Kubernetes Edge-Cloud Continuum Dataset

This repository contains a collection of synthetically generated datasets that model Kubernetes nodes and pods across the edge-cloud continuum.

Dataset Structure

Each dataset is organized into folders, named according to the following pattern:

<number_of_nodes>_nodes_<target_utilization_percentage>_util_<instance_id>

• <number_of_nodes>: The total number of Kubernetes nodes in the continuum.
• <target_utilization_percentage>: The target utilization of the nodes, expressed as a percentage (e.g., 70 for 70%).
• <instance_id>: A unique identifier for the dataset instance.

Folder Contents

Each folder contains the following files:

• nodes.csv: A CSV file containing information about the nodes, including their resource capacities (CPU, memory, etc.) and metadata. The header comments describe the parameters used to generate the dataset.
• node-latencies.json: A JSON file that provides latency data between nodes within the same cluster (intra-cluster latencies) and between nodes in different clusters (inter-cluster latencies).
• pods.csv: A CSV file containing resource requests and metadata for the pods. Like nodes.csv, the header comments explain the parameters used in dataset generation.
• workload-latency-requests.json: A JSON file specifying the latency requests between different workloads.