and software given that the two are potentially provided by multiple different su
291、ppliers.The optimisation team would need to build expertise on hardware and software separately.Separation of functions in disaggregation will also impact the operations team as they need to build expertise in all the new vendors that will be integrated in the network to ensure all issues will be ad
292、dressed or resolved.Impact on Transport Service provisioning process Due to the fact that in the disaggregated transport re-integration(of separated HW and SW packages)will take place prior to deployment,the service provisioning process for Version 2.0,20th September 2022 Page 47(101)disaggregated T
293、ransport remains largely unchanged in comparison with previous“integrated”models.4.3.4 Non domain specific Operational Activities 4.3.4.1 Integration of parts There are 2 options for each operator to address integration of different parts and vendors to have a complete solution.One is to develop an
294、internal team that will be responsible for integrating the solutions of all the participating vendors to build the full set-up.The other is to outsource or hire an external entity that will be tasked to do all the integration works and services.That same entity will be the one responsible for any fa
295、ult in the system,be it hardware or software for any part-CU,DU,RU as well for for the RAN-Transport systems and Core which might come disaggregated as well,as described in previous sections.The first one is quite tedious and costly because it will require hiring new resources that should stay for a
296、 long time,training them and building their competencies.Rather than making the organization lean,it requires now to beef up in order to support the disaggregated architecture.The second option might be easier and simpler although it will be very costly and would require governance 4.3.4.2 Interoper
297、ability and Compatibility Assurance As outlined in other chapters,interoperability is a huge challenge.Horizontal interoperability,i.e.between functions building up the network end-to-end remains business as usual.In addition,disaggregation brings the need for vertical interoperability.This is addre
298、ssed by integration and testing.Due to the increasing number of tests,this calls for automation,typically in a CI/CD(Continuous Integration/Continuous Delivery)pipeline.The industry is organising activities to achieve interoperability,by initiatives such as TIP international labs and testing efforts
299、 as well as cooperation being done between operators and vendors.Some hardware vendors are also doing certifications with software suppliers to ensure their hardware are always updated and are always capable of supporting any new functions and features that comes with the new software release.This w
300、ill help a lot,but then again this will not completely assure compatibility and interoperability.Telecom Infra Project(TIP)has 14 labs (https:/ by individual TIP participant companies that test interopeability.There are three types of labs:Version 2.0,20th September 2022 Page 48(101)TIP product labs
301、-focused on Proof of Concepts(PoCs).TIP integration labs-focused on end to end testing to evaluate a products maturity toward commercial readiness.TIP deployment labs-focused on people(education and technology spread out),processes and tools.TIP certification and badging(https:/ Badges and Ribbons a
302、re awarded depending on the level of maturity of products and solutions against the technical requirements that is evaluated.Only those products with market availability are qualified to be evaluated.Awarded products and solutions are then listed to the TIP Exchange(https:/ the corresponding awarded
303、 badges.Aside from badges,products on TIP Exchange may also be awarded ribbons.Badges:Supplier Validated Product(Bronze)This is being awarded to the products that technology suppliers have tested in their own laboratories,This is primarily applicable to individual network products and components.Pro
304、ducts are required to be commercially available even on its early stages TIP Validated Product(Silver)This badge is awarded to integrated network layers.It is awarded to the products that were validated in a TIP Community laboratory or it could also be validated by an approved 3rd party laboratory.P
305、roducts has to be commercially available with the minimum product support in order to qualify for this badge.TIP Validated Solution(Gold)Gold badge is awarded to products that were validated in a TIP Community laboratory or it could also be from an approved 3rd party party laboratory.This is primari
306、ly applicable to end-to-end solutions,integrated network layers,or it could be individual products tested in an end-to-end environment that is representative of actual service provider conditions.Solutions are required to be commercially available with full product support to qualify Version 2.0,20t
307、h September 2022 Page 49(101)Ribbons:Operator tested Ribbons Ribbons are awarded to listed products on TIP Exchange that were tested in an actual field trial conducted by an operator.Requirements Compliant This is the minimum requirement to get listed to the TIP Exchange.Individual network component
308、s or products should be compliant to the requirements that was set by the associated project group.For a more detailed information,it is advised to visit the TIP website(https:/ Joint DevSecOps Pipeline Disaggregation is very different from the development model of existing networks.In order to ensu
309、re that the network quality,customer experience,implementation time,optimization and maintenance will not be affected by technological changes,the network operation mode also needs to be changed and new processes,skills and tools needs to be adopted.Operators will face challenges at two levels.First
310、,in terms of technical complexity,disaggregation increases the complexity of integration testing,compatibility,and security prevention and control.Second,in terms of organization and process,disaggregation will have a significant impact on procurement,integration,and operations.Therefore,DevSecOps i
311、s a necessary means to comply with the trend of Disaggregation.The introduction of DevSecOps will help operators improve efficiency and reduce costs.On the one hand,cross-organizational DevSecOps application scenarios are required to solve the contradiction between the complexity of network evolutio
312、n and network management efficiency;On the other hand,it is necessary to introduce the cross-organization standard pipeline general solution R&D tools into the testing and certification system to solve the problem that the test environment adapts to the frequent upgrade of the existing network equip
