KNative

Interactive Learning Environment

In this lab, we will examine the Docker CLI and execute our first container.

Kubernetes CNCF: Module 6 - KNative

All of our labs are powered by Katacoda and are located at this Profile.

Background

Event-Driven, sometimes also called serverless or functions as a service, is a computing execution model in which the infrastructure/provider dynamically manages the allocation of machine resources.

Code is deployed to a CSP or elsewhere
Instance does not exist until invoked, spins up and scales as required, scales to zero when complete
Changes IT requirements related to security, administration (patching) and architecture

Currently there are various standards with little interoperability between Cloud Service Provider’s. Common examples include the following:

IBM OpenWhisk
AWS Lambda
Google Cloud Functions
Azure Functions

Overview

KNative Architecture

KNative can be used to deploy serverless-style functions, applications, and containers to an auto-scaling runtime on Kubernetes. Additionally it has many other features:

Deploy multiple versions
Perform custom tasks on your app’s source code
Build reusable templates

Serving

How your code recieves requests and scales with them

Request-driven compute runtime
Scale-to-zero / scale out per load
Deploy from container registry
Multiple revisions of same app
Route traffic across revisions

Build

How your code is built and packaged as a container

Pluggable model to build containers from source code
Build in-cloud or on-cluster
Push image to registry
Templates available (buildpacks)

Eventing

How your code is triggered by events and executed

Apps and functions consume and publish event streams
Multiple event sources available
Encourages asynchronous, loosely coupled architecture

Note: Initially serverless, Functions as a service (FaaS) and event-driven architecture referred to a microservice that is run on public cloud compute engine only when invoked. Recently, serverless is also being used to describe managed container services (CaaS) and application deployments where the user is not responsible for the compute where/when the container runs (Azure AKS and ACI, AWS EKS and Fargate, Google GKE, Google App Engine, and IBM Kubernetes Service).