September 26-28 Topics

Owned by Kenny Paul
Last updated: Nov 13, 2017 by Shuki Yulzari
23 min read

Service Orchestration Workflow Development and Testing~The September Developers Event is for ONAP developers, by ONAP developers. Your contributions and needs are what will drive the agenda for all of these sessions.  To build the agenda for the we are soliciting Topics for discussion from the community.  Ideally these are questions that you need a deeper understanding of in order to make progress or deep-dive information sharing in a particular area.

Topic Template - Cut&Paste as needed

Topic/Question Name

  • Description: Description of the topic and the objective/outcome desired at the end of the discussion.
  • Topic Leader: The person that is knowledgeable enough to lead and moderate this discussion.
  • Volunteer Note Taker: First Last  email
  • Estimated Duration:
  • Link to data Source (if applicable)


INSTRUCTIONS:  

Please edit the section below and add any topics (using the template) that you would like to lead and/or attend. List the topic, and your name. If there is a topic that you would like discuss, but you do not feel that you can lead it, please list the topic below and indicate that you are "Interested in Attending", then indicate TBD as the "Topic Leader"

IMPORTANT DATES

Community Members provide Topic recommendations Aug. 3 - Sept. 10

Attendees self-select Topics of interest Sept. 11 - 15

Agenda set Sept. 18

TOPIC RECOMMENDATIONS



Amsterdam M4 Code Freeze Milestone Review  

  • Description: M4 Code Freeze Review of all Amsterdam Projects. TSC will vote on the projects and depending on the findings, TSC will need to decide on:
        1. projects who pass M4
        2. projects that will need to be removed from Amsterdam
        3. projects that will need some adjustements


Training: On the Importance of adopting Scrum and Estimating 

  • Description: Discussion on the benefits of applying Scrum techniques and Poker Planning for Estimating.
    1. Why is Scrum important for the Team. How to apply Scrum for distributed Team
    2. You can't swallow more that you can shew: Estimating with Poker Planning will make the team more realistic.

Beijing Release Calendar Proposal

Enforcing an "Upstream first" approach to ONAP

  • Description: Description of the topic and the objective/outcome desired at the end of the discussion.
  • Topic Leader: The person that is knowledgeable enough to lead and moderate this discussion.
  • Volunteer Note Taker: First Last  email
  • Estimated Duration:
  • Link to data Source (if applicable)
  • Interested to attend:


External Open Source collaboration (OPNFV, ODL, etc.) and required actions  


External Standards Coordination and required actions with a different SDO   


Joint Architecture/Usecase subcommittees' meeting 


Joint Architecture/Usecase/Modeling subcommittees' meeting 


ONAP security subcommittees' meeting   

  • Description: A discussion and feedback on ONAP security topics:  Static Code Scanning; CII Badiging program certification feedback (CLAMP team); Credential Management
  • Topic Leader: Alla.Goldnerephen Terrill; (together with a CLAMP team memember)
  • Volunteer Note Taker: Pawel Pawlak pawel.pawlak3@orange.com 
  • Estimated Duration: 1 hour
  • Link to data Source (if applicable)
  • Interested to attend:


Operator Perspectives on R2 (Beijing) Architecture 


R2 (Beijing) use cases/platform capabilities first presentation to the TSC 


Usecase subcommittee meeting  


Installation hands-on


Kubernetes ONAP deployment overview and hands-on  


Modeling 


Residential vCPE Use-Case deep dive 


VoLTE Use-Case deep dive 


Microservice Bus Tutorial 


Introducing CLI Model engine &  Tutorial 


Integration / Testing 

  • Description: This session will give an overview of ONAP status from Integrtaion / Testing point of view:
    1. ntegration testing process and its status:
      1. UT
      2. CSIT
      3. Integration lab, this includes pair testing rules for each project, this includes a brief on how to use the lab in general, Stephen and Yang will give details later
      4. End to End lab status overview for vCPE, VoLTE  (deep dive will be covered in another two sessions) and vFW / vDNS
      5. Integration deployment strategy for Amsterdam: assets requirement and deployment topology, quick review of how to use heat template to deploy ONAP in Integration lab
    2. Integration testing practice and results from Orange open lab


How to use ONAP Community Lab 


R2: Non-functional requirements proposals  


ONAP Development Tutorial 


Control Loop - E2E 


LF Productivity and Process Town Hall Session  

  • Description:  Everyone has their favorite tool they want to use to work and communicate more efficiently.   The LF IT Department gets requests to add new tools or functiuonaluty on an almost daily basis, but accomodating all of these isn't feasable.  We need a more complete strategy for supporting your needs and this Monday session is intended to gather community input and share ideas on what is working well and can be improved upon in the area of tools and processes.  NOTE: This is not a technical "how-to" session.
  • Topic Leader:Kenny Paul  Tech. Program Manager, The Linux Foundation
  • Volunteer Note Taker: First Last email
  • Estimated Duration: 1hr
  • Interested to attend:


Documentation: how to describe detailed call flows 

  • Description: To undestrand the various ONAP component inteactio, call flows description are very impôrtant. They are currenly used in various projects  and use-cases. Various tools are used to describe call flows with various levels of details. This session will present a way to describe some detailed call flows with API calls between various components using PlantUML.
  • Date: TBD
  • Topic Leader: Eric Debeau - Orange
  • Volunteer Note Taker: 
  • Estimated Duration: 30'
  • Interested to attend:


Future Plan for the Development of Holmes 

  • Description:  The basic analysis capability of Holmes will be available by R1. We want to discuss how to enhance the functionalities to make Holmes be able to support more use cases and scenarios. Besides, we want to have a brief introduction on the vision of a smarter and more general analysis application based on the big data components provided by the ONAP community. 
  • Topic Leader: Guangrong Fu
  • Volunteer Note Taker: First Last email
  • Estimated Duration: 1hr
  • Interested to attend:

VoLTE E2E Service Design Demo

  • Description:  David Shadmi from SDC team gave a small group people a demo of VoLTE E2E service design by using early version of SDC, it was very helpful. We would like to see the same demo with the latest SDC code and bigger audience to get the feedback.  
  • Topic Leader: David Shadmi
  • Volunteer Note Taker: Yang Xu
  • Estimated Duration: 1hr
  • Interested to attend:


End-to-End Model Driven VNF Lifecycle Management/Modeling 

  • Description: As part of the VNF SDK project, VNF modeling subproject has been focusing on establishing a consistent end-to-end approach for VNF and VNF package modeling in R2. We have identified several starting point for standardization of these models. What we need to do next is to agree on what models and/or standards are going to be applicable at what stages of the VNF lifecycle.

    We will start with the VNF model, as it is seen by VNF developers, inclusive of compute/storage/network hardware requirements, telemetry and policy specification, and identified how this information is ingested and trasformed into other models as part as part of VNF on-boarding (SDC), instantiation (SO/VF-C/Optimization/Policy) and operation (APPC/VFC/Optimization/Policy).

    We will also need to understand how policies and telemetry apply across the VNF management lifecycle, at what point they are defined, who gets to define them and how they are blended/federated with policies already provided during VNF on-boarding.


Enhancing ONAP to orchestrate container VNFs using container orchestration as VIMs 

  • Description: Release 1 ONAP instantiates VNFs as VMs in compute farm using VIMs such as Openstack which manages VM. Increasing VNF density on each node, need for multiple networks/slices, Services using SFC and reduced bringup & network latency requirements in market such as 5G and vCPE  are driving container based VNFs.  This project proposal is to enhance ONAP to support VNFs as containers in addition to VNFs as VMs by container orchestration as VIM instead of VM manager like openstack.
    • Objective/Outcome desired:
      • Get consensus on crawl/walk/run direction.
      • Get agreement on K8S as container VIM in R2
      • Get feedback on proposed architecture and changes to various ONAP projects.
      • Get an understanding that VNFs as container is needed for ONAP R-2 use cases(5G-RAN and vCPE)


DCAE New Features and Enhancements for Beijing and Future Releases 


SDC: R2 Backlog Discussion 


VID: R2 Backlog Discussion  

  • Description: Presenting and further gathering of R2 candidate requirements
  • Topic Leader: David Shadmi 
  • Volunteer Note Taker: 
  • Estimated Duration: 1 hours
  • Link to data Source (if applicable)
  • Interested to attend:

AAI: R2 Backlog Discussion  

  • Description: Presenting and further gathering of R2 candidate requirements
  • Topic Leader:Colin Burns
  • Volunteer Note Taker: 
  • Estimated Duration: 1 hours
  • Link to data Source (if applicable)
  • Interested to attend:

AAI: Graph DB Evolution 

  • Description: Discuss Beijing plans to move to JanusGraph, migration strategies, etc.
  • Topic Leader: Jimmy Forsyth
  • Volunteer Note Taker: 
  • Estimated Duration: 1 hours
  • Link to data Source (if applicable)
  • Interested to attend:
         


Service Orchestration Workflow Development and Testing  


TOSCA Native Service Orchestration Design 


ONAP Modeling: long-term strategy vs short-term tactics  


  • Description: Discuss goals, challenges and approaches to define an ONAP modeling long-term strategy (R2 and beyond), with the emphasis on E2E service automation, in sync with long-term architecture and driven by ONAP Service Provider requirements, and the short-term tactics (R2) in support of the long-term strategy; consider requirements, architecture, info models, data models, and DSLs in this order, and discuss the landscape of standards and open source communities contributing to them. Make specific long-term strategic, as well as specific short-term tactical recommendations for ONAP and other SDOs/open source communities.
  • Topic Leader: Michael Brenner, Cloudify, michael@cloudify.co
  • Volunteer Note Taker: First Last  email
  • Estimated Duration: 1.5 hrs
  • Link to data Source (if applicable)
  • Interested to attend:

Workshop Title: ONAP Multi Cloud Architectural vision for R2 and beyond

Goal:                               Advance Multi Cloud beyond the current proxy implementation to address ONAP platform level issues across all use cases

Workshop Lead:            VMware

Primary Contact:           Ramki Krishnan, VMware, ramkik@vmware.com

Estimated Duration:      3 hrs.

Desired Scheduling:     Before Architecture and TSC meetings. After lunch scheduling is desired since some folks are connecting from US and may not be able to attend in person. 

Interested in Attending:

Slides: ONAP-mc-workshop-agenda.pdf

Minutes on Multi-Cloud follow up and R2 scope discussion (09/28/2017): Multi-Cloud follow up and R2 discussion.pdf

Subtopic 1:   Multi-Cloud Architectural Vision Introduction 

Description:   This session will focus on discussion of the evolution of the Multi-Cloud framework for R2 and beyond, to address some of the platform gaps and move towards a more extensible and consistent cloud mediation layer. We would like to introduce a few key focus areas for this evolution, both in term of the use cases, architectural design principles, and integrations.

Architectural focus: Model Driven API for cloud infrastructure, Standardized cloud telemetry management, and policy driven cloud agnostic deployment.

Deep-dive subtopics (2, 3 and 4) towards achieving this vision are listed below for convenience

  • Performance & Isolation - "Towards a performance-aware and portable cloud-agnostic Infrastructure"
  • SDN - "Architectural options for Multi-vendor SDN Controller and Multi Cloud Deployments in a DC"
  • FCAPS Telemetry - "Standardized Infrastructure class statistics Model" 

Suggested Audience:      ONAP-OF, OOM, SDN-C, SO, VF-C, APP-C, DMaaP

Estimated Duration:        1.5 hrs

Subtopic Lead:                 VMware

Presenters:                       AT&T, VMware, Wind River, Intel

Slides:                               ONAP-mc-intro.pdf


Subtopic 2: Performance & Isolation - "Towards a performance-aware and portable cloud-agnostic Infrastructure" 

Description:                    Multi-vendor cloud portability and interoperability while delivering performance is a mandatory feature of distributed DC deployments. The practical challenge in achieving this goal is the lack of standardization of Platform-aware and QoS features, for example extra specs in OpenStack, which results in a vertically integrated solution. This session will focus on how Multi Cloud can offer policy standardization and translation as a microservice for platform-aware and QoS features (hard-guarantee, min-guarantee, best effort etc.) for addressing this challenge and how NFV application classes such as IMS/EPC Control/Data Plane can benefit from this framework.