313、ment.4.3.4.4 Upgrades and expansions New education must be done in doing upgrades as it is not the same anymore as the way we do it with traditional solutions.It is not simply upgrading the hardware and buying software and licenses.Now,the number of processors must be accounted and dimensioned Versi
314、on 2.0,20th September 2022 Page 50(101)accordingly.It is now like buying computer for your home and ensuring you bought the correct processor specification and quantity,memory and storage that will suffice for your everyday needs.So our traffic requirements and projections should be translated to ha
315、rdware requirements up to the chip level.4.3.4.5 Troubleshooting and disaster recovery Once we implement disaggregation,as complex as it may be,we are compelled to have a troubleshooting and disaster recovery plan.Troubleshooting is included in the main course and action of operations and daily acti
316、vities.This is why troubleshooting for every group of hardware and software should be planned well.There should be enough manpower to do that and they should be equipped with the right knowledge and tools.Each of the brands,hardware type and software type might need different types of tools.If there
317、 is a single or universal one that could support all types of hardware and software,then that would be the most ideal.Otherwise,there should be proper grouping and minimization of tools used so there is not too many that the troubleshooters need to learn.This will avoid confusion which could potenti
318、ally lead to errors later.Due also to the new architecture and splits,a new resilience plan would have to be developed.This is especially true if CUs and DUs are centralized.Since more sites will be home to a single DU and/or CU,there has to be a secondary homing plan to a different DU and/or CU tha
319、t could support transferred traffic in cases of disaster or any down time.Nowadays,we experience more and more devastations from storm,earthquakes,volcanic eruptions.Therefore,it is a necessity to have a very strong disaster recovery plan that could minimize the service down time no matter how bad t
320、he situation is.Operations team should know each and every hardware and software by heart so the decision-making during disasters would be fast and easy.Fast decision making and correct judgement are the two important things during disaster recovery.Our teams should be equipped in ensuring these two
321、 are achieved even when the architecture of the network has significantly changed.4.3.5 Support for Green Technology In all the improvements being done to the network ensuring flexibility and efficiency,minimization of carbon emissions and energy efficiency should always be considered as these are t
322、he key factors to sustainability.To enable MNOs and the wider mobile and IT industry to meet their sustainability goals,the design,manufacture,deployment and operation of Version 2.0,20th September 2022 Page 51(101)disaggregated networks will need to go hand in hand with the adoption and advancement
323、 of new green technologies and processes.NGMN Alliance,through its Green Future Networks strategic programme is providing industry leadership in this area and has already published a number of white papers outlining the overall challenges and opportunities 5 as well as addressing how to make network
324、s more energy efficient 6.Further work in this programme is expected to address wider issues related to how MNOs and the entire mobile industry value chain can reduce carbon emissions;use advanced technologies and processes to further improve network energy efficiency;and provide guidance on how the
325、 industry can reduce its overall environmental impact.The impact of network disaggregation on the energy efficiency and carbon emissions of the network requires further study.At a high-level,the key issue is the extent to which energy efficiency gains from pooling lots of different network function
326、compute tasks in the cloud(where they are run on general purpose compute resources)are offset by the energy efficiency losses in moving these tasks from highly optimised(and presumably energy efficient)dedicated compute resources(often using systems on chips designed for the specific task).Although
327、it is anticipated that future phases of NGMNs Mastering the Route to Disaggregation programme will address specific disaggregation challenges in relation to Green Future Networks it is worthwhile pointing out two existing industry initiatives that address data centre energy efficiency.These are:Redf
328、ish,a REST API used for platform management and standardized by the Distributed Management Task Force,Inc.7 Scaphandre,an open-source metrology agent that can be deployed on a CaaS platform(Kubernetes)to collect power metrics related to the overall cluster and the individual CNFs running on it.8 4.3
329、.6 Integration to End-to-End Service Orchestration and Common Management and Service Assurance System While parts are increasing,and network functions are being distributed to more nodes,we need to ensure that all network elements are connected and covered by the overall management and service assur
330、ance system.This will ensure proper monitoring,alarm management,troubleshooting and control.This will also ensure proper operation of all elements as well as overall visibility ensuring expected and target availability.Version 2.0,20th September 2022 Page 52(101)All functions should also be connecte
331、d to and covered by End-to-End Service Orchestration to ensure all parts of the network are participating in the automation process and services are instantiated and defined in all parts of the network RAN,Core,Transport,billing,etc.Version 2.0,20th September 2022 Page 53(101)Cloudification is the n
332、ext step in the evolution of a disaggregated network.This evolution is not only on network functions design and implementation,Cloudification also brings new tools supporting automation and orchestration that have a relevant impact on operation.NGMN published in 2021 an extensive study on the target
333、 picture provided by Network Disaggregation and Cloudification for the overall Telco Platform 9.This chapter summarises the main aspects on the Cloudification of a disaggregated network and suggests possible impacts/challenges on operation.5.1 The overall Blueprint 5G is designed with open interfaces and a service-based architecture allowing services to be delivered via a network that is disaggreg