Infrastructure Configuration
This section explains how to configure the computing infrastructure required for an experiment in Plaza6G, including Virtual Machines, Kubernetes clusters, and Bare Metal servers.
1 Overview
Each Plaza6G experiment is built using Virtual Machines (VMs), Kubernetes Clusters, or Bare Metal servers depending on performance requirements and experiment complexity.
All resources are automatically deployed through Plaza6G’s orchestration backend, powered by OpenStack and Terraform (Tofu). Users do not interact with these tools directly, but benefit from reliable and reproducible deployments.
Every infrastructure element is preconfigured with monitoring agents that feed Grafana dashboards for CPU, memory, disk, and network traffic metrics.
2 Virtual Machines (VMs)
Virtual Machines are the most commonly used component for running network functions, traffic generators, analytics, and core network elements such as Open5GS.
How to Create a VM
- Open Setup Experiment → Infrastructure.
- Select Virtual Machine.
- Fill the configuration fields.
- Click Add Virtual Machine.
Configuration Parameters
| Parameter | Description | Example / Limits |
|---|---|---|
| Name | Unique identifier for the VM. | open5gs-core, dnn-vm |
| Operating System | Select VM OS image. | Ubuntu 20.04 / 22.04 |
| CPU | Number of vCPUs. | 2–8 |
| RAM | Memory per VM. | Up to 32 GB |
| Disk Size | Storage allocation. | 20–200 GB |
| SSH Access | Enable or disable SSH. | Yes / No |
| Internet Access | Allow external connections. | Yes / No |
| Monitoring | Default system metrics. | CPU, Memory, Disk, Traffic |
Example use case: A VM running Open5GS for 5G Core and another acting as a DNN traffic endpoint.
3 Kubernetes Clusters
Kubernetes clusters support distributed or containerized workloads such as scalable 5G cores, microservices, AI pipelines, or network slicing experiments.
How to Create a Cluster
- Open Setup Experiment → Infrastructure → Kubernetes Cluster.
- Set a cluster name and resource parameters.
- Select number of nodes and node sizes.
- Click Add Cluster.
Configuration Parameters
| Parameter | Description | Example |
|---|---|---|
| Cluster Name | Identifier of the cluster. | k8s-slicing |
| Nodes | Number of cluster nodes. | 1–5 |
| vCPUs | CPU per node. | 2–8 |
| RAM | Memory per node. | Up to 32 GB |
| Disk | Storage per node. | 20–200 GB |
| Communication Plugin | Networking backend. | Calico |
| SSH Access | Administrative access. | Enabled |
4 Bare Metal Servers
Bare Metal nodes offer maximum performance with direct hardware access, ideal for real-time computing, high-throughput networking, and experiments requiring dedicated CPUs or fast storage.
How to Request Bare Metal
- Open Setup Experiment → Infrastructure → Bare Metal.
- Select Request Bare Metal Server.
- Enter CPU, RAM, and Storage requirements.
- Submit the request.
Typical Specifications
- Multi-core high-performance CPUs
- Up to 64 GB RAM
- SSD or NVMe storage up to 1 TB
- High-speed Ethernet or InfiniBand
5 Monitoring and Visualization
All infrastructure elements integrate automatically with Prometheus and Grafana. Users can monitor:
- CPU utilization
- Memory consumption
- Disk I/O
- Network throughput
No manual configuration is needed — dashboards are generated automatically.
6 Roadmap Features
- Standalone containers without Kubernetes
- Extended metrics (power, temperature)
- Physical network hardware control
- REST API automation
Summary for Users
- Configure infrastructure in Setup Experiment → Infrastructure.
- VMs, Kubernetes clusters, and Bare Metal resources are available.
- All nodes are automatically monitored via Grafana.
- Provisioning relies on OpenStack and Terraform automation.
Example Q&A
Q: What types of infrastructure can I deploy?
A: Virtual Machines, Kubernetes clusters, and Bare Metal servers depending on experiment requirements.
Q: What metrics are collected automatically?
A: CPU, memory, disk I/O, and network traffic are monitored by default through Grafana.
Q: How do I monitor infrastructure performance?
A: Use the Grafana dashboards available under My Experiments.
Q: Can I deploy standalone containers?
A: Standalone container support is planned in upcoming versions.
Q: What tools manage the deployments internally?
A: OpenStack provisions the resources while Terraform (Tofu) handles declarative orchestration.
Q: What are the resource limits?
A: Up to 8 vCPUs, 32 GB RAM, and 200 GB disk for VMs. Bare Metal may support higher limits.
