Date: Fri, 29 Mar 2024 02:14:37 +0000 (UTC) Message-ID: <2001852433.29626.1711678477251@aws-us-west-2-lfn-confluence-1.web.codeaurora.org> Subject: Exported From Confluence MIME-Version: 1.0 Content-Type: multipart/related; boundary="----=_Part_29625_916845206.1711678477251" ------=_Part_29625_916845206.1711678477251 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Content-Location: file:///C:/exported.html
This session is to introduce the survey drafted= by LFN projects Anuket and XGVela together. Some of the key objectives of this Survey are:
We would discuss the details of the topics and questioner and the n= ext course of action.
https://wiki.anuket.io/display/HOME/Draft+Pared+Down+--+LFN+PaaS= +Survey
Follow up to NFV testing White Paper =E2=80=93 What test suites are need= ed in the Lab and what other work needs to be done, which will include:
1.What are the common requirements/ information for doing Cross-o= rganization automatic testing
2.What is the requirements for supporti= ng NFs in the testing process considering 5G background
3.AI/ML testi= ng platform
4.Other potential requirements
Working Session for the AI/ML White Paper =E2= =80=93 gather requirements from the community for targeting projects , incl= ude:
Overview of of open sou= rce participation at a high level as well as an overview of LFN's structure= . Much of the event will rightly be focused on deep dives and moving t= he ball forward on specific work items, but for new members, new folks at e= xisting members, and interns, some background context and overall shaping o= f how progress occurs in such a large and diverse community can be quite us= eful.
Welcome and opening co= mments
Regular EUAG Meeting =E2=80=93 Needs to be= scheduled Tuesday 10AM EST (the regular meeting time)
This page will be used to provide a brief summary of the daily = topics, provide feedback to the LFN staff, and plan for the next day.
Meeting minutes and action items should be recorded on your Topic page
=ODL - Our Recent Testing with Container Deployment and = Observations
Ove= rview of the changes delivered by the platform projects in the Phosphorus t= imeframe.
The Simulta= neous Release is supported by the autorelease project, which now hosts only= a few projects. This will be an open discussion as to what options we have= and how can we evolve our tooling.
This is a placeholder for open-mic discussion arou= nd topics which either cropped up during other sessions or did not fit in t= hem.
A new TSC Task Force was created on = January 20th, 2021 in order to define ONAP added-value for Enterprise Busin= ess.
This session will = share the role of ONAP in the 5G Super Blueprint<= /a> and share how the ONAP Platform will interact with the Magma open source platform
During this s= ession, semi-live demonstration of the latest capabilities introduced = in Guilin (CNF Adapter and Helm type) and developed in Honolulu (CNF Day2 s= upport) will be presented. The demonstration will show the CNF/Helm onboard= ing and deployment process of vFW CNF use case which since the Honolulu rel= ease is enriched with config-assign/deploy workflow that performs the CNF c= onfiguration process. Moreover, the demonstration will show how to track an= d validate the deployment status of CNF, and how to fetch information about= the status of all resources deployed on the k8s cluster.
Most practical deploymen= ts of edge infrastructure and applications are hybrid in nature, where an a= pplication deployed at the edge often needs to access services residing in = the core cloud. In addition, there is a need for efficient and performant i= nterconnection between edge and cloud as well as between the distributed ed= ges proximal to end users. However, each individual service domain (e.g. ed= ge, cloud, network fabric) often presents their own specific APIs and/or ot= her provisioning methods (e.g. CLI), thus making end-to-end deployment chal= lenging both in complexity and in time. Therefore a multi-domain orchestrat= ion solution is required to handle edge, cloud and interconnection in a uni= form and consistent manner. Terraform is rapidly emerging as a common way t= o orchestrate and configure services rather than having to deal with APIs f= or each different domain: public or private cloud, edge and interconnection= . In this presentation and hands-on demo, Equinix and Aarna Networks will s= how how an end-to-end application can be deployed, configured and interconn= ected between the edge and core clouds by using ONAP/CDS orchestratio= n in combination with Terraform.