Suggested Audience:     VNF Requirements/Modelling, A&AI, Policy, SO, VF-C, APP-C, SDC

Estimated Duration:       .5 hrs.

Subtopic Lead:                Bin Yang, Wind River, bin.yang@windriver.com 

Presenters:                      Wind River, Intel

Slides:                              ONAP-mc-performance-isolation.pdf

Subtopic 3:  SDN - "Architectural options for Multi-vendor SDN Controller and Multi Cloud Deployments in a DC" 

Description:                    Multi-vendor components in a DC including underlay, overlay, gateway and SDN Controllers for the same pose various architectural challenges in terms interoperability, security etc. This session will focus on an in depth analysis of the architectural options including related efforts such as OpenStack Gluon. This proposal also addresses communication latency-bound multi-side physical DCs which can be treated as a single virtual DC for management purposes.

Suggested Audience:    SDN-C, SO, VF-C, APP-C, SDC, OOM

Estimated Duration:       .5 hrs.

Subtopic Lead:               AT&T

Presenters:                     AT&T, Intel, Huawei, VMware

Slides:                             ONAP-mc-sdn.pdf

Subtopic 4:   FCAPS Telemetry - "Standardized Infrastructure Class statistics Model" 

Description:                   This session will focus on a hierarchical cloud-platform-aware architectural framework with separation of collection, storage, processing functions and real-time vs historical analytics components. In this framework, Multi Cloud will deliver a standardized infrastructure class statistics model for driving ONAP component/VNF placement/change management across distributed DC multi cloud instances through ONAP-OF, DCAE and other components. The benefit to ONAP platform, across all use cases, will be delivering the best performance and security while minimizing cost.

Suggested Audience:      ONAP-OF, DCAE, A&AI, Policy, OOM, SDN-C, DMaaP

Estimated Duration:        .5 hrs.

Subtopic Lead:                 AT&T

Presenters:                       VMware, Wind River, AT&T, Intel

Slides: ONAP-mc-fcaps.pdf

Kubernetes vs Dockers Swarm supporting ONAP-OOM on multi-cloud multi-stack environment 

Description:  ONAP was set originally to support multiple container platform and cloud through TOSCA. In R1 ONAP and OOM is dependent completely on Kubernetes. As there are other container platforms such as Docker Swarm that are gaining more wider adoption as a simple alternative to Kubernetes. In addition operator may need the flexibility to choose their own container platform and be open for future platform. We need to weight the alternatives and avoid using package managers as Helm that makes K8s mandatory.

The use of TOSCA in conjunction with Kubernetes provides that "happy medium" where on one hand we can leverage Kubernetes to a full extent while at the same time be open to other alternative. In this workshop, we will compare Kubernetes with Docker Swarm and walk through an example of how ONAP can be set to support both platforms using TOSCA.

            

Suggested Audience:    SDN-C, SO, VF-C, APP-C, SDC, OOM

Estimated Duration:       1 hrs.

Desired Scheduling:      Before Architecture and TSC meetings.

Topic Lead               :      Cloudify

Presenters               :      Cloudify, AT&T (to confirm)

Interested in Attending:

SLIDES 

Amsterdam Release Lesson Learned

Description:  The purpose of this topic is to share preliminary Amsterdam Release Lesson Learned, to reflect together on how everything went and then to decide what changes we want to make in the next ONAP release (i.e. Beijing), answering the following questions:

  • What worked well?
  • What did not work well?
  • What actions can we take to improve our process going forward?

Suggested Audience:    Anyone but ONAP Release Manager and PTLs are expected

Estimated Duration:       1 hrs.

Desired Scheduling:      No preference

Topic Leader:     Catherine Lefèvre (AT&T)

Volunteer Note Taker: Gervais-Martial Ngueko (AT&T)

Interested in Attending:



VID: R2 Backlog Discussion  

  • Description: Presenting and further gathering of R2 candidate requirements
  • Topic Leader: David Shadmi 
  • Volunteer Note Taker: 
  • Estimated Duration: 1 hours
  • Link to data Source (if applicable)
  • Interested to attend: