FogDEFTKube: An Extended FogDEFT Framework Supporting Kubernetes

Project Overview

FogDEFTKube is an extended framework built upon the OASIS - Topology and Orchestration Specification for Cloud Applications (TOSCA) tailored for fog computing. This framework provides a user-friendly paradigm to model and dynamically deploy fog services remotely and on-demand. FogDEFTKube maintains four essential layers of abstraction:

1. Platform Independence: Utilizes Docker Containerization technology for seamless deployment across different platforms.

2. Interoperability: Leverages Kubernetes features to establish seamless inter-service communication and orchestration.

3. Portability: Incorporates TOSCA, a vendor-neutral modeling language, ensuring flexibility and compatibility with various cloud providers.

Dynamic Deployment Workshop Demo

This repository contains the code for the IntelliClimate case study, showcasing dynamic deployment on fog infrastructure using xOpera orchestrator.

Setup xOpera on your workstation/PC

1. Install xOpera from pip directly.

   sudo pip install opera==0.6.8

   (OR)

2. Install xOpera inside python virtual environment.( xOpera-CLI)

   #Update package information and install required packages:
   sudo apt update
   sudo apt install -y python3-venv python3-wheel python-wheel-common
   
   #Create a directory for xOpera and set up a virtual environment:
   mkdir ~/opera && cd ~/opera
   python3 -m venv .venv && . .venv/bin/activate
   pip install --upgrade pip
   pip install opera==0.6.8

3. By default, xOpera works with user-name : centos. Make sure that opera's username alligns with the fog infrastructure user, you wish to work on (In this case study, it's root user).

   #set the opera to work with root user.
   export OPERA_SSH_USER=root
   #check the opera user.
   echo $OPERA_SSH_USER

NOTE: You can also add this to the bashrc file directly for persistence across the sessions.

Clone the Repository

git clone https://github.com/cloud-and-smart-labs/IntelliClimate_Case-Study_deployment.git
cd IntelliClimate_Case-Study_deployment

Change the inputs.yaml file to have your allocated cluster set of IP addresses.

#Refer this fro getting your cluster set.
**Cluster-set 1**
node_1: 192.168.0.204 # Kubernetes master/control plane
node_2: 192.168.0.233 #kubernetes slave-1
node_3: 192.168.0.75   #kubernetes slave-2 and actuator node(attached with servomortor)
node_4: 192.168.0.251 #Publisher node ( attached with DHT11 sensor)


**Cluster-set 2**
node_1: 192.168.0.23 # Kubernetes master/control plane
node_2: 192.168.0.106 #kubernetes slave-1
node_3: 192.168.0.205  #kubernetes slave-2 and actuator node(attached with servomortor)
node_4: 192.168.0.103 #Publisher node ( attached with DHT11 sensor)


**Cluster-set 3**
node_1: 192.168.0.97 # Kubernetes master/control plane
node_2: 192.168.0.30 #kubernetes slave-1
node_3: 192.168.0.18  #kubernetes slave-2 and actuator node(attached with servomortor)
node_4: 192.168.0.118 #Publisher node ( attached with DHT11 sensor)

Setup Passwordless SSH from workstation/PC to all fog nodes(inputs.yaml)

1. Generate SSH key:

   ssh-keygen

2. Copy SSH key to remote fog infrastructure i,e all nodes specified in inputs.yaml file:

   ssh-copy-id root@fog-node-1_IP
   ssh-copy-id root@fog-node-2_IP
   ssh-copy-id root@fog-node-3_IP
   ssh-copy-id root@fog-node-4_IP

Deploy Services

Use the xOpera CLI commands to deploy services onto fog infrastructure:

1. Validate the deployment:

      opera validate -i inputs.yaml service.yaml

2. Deploy the services

   opera deploy -i inputs.yaml -w 2 service.yaml

3. Undeploy the services

   opera undeploy

NOTE: For more details on xOpera commands and options, visit xOpera-CLI

Observations

After deploying the IntelliClimate case study, observe the following:

On node1 (Master Node)

Check the k3s cluster status:

kubectl get nodes

View kubernetes objects.

kubectl get deployments
kubectl get pods
kubectl get services

On node4

Check the status of mqtt-publisher-service:

systemctl status mqtt-publisher

On node3

Check the status of the actuator service:

systemctl status actuator

Sensor Data Visualization

Install mosquitto-clients on any fog node:

apt-get install mosquitto-clients

Subscribe to sensor/data topic:

mosquitto_sub -h node1_IP -p 30001 -t sensor/data

Here, node1_IP corresponds to the IP of kubernetes master node.

Now, you can observe the sensor-generated values on the mqtt-server’s sensor/data topic.

We had setup the thresholds of 30 and 70 for temperature and humidity respectively. So, whenever the values goes beyond the threshold, you can see the actuator (here the servomotor) performing some work (rotates on to left for opening). Whenever, it feels temperature and humidity values are normal (below thresholds) for 30 secs, rotates towards right for closing.