Istanbul CPS development updates for NCMP and Te= mporal
This session will provide a brief overview of E2E Net= work Slicing use case - what has been achieved so far and the road ahead. A= Demo of a part of the functionality realized up to Honolulu will be given.=
This session will describe the proposed new fea= ture in SDC to support design/onboarding of services/resources natively in = models other than SDC data model
Presentations on topics of next generation ONAP Security and= Logging
1) Analysis between AAF and Service-Mesh Security
2) S= ervice-Mesh Security architecture, leveraging Istio, KeyCloak, CertManager,= Ingress
3) Design, implementation and roadmap
4) Logging Archi= tecture, leveraging open-source and standard technologies
DB plans, Control Loops via TOSCA,= Policy-clamp UI demo
This session will = present the features developed in Honolulu and will share the requirement c= andidates for the Istanbul release, followed by a proposal of "Simplified CNF modeling".
<= /span>With FTTx rollouts progressing, there is an increasing need= to manage 10s or 100s of thousands of OLT switches. CDS provides an elegan= t solution to manage the configuration of these OLT switches. In this talk = and hands-on demo we will discuss the CDS Blueprint Archive we developed to= solve this issue. After watching this presentation, you will be able to un= derstand how to use CDS to configure PNFs using Telnet.
Share T= ransformation updates from Honolulu release and review planned scope for Is= tanbul.
Presentations on topi= cs of common interest between ONAP and SDOs
1) Hot topics on automati= on in 3GPP SA5
2) Closed Control Loop Automation - ETSI ZSM, IRTF, ONAP
30min
Presentations on topics of SO Custom Workflow= Onboarding and Orchestration
1) SO Custom Workflow Architecture, Des= ign and Implementation plan for Istanbul +
2) Scenarios that allow th= e user to add custom workflows to ONAP SO
3) Custom Workflow package = and version control
Intent-based network is deve= loped to support the smart applications of usecases in ONAP.
First pa= rt of the topic gives the progress of R8 intent translation and the proposa= l of R9 intent instance are introduced to support E2E Slicing and CCVPN use= cases.
Second part of the topic introduces the proposal of closed-loo= ped CCVPN intent guarantee.
This session is an open forum to collect feedback, raise any question= and provide suggestions from the people who are attending the DDF event to= the ONAP "Cloud Native" TSC Task Force.
This session is an open forum to collect feedback= , raise any question and provide suggestions from the people who are attend= ing the DDF event to the ONAP TSC.
In= this session, we will discuss work under way within the ONAP CCSDK and SDN= C projects to "decouple" our code from OpenDaylight. The goal is to a= llow ONAP users the flexibility to deploy CCSDK/SDNC with a newer (or older= ) OpenDaylight release than the release used during ONAP release developmen= t.
In this session, = we will show how to integrate ONAP network slicing with a commerc= ial 5G core. We present the options that are available in ONAP for suc= h an integration, and show a demo of the feature, by creating network slici= ng templates from ONAP SDC, create a core network slice, activating it (fro= m ONAP UUI), and running a test using a commercial UE/gNB simulator. We als= o present the future roadmap, and what other possible integration options a= re available with ONAP Network slicing.
This session shall provide a brief ov= erview of what has been realized in the SON use case so far. It will then d= iscuss the road ahead for the use case, and the considerations w.r.to (O) R= AN from a broader ONAP perspective.
