2021-02-22 Multi-SDO Autonomous Networks Workshop #2 Meeting Notes

Not fully curated

Date

22 Feb 2021

Apologies

Input Material

Meeting Agenda

Call for IPR Declaration – IPR Statements for Use in Meetings

Recordings

Video:   https://tmf365-my.sharepoint.com/:v:/g/personal/dmilham_tmforum_org/EeYB-E4nKMROqdnwSNohHNcBlH-2c14Au55zR93uI_ZU5g?e=Gede3n   Password:  ANMulti%SDO2 Expies Aug 21, 2021

Audio:   https://tmf365-my.sharepoint.com/:u:/g/personal/dmilham_tmforum_org/EYNpWdamJYhKuiIZX84bq-oBAyZBVnZo1KE60DfdXAsRpg?e=rp5ONN  Password:  ANMulti%SDO2 Expies Aug 21, 2021

Chat Log:   https://tmf365-my.sharepoint.com/:t:/g/personal/dmilham_tmforum_org/EfC2weih379GubdfWhr-JVwBqGAtWlDAJEL9e3NeS0kgEw?e=Gq9lG5  Password:  ANMulti%SDO2 Expies Aug 21, 2021

Meeting notes-summary 

Introduction by Dave Milham

Many attendees were new and had not attended previous workshop ~1.

Presented the previous workshop minutes location - 2020-09-28 Multi-SDO Autonomous Network Coordination Workshop Meeting Notes/Video - AN-SDO Collaboration - TM Forum Confluence  - and key agreement at end of Workshop #1:

• "Start from the consensus and alignment of basic concepts, terminologies, framework, and key perspectives including definition of Autonomous Networks, closed loop, Autonomous Networks Levels, Autonomous domains, Intent driven interaction and open discussions.
• Further deep dive on the development of key technical mechanisms, framework and interfaces, as well as use cases/PoCs and testing and verification mechanisms to support autonomous networks."

Presentations:

• AN SDO Workshop Contributions - AN-SDO Collaboration - TM Forum Confluence

Feedback received on detailed  ETSI TC and IETF group titles.

Any correction should be email to [email protected]

Worksheet proposal - Kevin McDonnell

Kevin proposed a worksheet to establish which SDO and individuals were prepared to contribute to the topics agreed in Workshop #1.

Notion was to establish which groups had interests in which areas.

However was not possible to fully progress I the meeting due to a large amount of feedback from both those that attended Workshop #1 and those that had not.

Updated Spreadsheet above: Multi-SDO Autonomous Networks Meeting 2 MoM 22-FEB-2021.xlsx

Feedback Topics

A number of comments were made about the process of moving the work forward which differed from the output of the previous workshop.

These are summarised below and are not attributed to individuals or SDOs.  teh recording above give the chronological record 

Scope 

• Need to establish a common view of the objectives and benefits of such coordination.
• Query about objectives of the worksheet:
  • Confirm the topics identified from Workshop #1.
  • Who is interested in what topics? Frameworks specific meaning of definitions / terms
  • Who might be planning to contribute?
  • Alignment by sharing of work
  • Possibly jointly developed document.
  • Approach is to establish which groups have interest in which  areas.
• Noted that topics are similar and different groups may be working on similar concepts but at different levels of the network.
• General consensus to work out what to do together in concrete form

Concerns

• Concern if the work addresses standardisation as that will take much time.
• Joint specification would require discussions on rules and procedures.
  And is the purpose of the parallel Inter SDO meeting for later in March 2021.
• Over arching document might be feasible.  E.g. White paper
  • Could be definitions and or frameworks
  • Use case
  • Test and verification
• Different group will have more of less engagement and some topics might be of higher or lower priority.
• Autonomous levels is this separate from maturity levels? Consensus was yes.
• Conceptually alignment whilst necessary might still not create aligned outputs for example APIs models split between common /non domain specific and domain specific.
• Intent APIs are one example where divergence across SDO would be unhelpful.
• Concerns about establishing a mega organisation for create common framework
• Multiple things done in organisation trying to create consistency kwork in collaboration using Liaison offices asn liaisons
• Need avoid creating obstacles and impeding SDO progressing their works
• Idea of PoC looks good and was supported by multiple participant
• Concerns about setting a scope of work that is too large to be done within reasonable timeframe and effort.
• Starting point should be the existing SDO material and consider motivation and issues they plan to solve. E.g. whether Autonomous Network means the same in all organisations. Should try and find the commonalities  and how complementary work.

CSP Concerns

• Major Obstacle for CSP is to follow and analyse the activity of the SDOs organisation.
• Suggestion that CSP would benefit initially from a document setting out the  landscape of AN activities across SDOs to create a common understanding, more elaborated understanding of the terminologies (some are technical e.g. Autonomous Networks Autonomous Domains;  and some are types of deliverables) and get a picture of where the industry is going i.e. form of roadmap (attracted several supporting comments).

Technical Concerns

• Would be useful to know the current detailed work on the the list of topics where coordination could be useful.
• Common operational principle: Need to establish what is common across SDOs. And would need metamodels.  Reference t o IEEE Common overarching operational principles. Contributed by IEEE, AFI and TM Forum.
• Metamodels strong opinion were expressed both against and for Metamodels:

• • Against : too complex and create another model to be supported and of unknown value.
  • For: Suggestion to extract common model by extraction form multiple existing models with create criteria.
    Metamodels linking SDO Concepts is essential for full automation as envisaged by AN I.e. Zero-touch, Self-healing.

• Common information models:
  • Federation / merge of information models was successful for Network models as developed between 3GPP and TM Forum and common models lead by Cablelabs.
  • Concern about creating new over arching model as that creates N+1 standards.
  • Should try and establish what was done in the past that worked well.
• Concern that even aligning principles guidelines across SDOs could be a challenge as implies changes to existing SDO specifications.
• Formal contributions might be best done by company contributions as easier to handle in the SDO Working Groups.
• Outcomes could be:
  • sharing of information.
  • Sharing of definition then figure out if there a was mutual interest to create shared understanding which could then become the Landscape/ Roadmap.

Next meeting:

proposed to run in about 4 week at the time of 11:00-12:30 GMT/UTC
( Ed note proximity to Easter vacations needs to be checked.

We will ask attended to fill ourt the spreadsheet on line.

• Chat Log 
  • .

