Prerequisites

• A Kubernetes Cluster
• Be the Admin of the cluster, with access to the default Service Account.

Prerequisites for Testing as a Standalone Component

• All the above Prerequisites
• Synthetic Data Generator Installed (Mandatory) [1]

Description

The code in this repository is using all the different Images of the subcomponents from the CODECO PDLC Component. Using the bash files and a Kubernetes Cluster, an administrator can install the PDLC Subcomponents.

Installation

Before you install PDLC you must do the following.

Configure PDLC-DP

Go to the following CRD

• data_preprocessing/pdlc-dp-deployment.yaml

Write the name of the Cluster in the Enviromental Variable named: CLUSTER_NAME

Configure PDLC-RL

Go to the following CRD

• rl_model/rl-model-deployment.yaml

Add the amount of nodes being used in the Environmental Variable named: NODE_NUMBER.

Specify the Node Name for Deployment

Specify the node that PDLC is going to be deployed on. Go in the field spec/spec/nodeName in the following files and provide the name of your node:

• data_preprocessing/pdlc-dp-deployment.yaml
• context_awareness/pdlc-ca-deployment.yaml
• gnn_model/gnn_controller.yaml
• gnn_model/gnn_inference.yaml
• rl_model/rl-model-deployment.yaml

Run the Installation Scripts

Run the following commands to use the installation script:

chmod -R 777 apply_yamls.sh
./apply_yamls.sh

How to Uninstall PDLC

Run the following commands to use the deletion script:

chmod -R 777 remove_yamls.sh
./remove_yamls.sh

References

1. Syntetic Data Generator

Authors

• Panagiotis Karamolegkos (UPRC)
• Pepi Paraskevoulakou (UPRC)