Presentation of what has been done in OOM for H releas= e
Plans for OOM in Istanbul release and afte= r
BluVal (Blueprint Val= idation) Framework is a cloud native full stack test automation tool based = on a disaggregated and layered approach for test and validation of implemen= tations to ensure conformance. BluVal disaggregates the implementation and = associated manifests into their layers and provides automated validation fr= amework to test each layer one by one to achieve end-to-end validation of t= he implementation. All results are standardized by using Robot Framework wh= ich creates an html file of test results that is easy to read and interpret= . By virtue of its modular design, the BluVal framework is extensible and p= rovides a simple, declarative test harness for authoring and executing test= suites. New functional tests can be written up from scratch in Robot, with= simple human-readable commands, or existing upstream tests can be quickly = integrated with a few lines of template code that can call any other testin= g or scripting tool. BluVal framework also provides a portal (User Interfac= e) for viewing executed test result.
Overview of an academic study/experiment done around OpenStack= management and operations. Openstack is a fairly complex open-source cloud= software with many loosely coupled components. This presentation covers to= pics around OpenStack architectural design and cloud-native values that can= be built into OpenStack. A cloud-native OpenStack model can decrease the m= anagement and operations requirements of OpenStack, make things easy to ope= rate and automate.
There are several leftovers from the CNTT - OPN= FV merger what we should work on. In this session I would like to collect t= hese items and propose resolution if we can.
Leftovers from CNTT what= I can remember:
This = is the regular meeting of the Anuket TSC, where much of the June 8 agenda w= ill likely be dedicated to the impending Kali Release decisions.
Security is k= ey within a virtualized environment, it's an important topic for Anuket whi= ch aims to specify a trusted Cloud Infrastructure for workloads.
Duri= ng this session, we will discuss the updates made to the Reference Model fo= r Kali release: open source software security, automation security challeng= es...
Traffic generators and load generators are critica= l for different performance testing. In this talk we present the opportunit= ies and challenges of containerizing these generators for Kubernetes perfor= mance testing.
Discussing multi- and hybrid cloud requirements and= choices as it is driven by edge and other use cases.
The session is = a continuation of the similar discussion at the recent PTG: https://etherpad.opendev.org/p/ecg-ptg-april-2021
=Recording of the PTG discussion: https://zoom.us/rec/share/yMN8QANWzI7FJ_Hh-ys166= _sECX2XgSgX-B_jKnnlYh4UjCAv9S32s-9U4mzjUGf.HZIErlBO6LOIMrJz
Relev= ant RM sections:
RM 3.5.2 Network Layering and Concepts https://github.com/cntt-n/CNTT/blob/mas= ter/doc/ref_model/chapters/chapter03.md#3.5.2
RM 3.6 Storage https://github.com/cntt-n/CNTT/blob/master/doc/ref_model/chapters/c= hapter03.md#3.6
RM 3.8 Hardware Acceleration Abstraction https://github.com/cntt-n/CNTT/blob/mas= ter/doc/ref_model/chapters/chapter03.md#3.8
RM 3.8.3 Workload Placement https://github.com/cntt-n/CNTT/blob/master/doc/ref_m= odel/chapters/chapter03.md#3.8.3
RM Ch 8 Hybrid Multi-Cloud: Data Center to Edge https://github.com/cntt-n/CNTT/blob/ma= ster/doc/ref_model/chapters/chapter08.md
Relevant RA sections:
Wh= ich OpenStack release to select as a baseline for the next RA1 release?
=Following RM's developments on acceleration, RA1 must evolve to integrat= e these new requirements.
Workloads expect certain capabilities and performance from the resource=
s that they will execute on. Creating workload specific configuration=
s becomes not only unmanageable but may render the resources un-shareable.&=
nbsp; A profile is a set of resource configurations that meet need=
s of a family of workloads. Profiles partition the infrastructure resources=
and, thus, the number of profiles affects resource utilisation. Newe=
r workloads, such as for 5G or O-RAN, leads to demand for new profiles or p=
rofile extensions. In Anuket, we recently decided on the concept of profile=
extensions. With the new definition of two high level profile abstractions=
and associated profile extensions, flavours have also been redefined. This=
session will seek a discussion on the efficacy of the adopted approach, al=
ternatives, definition of additional profile extensions, and workload flavo=
urs.
