Epic: Nephio Concepts Series

[johnbelamaric]

Develop a series of short videos, decks, blogs and user exercises to explain Nephio and the concepts underlying it, using R1.

Each episode should include:

• 1-3 slide deck explaining the content
• 10 minutes or shorter video
• Workshop exercises to try
• Which other episodes' content is a prerequisite to understanding this episode

Episodes

• Series Introduction (2 minutes)
  • What is Nephio - in brief
  • The project, the community
  • About the tutorial series
• Why Nephio (10 minutes)
  • Prerequisite: Series Introduction
  • What problems does it solve
• Community (10 minutes)
  • Prerequisite: Series Introduction
  • How does the community work
  • How can you get involved
• Demo Environment (15 minutes)
  • Prerequisite: Series Introduction
  • What’s in it
  • What its purpose is
  • How to get it up and running
  • How to reset it
  • Exercise: get the environment up
• Core Concepts
  • Prerequisite: Nephio Demo Environment
  • Management cluster vs workload clusters
  • Declarative vs imperative
  • Kubernetes and KRM
  • Packages, Repositories, Upstream, Downstream
  • Exercises:
    • Clone a cluster creation external package into an organizational package
    • Customize the organizational package with the UI
    • Publish the package and deploy it to the management cluster
    • Observe the creation and registration of the workload cluster
• Deploying a Package Many Times
  • Fan out
  • Fan out with variations
  • Exercises:
    • Creating more workload clusters
    • Deploying a workload across all workload clusters
• Collaborative Package Configuration
  • Nephio package configuration lifecycle overview
  • Package conditions and multiple user interaction
  • Specializers and external system integrations
• Deep Dive: PackageVariant and PackageVariantSet
• Deep Dive: Nephio Package Configuration Methods
  • Different ways to mutate a package as it goes from the original source, to your workload cluster
  • How and when to use different methods, their pros and cons
  • Mutations on clone, KRM functions, injectors, package fanout, apply time mutation, mutating webhooks, runtime operators
• Deep Dive: KRM Functions

Future Potential Topics

• Package Dependencies
• Nephio Network Function APIs
• Nephio Infrastructure Management APIs
• General tools and techniques for reducing complexity in configuration

[electrocucaracha]

I'm not sure if this is the right place to comment, but I suggest to use Divio documentation system for this project. Essentially, it doesn't change anything, but it helps to structure the documentation.

[johnbelamaric]

I would leave it up to SIG Release to determine how we manage and package docs.

[johnbelamaric]

/retitle Epic: Nephio Concepts Series

[johnbelamaric]

/area docs

[debeaueric]

I think that we should add

• a video explaining NFDeployment concepts and to illustrate how a package is enriched via kpt pipeline for SMF
• a video explaining how R1 is using KRM for networking between clusters

[rravindran123]

• Developing Use Case Specific CRDs using Nephio's Automation platform to provision Clusters and Network Functions, e.g. CRDs that were developed in R1 to handle 5GC (AMF/SMF/UPF) NFs in Free5gc

[johnbelamaric]

Some more ideas:

• What's in R1 (use cases, capabilities, demo)
• How data gets into apiserver(s)
  • imperative / kubectl create
  • declarative / kubectl apply
  • helm
  • operators
  • gitops
  • pkg mgmt and specialization is how data gets into git (backing up the process further)
• Nephio and GitOps (either one video or maybe multiple for different tools - ConfigSync, Argo, Flux)
• Nephio and Crossplane
• Nephio and Helm

[tliron]

Proposing two workload-cluster videos:

• Using the free5GC operator to deploy UPF, AMF, and SMF (without using the rest of Nephio, link to videos that show the complete end-to-end flow)
• How did we convert the towards5GS Helm charts to operators? Code and architecture overview

[efiacor]

Proposal for the UI lovers:)

• Closer look at the nephio web ui dashboard. Repos, blueprint types, etc. and their intended uses.

• Also (if useful) a look at how config-sync hooks the above up to the gitea backend?

[johnbelamaric]

In the TSC approval slide, I listed a few "General Purpose Config Management Capabilities". I think these are all covered above, but I am copying them here just so we can make sure.

• Bundle infrastructure and workload intent into packages
• Day 1 general purpose package management
  • Discovery, inventory, cloning, organizational, team, and multi-level trees of customization
  • Package deployment across a fleet, with variation based on region, location, and other deployment context
  • Delivery of fully configured packages via GitOps to edge clusters
• Day 1 package specialization
  • Collaborative package editing by automation and humans
  • Package specializers that integrate with external resource backends
  • Developer SDK and specializer test framework to simplify development of specializers
• Day 2 package management
  • Regulated, manual adoption of changes to configuration across the fleet
  • Discovery of updated packages and the affected downstream deployments
  • Propagation of input changes across affected downstream deployments

[johnbelamaric]

I think for each of these suggestions, we probably need to create a separate issues, and assign it to someone. The issues should each include an outline of the contents of that episode.

[johnbelamaric]

FYI, this is underway, see: https://wiki.nephio.org/display/HOME/Learning+with+Nephio+R1

So far I put together:

• Demo video
• Series introduction
• Why Nephio?
• The Nephio Approach

I will leave this open for now, but eventually we should make separate issues for each future episode and close this issue.