    11:05:51 From Kevin McDonnell : Minutes of Workshop #1 https://projects.tmforum.org/wiki/pages/viewpage.action?pageId=146292908
    11:11:47 From Nurit Sprecher : which WG in IETF? is it ANMIA?
    11:11:51 From Nurit Sprecher : ANIOMA?
    11:12:01 From Nurit Sprecher : ANIMA
    11:13:53 From Lingli CMCC : autonomic network
    11:17:30 From Nurit Sprecher : would be good to see companies but also orgs representing
    11:19:21 From Kevin McDonnell : https://docs.google.com/spreadsheets/d/1N7hSW9-cERX971_hxQeptA91zX50yKpMdL4fR94oClA/edit#gid=0
    11:19:38 From klaus Martiny (DTAG) : rh
    11:20:01 From Dave Milham : draft meting minutes whcih has as input the slides https://projects.tmforum.org/wiki/display/ASC/2021-02-22++Multi-SDO+Autonomous+Networks+Workshop+%232+Meeting+Notes
    11:31:51 From Dr.Muslim Elkotob : PoCs are not run exclusively by companies as was just commented; there are coordinated PoC activities in the TC INT PoC series based on GANA Autonomics in 5G and Beyond; and each of the events of the PoC series we run involves several stakeholders in the value chain at the same time (Vendors, ISVs, CSPs, etc.)
    11:36:37 From klaus Martiny (DTAG) : that is what I mean. clarification of the common ground and understanfing: e.g. close loops,ettc.
    11:38:17 From Nurit Sprecher : RH
    11:38:27 From Lingli CMCC : rh
    11:40:48 From Emmanuel (Huawei) : rh
    11:42:36 From klaus Martiny (DTAG) : agree with Nurit
    11:42:46 From Emmanuel (Huawei) : +1
    11:42:56 From Zou Lan (Huawei) : agree with Nurit
    11:43:06 From Anatoly (Nokia) : RH
    11:43:07 From Zou Lan (Huawei) : RH
    11:45:39 From klaus Martiny (DTAG) : indeed, that is the reason why there were a couple initiatives tried the hamonized the various prpjects
    11:47:43 From Nurit Sprecher : RH
    11:51:15 From rRanganai chaparadza : Towards the Standardization of an Overarching Blueprint of Common Operational Principles of Autonomic/Autonomous Networks
    11:52:05 From Laurent Ciavaglia (Nokia) to Dave Milham(Direct Message) : Hi Dave, It seems people are using different "raise hand" systems... the built-in raise hand and "RH" in the chat… maybe useful to clarify to avoid issue with queue management ;)
    11:53:22 From Zou Lan (Huawei) : which tool we use for RH?
    11:53:54 From Kevin McDonnell : lets use chat
    11:56:58 From Jörg Niemöller : rh
    11:58:13 From Abdul Majid Hussain (Telstra) : In the interest of the time, providing my feedback via this chat
    - Agree with most of the speakers, we need to move forward and work towards alignment of taxonomy, architecture, principles, etc
    - We need to agree with what we need to do next and create streams of work
    12:02:04 From klaus Martiny (DTAG) : rh
    12:02:44 From rRanganai chaparadza : Federated Models are part of Modeling and there is a misunderstanding on what is implied by Meta-Model
    12:03:05 From Jörg Niemöller : Federated models is what I have in mind when asking to define responsibilities. I still see that we need to define common modelling and API principles. This is in my understanding the meta-model.
    12:03:35 From rRanganai chaparadza : Fully agree with you Jorg
    12:04:57 From rRanganai chaparadza : RH
    12:05:08 From Emmanuel (Huawei) : +1 Dave
    12:06:14 From Dr.Muslim Elkotob : Fully agree; the collective set of connections in a federated hyper-model that connects all streams or contributions forms jointly a strucuture to which Ranganai referred to as a meta-model
    12:06:59 From Anatoly (Nokia) : @Ranganai - we have been through the "meta-model" discussions (quite literally). The outcome was not federated modeling approach... The FNIM is an example of umbrella model where other models "plug-in"... a meta-model did evolve into "meta-policy" and failed to "fly"
    12:07:02 From Thomas Tovinger (3GPP SA5 chair, Ericsson) : RH
    12:07:22 From Emmanuel (Huawei) : +1 Nurit
    12:07:36 From Leon Wong (Rakuten Mobile) : RH
    12:08:18 From Emmanuel (Huawei) : +1 +1 +1 Nurit
    12:12:07 From Jörg Niemöller : fully agree to Rangannai
    12:12:18 From Anatoly (Nokia) : RH (as an answer to Rangannai)
    12:14:07 From James O Sullivan (Huawei, TMF AN) : Enable distributed ownership with shared conceptual and taxonomic models, not full info models.
    12:14:32 From Jörg Niemöller : If every SDO sticks to their guidelines as they are we will get overlaps and the same things done differently.
    12:15:27 From Raymond Forbes ISG ENI, Huawei UK : Rangannai is correct Frameworks are NOT APIs - Thomas is correct that common documents will take a long time
    12:16:38 From Jörg Niemöller : I also believe that the challanges for realizing autonomous networks will require new approaches and and concepts and not just applying standard telco guidelines to be applied again.
    12:16:59 From Diego Lopez : Some time ago, we went through a similar process in the network virtualization arena. The agreement was to build a map (or “landscape” if you want), trying to avoid overlaps and to identify and align the contact points in a P2P way. I think this fits well with the federation approach proposed before
    12:17:25 From rRanganai chaparadza : Very true Ray.
    12:18:45 From rRanganai chaparadza : Jorg, its is true that the traditional approaches in SDOs is no longer viable in the era of automation
    12:20:37 From Lingli CMCC : with pleasure@kevin
    12:22:26 From rRanganai chaparadza : What we are saying is that FEDERATED Models need to be described in a formal way not using paper text description
    12:24:39 From rRanganai chaparadza : We are not contracting each other as such. Modeling is at the heart of Federated Models as well
    12:24:57 From rRanganai chaparadza : we are not contradicting each other
    12:26:17 From Raymond Forbes ISG ENI, Huawei UK : RH
    12:26:43 From Nurit Sprecher : RH
    12:26:55 From rRanganai chaparadza : There is simply a misunderstanding on interpretation of what a Meta-Model
    12:27:17 From klaus Martiny (DTAG) : i Have to leave sharp at 13:30
    12:28:44 From Jörg Niemöller : I also think that it is not only a matter of which information models survive. This will definitely happen. But something like intent based operation marks a shift from action/process triggering API towards API for life-cycle management of knowledge objects (intent are such objects). These are conceptually different ways of doing API and I believe we have to agree on a common one. Just one example where missing alignment and common concepts will lead to conceptual incompatibility.
    12:29:23 From Ton Brand (ETSI) : same here - need to leave at 13:30 sharp. can we share this google sheet as I am not able to access it via the link.
    12:29:55 From Manchang Ju(ZTE) : fully agree to Lingli
    12:30:14 From rRanganai chaparadza : The problem is that there is often lack of engagements on people or SDOs deeply competent in Modeling
    12:30:24 From Thomas Tovinger (3GPP SA5 chair, Ericsson) : I can't access the google sheet link either, although I have a TMF account.
    12:30:48 From klaus Martiny (DTAG) : cheers
    12:32:26 From rRanganai chaparadza : What we mean by Overarching Blueprint of Common Operational Principles of Autonomic/Autonomous Networks, involves Mappings fore Federated Models as well
    12:32:53 From rRanganai chaparadza : Including Ontologies for Taxonomy Harmonization
    12:33:39 From Anatoly (Nokia) : +1 to Nurit - creating a "meta-organization" attempt would be a real disaster

Action List

Next Meeting details

Scheduled Date:        

Link to Project Calendar:

Apologies for next week’s meeting: 

Requested topics for next meeting agenda: