ASTON
Testbed
The ASTON Testbed is located in Aston University, Birmingham, England, United Kingdom. The testbed is managed and operated by the ASTON CyberHub Research Center for Cyber Security, Privacy and Trust.
Figure 1. UWS (ASTON) Testbed: Overview of infrastructure
Infrastructure
shows the topology for the ASTON infrastructure, consisting of several layers, including cloud-managed infrastructure (edge and core), Transport, Management and control, and Interdomain (Access layer) for Internet access outside the university. It is noted that edge nodes are deployed in the infrastructure, which will be further developed to support the demanding XR application in this use case. RAN capabilities are subject to upgrading to offer more reliable and capable performance in line with beyond 5G (towards 6G) expectations.
The testbed comprises two distinct sets of machines: those dedicated to managing the infrastructure and those used for managing the infrastructure itself. Two different 5G software are installed: “Open Air Interface 4G RAN” and “Open Air Interface 5G RAN”, with versions 2023.w42 and 2023.w42, respectively. Only one can be instantiated at a given time, so 4G or 5G networks can run simultaneously.
In one of the “Core” machines, there are four VM machines:
- 1 VM for the complete 4G Core, using the OpenAir-CN version develop branch.
- 1 VM for the complete 5G Core, using OpenAir-CN version v1.4.0.
- 1 VM for the exposition of RAN network slicing capabilities in 4G networks, using the FlexRAN version[1]
- 1 VM for the exposition of RAN network slicing capabilities in 5G networks, using FlexRIC version 8534bcbb.
When both the Linux Container (LXC) 5G RAN and Kernel-based Virtual Machine (KVM) VM 5G Core are fully operational, the complete 5G infrastructure becomes available. The setup includes prepopulated cryptographic information within five Subscriber Identity Module (SIM) cards located on our premises for both 4G and 5G Core networks.
The infrastructure is further supported by two UE devices:
- 1 UE: OPPO RENO6 5G
- 1 UE: OnePlus 8T
These mobile phones are connected to the Ettus Universal Software Radio Peripheral (USRP) B210 when the RAN and CORE are linked. They receive the necessary configuration to access both any VM deployed within our infrastructure and the Internet.
All machines in the setup are Dell T5810 models, each equipped with 12 cores (24 hyper-threaded), 2 TB HDD (Hard Disk Drive), and 32 GB of RAM (Random Access Memory).
There are five physical machines designated as Compute (Nova) nodes for deploying services within the infrastructure. Four of these machines use KVM as the hypervisor, while the fifth employs LXC as the hypervisor.
The LXC physical machine is logically assigned to the zone labelled “Edge 1”. Another physical machine running KVM is assigned to the zone “Edge 2”, and the remaining three machines are assigned to the “Core” zone. These zone assignments are configurable and can be adjusted as needed.
Two machines in the “Core” zone and the machine in the “Edge 2” zone are currently free and available for service deployment.
The LXC Compute machine is connected to an Ettus USRP B210 Software Defined Network Radio Interface via USB (Universal Serial Bus). A live-detection and attachment system for the USB has been built and deployed at our premises to enable USB passthrough to the LXC containers.
All machines are interconnected using a 10 Gb network, serving as the data plane.
Services
Figure 25 shows the services offered in the testbed. The system comprises three physical management machines: “Controller,” “David,” and “Goliah.”
Controller: This machine operates exclusively as a virtualisation layer using KVM, facilitating the creation of multiple virtual machines for management purposes. It hosts four distinct virtual machines:
- MaaS Controller VM: Equipped with Ubuntu Metal-as-a-Service (MAAS) version 2.9.3 and its MAAS client. This VM manages the physical infrastructure.
- Juju Controller VM: This VM runs Ubuntu Juju Controller version 2.9.43 and manages the services within the infrastructure.
- 6G-PATH Advanced Services: This VM contains all components necessary to deliver the advanced features available in 6G-PATH.
- 6G-PATH Controller: This VM hosts the components required for the 6G-PATH dashboard and provides access to the testbed.
David and Goliah: These machines are configured with all necessary services for operating a high-availability OpenStack infrastructure (Wallaby 5.6.0 release). They support advanced capabilities such as OVN networking, Vault for encryption of the management plane and high availability with RabbitMQ and MySQL quorums. Services on these machines are deployed in LXC containers, ensuring a microservice architecture where each container is isolated and dedicated to a single function.
Virtualisation, Management and Orchestration
Primarily, VMs are used in the testbed for hosting services. OpenStack manages those VMs and new VMs can be created on demand to host additional services.
Figure 25. UWS (ASTON) Testbed: Overview of services
Monitoring
The customised monitoring platform has been deployed to collect metrics related to devices, network ports, services, and network flows.
Security
For enhanced security, the entire testbed is protected behind a Virtual Private Network (VPN). Authorisation and protection mechanisms are in place to enable authorised remote access while blocking any attempt to get into the infrastructure from non-authorised users.
Remote access
Key services are made accessible through VPN capabilities using WireGuard.
Key Features and Assets
The testbed has the following key features:
Metal-as-a-Service, managing physical machines in a cloud-style environment. This feature is well-suited for vertical applications that may require access to bare-metal hardware resources, e.g., businesses leveraging high-performance computing, data analytics, and AI training.
OpenStack, a widely deployed open-source cloud operating system, managing virtual infrastructure including pools of compute, storage, and networking resources through APIs or a dashboard.
Juju, an open-source orchestrator that enables scalable deployment, integration and lifecycle management of applications using charms.
Advanced network and application monitoring capabilities at various levels, including devices, device ports, and application flows.
Initial prototype of advanced end-to-end network slicing, offering fine-grained data plane traffic classification and prioritisation across edge, transport and core networks.
Open Air Interface with FlexRIC RAN slicing support, compliant with ORAN approach.
Testbed Extensions from Open Call #1
6G-RESCUE (Newcastle University)
6G is expected to bring transformative improvements, significantly benefiting Edge MLOps (Edge Machine Learning Operations) for always-on battery-powered edge devices by enhancing customised hardware awareness, model deployment, real-time data processing, and distributed machine learning workflows. The project provides a software extension to the 6G for edge ML models training, management and deployment, which will be validated in a smart robot use case scenario that extensively leverages 6G functionalities.
- Improved capabilities of end-to-end network slicing.
- Integration with 6G-PATH experimentation facilities underway.
Targeted Use Cases
- All verticals are accepted
- All use case topics are accepted. Testbed prefers use cases targeting one or more of the following topics:
- eHealth and medical applications enhancing communication, diagnostics, and care delivery through advanced networking;
- Assisted living and remote healthcare support for vulnerable or ageing populations using sensing and connectivity solutions;
- Edge AI use cases improving service quality and user experience through edge-cloud continuum orchestration;
- Intelligent robotics and human-robot collaboration for smart manufacturing, agriculture, and other industrial domains;
- Swarm robotics leveraging B5G/6G for coordinated, distributed intelligence in dynamic environments.