Introduction✍️

Networking in k8s is such a crucial topic to understand and a must for working in k8s. Here we will discuss networking, particularly in the work use of k8s directly. So before you come to this topic you must have knowledge of basic computer networking.

Switching
Routing
Gateways
Bridges

PODS📦

Before jumping right into the Kubernetes networking stuff. First, let's recall the Pods concept because we will be using Pods in every sentence further.
Whatever we are doing, the main goal is to deploy our application in the form of containers on a worker node in the cluster which must be up and running.
As Kubernetes does not deploy the containers directly on the nodes.
The containers are encapsulated into an object known as Pods
The Pod is a single instance of an application.
A Pod is the smallest object that you can create in Kubernetes.
A simple yaml pod file:


      apiVersion: v1
      kind: Pod
      metadata:
        name: demo-pod
      spec:
        containers:
        - name: demo-container
          image: nginx
          ports:
          - containerPort: 80

Then run the kubectl apply command once you have created the YAML file to deploy it in the Kubernetes cluster with the necessary file name you have given.

kubectl apply -f demo-pod.yaml

This is how you can start creating PODs.

Network Policies📋

In Kubernetes, every routing network traffic instruction is set by the network policies.
A set of tools that defines the communication between the pods in a cluster.
It is a powerful tool used for the security of network traffic in k8s clusters.
Allowance of traffics from one specific pod to another one.
Restricting traffic to a specific set of ports and protocols.
Implementation is done by Network APIs.
It can be applied to namespaces or individual pods.
A simple network policy yaml file:

  apiVersion: networking.k8s.io/v1
  kind: NetworkPolicy
  metadata:
    name: demo-network-policy
  spec:
    podSelector:
      matchLabels:
        app: my-app
    policyTypes:
    - Ingress
    ingress:
    - from:
      - podSelector:
          matchLabels:
            app: allowed-app
      ports:
      - protocol: TCP
        port: 80

For the deployment of the Network Policy YAML file, use the kubectl apply command:

kubectl apply -f demo-network-policy.yaml

This is how you can define your own network policies.

Services🪖

Let's say, you deployed a pod having a web application running on it and you want to access your application outside the pod. Then how to access the application as an external user?

The k8s cluster setup in your local machine has an IP address similar to your system IP but the pod has a different IP address which is inside the node.

As the pod and your system share different addresses, there is no way to access the application directly into the system.

So we need something in-between to fill the gap and gives us access to web application directly from our systems.

Getting started with Kubernetes Services - Spectro Cloud

This is where Kubernetes Services comes into the picture.

k8s services activate the communication between several components inside-out the application.
It helps us to connect the application together with other applications and users.
It is an object just like PODs, Replicas and Deployments.
It listens to a port on the node and forwards the requests to the port where the application is running inside the pod.

A simple service yaml file:

apiVersion: v1
kind: Service
metadata:
  name: demo-service
spec:
  selector:
    app.kubernetes.io/name: MyApp
  ports:
    - protocol: TCP
      port: 80
      targetPort: 9376

Services Types📑:

NodePort Service
ClusterIP Service
LoadBalancer Service

1) NodePort Service - It makes an internal POD accessible to the Port on the Node.

the port number inside the pod is the target port
the port number in the service area is the 2nd Port
and the port on the node is NodePort where we're going to access the web server externally.
the valid range of NodePort (by default): 3000 to 32767
labels & selectors are necessary to describe in the spec section in case of multiple pods in a node or multiple nodes in a cluster.

2) ClusterIP Service - It creates a virtual IP inside the cluster to enable communication between different services.

a service created of a tier of your application like backend or frontend which helps to group all the pods of a tier and provide a single interface for other pods to access this service.

3) LoadBalancer Service - It exposes the service on a publicly accessible IP address in a supported cloud provider.

A demo NodePort yaml file:

apiVersion: v1
kind: Service
metadata:
  name: my-demo-service
spec:
  type: NodePort
  selector:
    run: nginx
  ports:
    - port: 80
      targetPort: 80
      nodePort: 30007

now run,

kubectl create -f myservice.yaml

This is how you can create your service of NodePort type. By changing the spec type to ClusterIP or LoadBalancer, whatever you wish to work with just change it with the name and port according to it.

CNI (Computer Network Interface)🌐

CNI is a set of standards that defines how to configure networking challenges in a container runtime environment like Docker and Kubernetes.
It is a simple plugin-based architecture.
It defines how the plugin should be developed.
Plugins are responsible for configuring the network interface.
CNI comes with a set of supported plugins already. Like bridge, VLAN, IPVLAN, MACVLAN.
Docker does not implement CNI. It has its own set of standards known as CNM ie, Container Network Model.
We cannot run a docker container to specify the network plugin to use CNI.
But you can do it manually by creating a docker container without any network configuration

docker run --network=none nginx
then invoke the bridge plugin by yourself.

CNI in k8s🕸️

Kubernetes is responsible for creating container network namespaces
Attaching those namespaces to the right network
You can see the network plugins set to CNI by running the below command

ps -aux | grep kubelet

The CNI plugin is configured in the kubelet service on each node in the cluster
The CNI bin directory has all the supported CNI plugins as executables.

ls /opt/cni/bin

k8s supports various CNI plugins like Calico, Weave Net, Flannel, DHCP and many more.
You can identify which plugin is currently used with the help of this command

ls /etc/cni/net.d

kubelet finds out which plugin will be used.

DNS in k8s📡

Kubernetes deploys a built-in DNS server by default when you set up a cluster.
Let's say we have three nodes k8s cluster having pods and services deployed in it having node name and IP address assigned to each one.
All pods and services can connect to each other using their IPs and to make the web app available to external users we have services defined on them.
Every new entry of a pod goes into the DNS server record by

cat >> /etc/resolv.conf

The k8s DNS keeps the record of the created services and maps the service name to the IP address. So anyone can reach by the service name itself.
DNS implemented by Kubernetes was known as kube-dns and later versions CoreDNS.
The CoreDNS server is deployed as a POD in the kube-system namespace in the k8s cluster.
To look for DNS Pods, run the command

kubectl get pods -n kube-system

Ingress🔵

Let's say you have an application deployed in the k8s cluster with a database pod having required services for external accessing with the NodePort in it. The replicas are going up and down as per the demand of the application.
You configured the DNS server so that anyone can access the web app by typing their name instead of IP address every time
like, http://my-app.com:port instead of http://<node-ip>:port and if you do not want others to type port number also then you served a proxy-server layer between your cluster and DNS server. so that anyone can access it by just
http://my-app.com
Now, you want some new features to be added to your application. You developed the code and deployed the webapp again for the new features of pods and services into the cluster. Again setting up the proxy server in-between to access the webapp in just a single domain.
Maintaining all these outside the cluster is a tedious task to do when your application scales. Every time a new feature adds on, you have to do all the layerings again and again.
This is where Ingress comes in. Ingress helps users to access your application with a single URL.
Ingress is just another k8s definition file inside the cluster
You have to expose it one time to make it accessible outside the cluster either with NodePort or the cloud provider like GCP which uses the LoadBalancer service.

Ingress Controller🔹

In the case of using ingress in a k8s cluster, an ingress controller must be deployed and configured.
Mostly used ingress controllers are Nginx, Traefik, and Istio.
Deploy the application on the k8s cluster and configure them to route traffic to other services.
The configuration involves defining URL Routes, Configuring SSL certificates etc. This set of rules to configure ingress is called an Ingress Resource.
Once the controller is running, resources can be created next and configured to route traffic to different services in the cluster.
It is created using definition files like the previous ones we created for Pods, Services and Deployments.
Ingress Controller is not built-in in them by default. You have to create it manually.
A simple ingress yaml file for the sake of explanation:

      apiVersion: networking.k8s.io/v1
      kind: Ingress
      metadata:
        name: minimal-ingress
        annotations
          nginx.ingress.kubernetes.io/rewrite-target: /
      spec:
        ingressClassName: nginx-example
        rules:
        - http:
            paths:
            - path: /testpath
              pathType: Prefix
              backend:
                service:
                  name: test
                  port:
                    number: 80

Create a single ingress called 'simple' that directs requests to foo.com/bar to svc # svc1:8080 with a tls secret "my-cert"

      kubectl create ingress simple --rule="foo.com/bar=svc1:8080,tls=my-cert"

This is how you can start working with Ingress in Kubernetes.

Thankyou!🖤

Kubernetes Networking