Anuket is promote HW i= ndependent RM, RAs and RC. For accelerators for RAN workload oRAN Alliance = is defining Acceleration Abstraction Layer and APIs - htt= ps://oranalliance.atlassian.net/wiki/download/attachments/872841331/O-RAN.W= G6.AAL-GAnP-v01.00.pdf?api=3Dv2. However Kubernetes currently does not = have an abstraction for accelerators and each accelerator is handled and ne= ed to be programmed to independently. This session will discuss current sta= te of affairs and what can be done to remedy it.
Let's review where are with RA2 and what will be our targets for= Lakelse release.
Review and discuss newly developed Anuket release process. &= nbsp;If time permits, we will also look at the schedule for the Lakelse rel= ease.
In thi= s session we will review the current status of the RI2 Kuberef and Airship = projects and plan the next steps forward. A special focus will be on the re= quirements, gaps and conformance and how we can best work together.
Documentati= on of Anuket is developed in different places and delivered to different pl= aces. I think we should keep the option to develop the documentation in dif= ferent places, but we should deliver the documentation to a single place wi= th similar look and feel. Let's discuss if this is an agreable target and w= hat are the alternatives to achieve it.
Anuket RM and RA2 define the prope= rties for Kubernetes based infrastructures to run CNF-s to ensure quick and= easy integration of CNF-s and the infrastructure. With the the different p= ublic cloud providers offering CaaS services for telecom applications a new= integration front is opened. Is it guaranteed that a CNF compliant with RC= 2 will run on the hyperscaler CaaS offerings? How well RM and RA2 modells t= he infrastructure of hyperscalers and what is the best way to ensure intero= perability between private and public CaaS offerings?
This session will cover working with the TF Bu= ild Environment, how to get started building TF from scratch, and how to pr= oduce kernel modules and other artifacts to add support for new distros and= kernels to TF.
Diving into Tungsten Fabric = -Bite size environments for Kubernetes and OpenStack Labs.
This talk = will cover the common installation options required to deploy OpenStack and= Kubernetes environments for use with Tungsten Fabric. Using the tf-ansible= -deployer and container-builder manifests we will explore the common deploy= ment architectures of TF that can be used for lab/validation. We will also = provide an overview of how these examples can be modified for use in higher= scale/complexity labs and deployments.
This topic covers the updated Tungsten Fabric release process= and plans for our releases in the upcoming 6 months.
The trend in edge computing and Kubernetes is to deploy a large numb= er of clusters across a wide area network, or across public clouds. This to= pic will focus on the capabilities of TF that can stitch together Kubernete= s clusters and provide seamless networking across disparate clusters.
= span>This session will provide an= overview of how scalable application centric policies can be used to deliv= er multi-cluster and multi-orchestrator security. An overview of multiclust= er overlay networking, and application security will be provided before div= ing into key enhancements in multicluster policy management.
This topic will cover the features and enhancements in TF that are allo= wing Telco Cloud deployments to adopt modern Cloud Native practices. This w= ill include an overview of dynamic data plane learning and aliveness valida= tion, BGPaaS for containers, and policy aware routing for multi-interface c= ontainers. This session will also cover key advanced networking concepts re= quired to deliver network slicing.
In this session, XGVela TSC will give upd= ate on project progress and answer cummunity questions.
In this session, = Qihui and Saad will introduce several use cases about PaaS in operators' cl= oud to show where we need PaaS, what functionalities are requried and how t= hese requirements are aligned with XGVela seedcode functionalities. = p>
In this session, Sandeep and Va= nce will introduce Information model and API approach for Telco-PaaS functi= onalities CMaaS, TMaaS, FMaaS and MMaaS.
What has the ODIM project been= up to since last event? What is ODIM's role in the 5G Super Blueprint? Wha= t are the new services and features planned for