ETSI GR NGP 006-2018 Next Generation Protocols (NGP) Intelligence-Defined Network (IDN) (V1 1 1).pdf

1、 ETSI GR NGP 006 V1.1.1 (2018-06) Next Generation Protocols (NGP); Intelligence-Defined Network (IDN) Disclaimer The present document has been produced and approved by the Next Generation Protocols (NGP) ETSI Industry Specification Group (ISG) and represents the views of those members who participat

2、ed in this ISG. It does not necessarily represent the views of the entire ETSI membership. GROUP REPORT ETSI ETSI GR NGP 006 V1.1.1 (2018-06) 2 Reference DGR/NGP-006 Keywords framework, next generation protocol ETSI 650 Route des Lucioles F-06921 Sophia Antipolis Cedex - FRANCE Tel.: +33 4 92 94 42

3、00 Fax: +33 4 93 65 47 16 Siret N 348 623 562 00017 - NAF 742 C Association but non lucratif enregistre la Sous-Prfecture de Grasse (06) N 7803/88 Important notice The present document can be downloaded from: http:/www.etsi.org/standards-search The present document may be made available in electroni

4、c versions and/or in print. The content of any electronic and/or print versions of the present document shall not be modified without the prior written authorization of ETSI. In case of any existing or perceived difference in contents between such versions and/or in print, the only prevailing docume

5、nt is the print of the Portable Document Format (PDF) version kept on a specific network drive within ETSI Secretariat. Users of the present document should be aware that the document may be subject to revision or change of status. Information on the current status of this and other ETSI documents i

6、s available at https:/portal.etsi.org/TB/ETSIDeliverableStatus.aspx If you find errors in the present document, please send your comment to one of the following services: https:/portal.etsi.org/People/CommiteeSupportStaff.aspx Copyright Notification No part may be reproduced or utilized in any form

7、or by any means, electronic or mechanical, including photocopying and microfilm except as authorized by written permission of ETSI. The content of the PDF version shall not be modified without the written authorization of ETSI. The copyright and the foregoing restriction extend to reproduction in al

8、l media. ETSI 2018. All rights reserved. DECTTM, PLUGTESTSTM, UMTSTMand the ETSI logo are trademarks of ETSI registered for the benefit of its Members. 3GPPTM and LTETMare trademarks of ETSI registered for the benefit of its Members and of the 3GPP Organizational Partners. oneM2M logo is protected f

9、or the benefit of its Members. GSMand the GSM logo are trademarks registered and owned by the GSM Association. ETSI ETSI GR NGP 006 V1.1.1 (2018-06) 3 Contents Intellectual Property Rights 4g3Foreword . 4g3Modal verbs terminology 4g31 Scope 5g32 References 5g32.1 Normative references . 5g32.2 Inform

10、ative references 5g33 Abbreviations . 6g34 Overview 6g35 Background 6g35.1 Continuous Evolution of Network 6g35.2 Functional and Systemic Requirement . 7g35.3 Rapid Development of Machine Learning Technologies . 8g36 Benefits of Introducing AI into Network . 9g36.1 Towards Fully Autonomic Network .

11、9g36.2 Response to the challenge of complexity . 9g36.3 Response to the challenge of variation . 10g36.4 Insights of the Network and Improve the Utilization . 10g36.5 To Be Predictive . 11g36.6 Potential Decision Efficiency . 12g36.7 Potential Business Model . 12g37 Design Goals of IDN 12g37.1 Goal

12、of IDN 12g37.2 Deployment models: Centralized, distributed and Hybrid . 13g37.3 Wired and wireless consideration . 14g37.4 Security and Privacy Considerations 16g37.5 Multi-objectives Resolution . 17g38 The proposed IDN Architecture . 17g38.1 Reference Architecture . 17g38.2 Comparing System design

13、21g38.2.1 Overview 21g38.2.2 Distributed Architecture . 23g38.2.3 Centralized Architecture . 23g38.2.4 Hybrid Architecture 24g38.3 Controlling Loop 25g38.3.1 AI-Enhanced Close Loop 25g38.3.2 AI-Enhanced Open Loop 27g38.3.3 Traditional Loop . 29g38.3.4 Internal Loop 29g38.3.5 UNI Loop 29g38.4 Core Su

14、pport Technologies 29g38.4.1 Network modelling . 29g38.4.2 Measurement and Data Orchestration . 30g39 Potential Standardization Works 31g39.1 Overview 31g39.2 Measurement 32g39.3 Data Centric standards 32g39.4 Control Centric standards . 33g3Annex A: Authors Essential, or potentially Essential, IPRs

15、 notified to ETSI in respect of ETSI standards“, which is available from the ETSI Secretariat. Latest updates are available on the ETSI Web server (https:/ipr.etsi.org/). Pursuant to the ETSI IPR Policy, no investigation, including IPR searches, has been carried out by ETSI. No guarantee can be give

16、n as to the existence of other IPRs not referenced in ETSI SR 000 314 (or the updates on the ETSI Web server) which are, or may be, or may become, essential to the present document. Trademarks The present document may include trademarks and/or tradenames which are asserted and/or registered by their

17、 owners. ETSI claims no ownership of these except for any which are indicated as being the property of ETSI, and conveys no right to use or reproduce any trademark and/or tradename. Mention of those trademarks in the present document does not constitute an endorsement by ETSI of products, services o

18、r organizations associated with those trademarks. Foreword This Group Report (GR) has been produced by ETSI Industry Specification Group (ISG) Next Generation Protocols (NGP). Modal verbs terminology In the present document “should“, “should not“, “may“, “need not“, “will“, “will not“, “can“ and “ca

19、nnot“ are to be interpreted as described in clause 3.2 of the ETSI Drafting Rules (Verbal forms for the expression of provisions). “must“ and “must not“ are NOT allowed in ETSI deliverables except when used in direct citation. ETSI ETSI GR NGP 006 V1.1.1 (2018-06) 5 1 Scope The scope of the present

20、document is to specify the self-organizing control and management planes for the Next Generation Protocols (NGP), Industry Specific Group (ISG). 2 References 2.1 Normative references Normative references are not applicable in the present document. 2.2 Informative references References are either spe

21、cific (identified by date of publication and/or edition number or version number) or non-specific. For specific references, only the cited version applies. For non-specific references, the latest version of the referenced document (including any amendments) applies. NOTE: While any hyperlinks includ

22、ed in this clause were valid at the time of publication, ETSI cannot guarantee their long term validity. The following referenced documents are not necessary for the application of the present document but they assist the user with regard to a particular subject area. i.1 https:/ i.2 https:/www.thes

23、un.co.uk/tech/4141624/facebook-robots-speak-in-their-own-language/. i.3 Reed S, Akata Z, Yan X, et al.: “Generative adversarial text to image synthesis“, in ICML 2016. i.4 Oord A, Dieleman S, Zen H, et al.: “Wavenet: A generative model for raw audio“, arXiv:1609.03499, 2016. i.5 LeCun, Yann, Yoshua

24、Bengio, and Geoffrey Hinton: “Deep learning“, in Nature 521.7553 (2015): 436-444. i.6 Kingma D P, Welling M.: “Auto-encoding variational bayes“, in ICLR 2014. i.7 Goodfellow, Ian, et al.: “Generative adversarial nets“, in NIPS 2014. i.8 Cisco White Paper. NOTE: Available at https:/ i.9 https:/arxiv.

25、org/abs/1701.07274. i.10 ETSI TR 121 905: “Digital cellular telecommunications system (Phase 2+) (GSM); Universal Mobile Telecommunications System (UMTS); LTE; Vocabulary for 3GPP Specifications (3GPP TR 21.905)“. i.11 ETSI TS 136 401: “LTE; Evolved Universal Terrestrial Radio Access Network (E-UTRA

26、N); Architecture description (3GPP TS 36.401)“. ETSI ETSI GR NGP 006 V1.1.1 (2018-06) 6 3 Abbreviations For the purposes of the present document, the abbreviations given in ETSI TR 121 905 i.10 and the following apply to scenarios that include mobile network architectures: 3GPPTM 3rdGeneration Parti

27、cipation Project AI Artificial Intelligence DHCP Dynamic Host Configuration Protocol E-W East and West (direction) IDN Intelligence-Defined Network IETF Internet Engineering Task Force IP Internet Protocol ISG Industry Specific Group ML Machine Learning NE Network ElementNGP Next Generation Protocol

28、s NMS Network Management System N-S North and South (direction) OAM Operation And Management OSPF Open Shortest Path First QoE Quality of Experience 4 Overview The Next Generation Protocols (NGP), ISG aims to review the future landscape of Internet Protocols, identify and document future requirement

29、s and trigger follow up activities to drive a vision of a considerably more efficient Internet that is far more attentive to user demand and more responsive whether towards humans, machines or things. A measure of the success of NGP would be to remove historic sub-optimised IP protocol stacks and al

30、low all next generation networks to inter-work in a way that accelerates a post-2020 connected world unencumbered by past developments. The NGP ISG is foreseen as having a transitional nature that is a vehicle for the 5G community and other related communications markets to first gather their though

31、ts together and prepare the case for the Internet communitys engagement in a complementary and synchronised modernisation effort. Therefore NGP ISG aims to stimulate closer cooperation over standardisation efforts for generational changes in communications and networking technology. The present docu

32、ment focuses on proposing a new Intelligence-Defined Network (IDN) architecture and a gap analysis of current architectures. The intelligence technologies can learn from historical data, and make predictions or decisions, rather than following strictly predetermined rules. On one hand, the IDN can d

33、ynamically adapt to a changing situation and enhance its own intelligence with by learning from new data. On the other hand, IDN can also aim at supporting human-based decision by pre-processing data and rendering insights to users through advanced user interfaces and visualisations. The integration

34、 with various network infrastructures, such as SDN, NFV it provides data analysis/learning services to a variety of applications, and generate network-level policies to intervene the network to run as the applications expected. ETSI ETSI GR NGP 006 V1.1.1 (2018-06) 15 The core of the Network Brain i

35、s mostly composed by a variety of AI relevant technologies such as machine learning algorithms, data-mining algorithms, expert systems, etc. As described in clause 6.3, the Network Brain also represents an interface to applications so that developer could easily create specific tasks without handlin

36、g the data analysis/mining/learning by themselves. Middle layer: Network Orchestration and Controlling This layer is to interpret the network-level policies generated by the Network Brain into device-level policies/configurations and deliver them to corresponding devices. SDN controller is an instan

37、ce of the middle layer. Based on the Network Brain generated policies, this layer could make some simple decision by itself so that a quick control loop could be formed to control the devices behaviour in a much more efficient way. Under layer: Network Devices There are different kinds of network de

38、vices: NFV infrastructure, SDN-managed devices, Intelligent Devices (which could directly interact with the Network Brain), and Traditional Devices. As the Figure 3 shows, the wireless part of IDN is composed of a number of distinct mobile intelligent network decision entities. One centralized cross

39、-domain IDN decision entity sits in upper layer. Below it in the network layer reside many distributed decision entities. Each of these entities is made up of four key components: wireless data collection, analysis RAN internal functional split is still under discussion, such Functions similar to E-

40、UTRAN as listed in ETSI TS 136 401 i.11. 7.4 Security and Privacy Considerations When security relevant decisions are made based on the use of intelligent analytics or automated intelligent decision making, care should be taken to understand the new security challenges. When, for example, more intel

41、ligent decisions are enabled through the collection of ever more data, it needs to be analysed how that potentially enables attackers to easier feed data that derails the intelligent system ability to distinguish good from bad behaviour. As i.1 and i.2 reported, many companies have been puzzled by t

42、he unprecedented “attack“ that the legal and safe operations and data cause unexpected result. The traditional security problems are caused by the bugs in design or implementation. These two may become the typical examples of potential new security challenges. The future system should include the ab

43、ility that can deal with the following problems but not limited. Negative Data: the training system should have the ability to recognize data which may induce the system to become in an unexpected form on purpose. This problem will be serious in every data-fed system and it never happened in the pas

44、t design. Conflict Data: the training system should have the ability that can recognize the data which may cause confliction to the current known strategies or states. This problem may happen when user or device execute its private rules in a large and share area unintentionally or intentionally. Co

45、ntent data transmitted through network contains private data about users. Whereas AI techniques are powerful tools to automate network management functions, systematizing the large collection and processing of data presents some risk regarding the privacy of users. Indeed, predicting the location of

46、 a mobile users can clearly help in allocating resources in a RAN but it is also a privacy breaches. A future AI-based system should take in considerations the two following: Limitation of private data. Privacy can be leaked out from collected or post-processed data. Decisions algorithms may need it

47、. Thus, any algorithms should clearly specify the mention of information it uses to get as input/output and clearly limits according to it. Therefore, acquiring non used data should be excluded. As a result, this will prevent to gather potential private information which is not use and also improve

48、the scalability of the collection process. Such a limitation has to be also compatible with legislation. ETSI ETSI GR NGP 006 V1.1.1 (2018-06) 17 Data security. The IDN architecture should include the necessary mechanisms to avoid unfortunate data leakage. First, collected data should be accessible

49、by only specific algorithms that use them in an appropriate way. Second, collected data could have a maximum lifetime to be then discarded when being meaningless for further processing. Third, the IDN architecture should prevent any algorithms (for instance provided by a third party) aiming at discretely supporting private data exfil