Karlstad University
Testbed
The Communication Advanced Research Laboratory Wireless Testbed (CARL-W) is located on and around the Campus of Karlstad University, Karlstad, Sweden, and is part of the NorthStar Testbed. CARL-W is managed by the Computer Science Department at Karlstad University in cooperation with Telia (operator) and Ericsson (HW supplier). The communication infrastructure, built on standard fibre and 10GbE, integrates a communication network with radio terminals (4G, 5G, WiFi and Starlink). The testbed connects several user equipment and technologies like mobile phones, IoT devices, and virtualisation and containerisation solutions (Kubernetes/Docker).
Infrastructure
CARL-W provides both indoor and outdoor coverage. It is equipped with a commercial indoor 5G system (Ericsson Radio Dot), with four radio dots deployed in Karlstad University’s engineering building and one in the nearby Innovation Park building, using 100 MHz in Band n78. Outdoor coverage on and around campus is provided through dedicated resources in the commercial Telia network in Bands n78 and n28.
The testbed provides 5G NSA and SA connectivity, and low-latency capabilities (less than 5 ms one-way delay) through so-called Local BreakOut (LBO) for local User Plane Function (UPF) deployment. The performance has been validated with a series of 5G devices and hardware components, such as Cradlepoint R19001, Miktrotik2, Huawei CPE, Quectel RM5003, and 5G SA-enabled mobile phones (e.g. CrossCal devices). The setup suits capacity-intensive services, low-latency applications, and other demanding use cases (VR/AR, smart industry, etc.).
The layout of the 5G section of CARL-W is depicted in the first figure above. The second figure shows some of the areas at Karlstad University covered by the indoor 5G section of the testbed. These include a large experimental space where various use cases can be staged.
Services
Virtualisation, Management and Orchestration
Kubernetes performs the testbed orchestration. CARL-W utilises specific descriptors to describe the service requirements (e.g., network technology, service type): testbed worker nodes are divided by the type of network technology they use, but the testbed is technology agnostic in the sense that it does not mandate a particular connection technology. However, all devices that should be part of the network orchestration need to support Wireguard, Kubernetes and have a routable IP address (i.e., be able to reach the Internet). The Kubernetes infrastructure is shown in the figure to the right.
KAU Testbed:
Example of visualisation of monitored data in CARL-W
Monitoring
The testbed has both infrastructure and service monitoring capabilities:
- We monitor the infrastructure by collecting data from different network technology endpoints (e.g., 5G, Starlink) and the status of the compute infrastructure components, e.g., CPU and Memory load. The monitoring of the computing infrastructure is based on the DESK monitoring framework and utilises a set of open-source components.
- Application and service performance monitoring evaluates network resources, service status, etc. Automated performance monitoring includes continuous monitoring of capacity and latency performance. Some eBPF-based tools for fine-grained latency and inter-packet delay monitoring are also available. These tools are currently started manually when fine-grained monitoring is required.
The monitoring information is collected in InfluxDB2. Part of it is also made available live through a Grafana web interface. The figure to the left displays an example of part of the live visualisation.
Security
No special security capabilities are available. Access to nodes within the KAU infrastructure is available through VPN access.
Remote Access
As mentioned, access to nodes within the KAU infrastructure is available through VPN access. Additionally, the testbed infrastructure is designed to facilitate the seamless integration of external (third-party) sensors, network devices, or services, in addition to supporting containerised services. For external devices or services to integrate with the testbed management system, they must meet specific requirements: they need to support WireGuard, function as a Docker/Kubernetes host, and have an internet connection.
Key Features and Assets
The testbed supports 5G SA as well as 5G NSA, both for indoor and outdoor deployments. Additionally, separate networks and gateways for Starlink and other communication types (such as Wi-Fi) are integrated into the testbed network, providing six possible internet connections: Wired, 5G SA, 5G NSA, LTE, Starlink and Wi-Fi.
The testbed offers support for low-latency services through its local breakout functionality and through RAN and core support for low-latency, low-loss, Scalable throughput (L4S). L4S is a 5G-Advanced Release 18 feature introduced to support XR and other low-latency services over 5G RAN networks.
The testbed offers multiple semi-static network slices that can be selected based on application requirements. Additionally, the testbed supports the CAMARA Quality on Demand (QoD) API. Using the QoD API, applications can activate the best networking configuration that suits better to the needs of the application in real time.
The platform supports the automated deployment of applications through containerisation technologies, such as microservices. The server infrastructure and data centre deploy all virtual machines and network functions running in the cloud, including Kubernetes and Docker Swarm. In addition, the data centre features bare-metal installations for time-critical workloads, along with edge nodes and user equipment (UEs) such as Raspberry Pi, Nvidia Jetson devices and various commercial mobile phones.
The testbed provides support for RedCap, allowing exploration of IoT use cases that do not require the full performance of 5G but instead benefit from reduced complexity and cost.
Testbed Extensions from Open Call #1
MMT (Quanta & Qualia)
The MMT project is set to revolutionise professional training by seamlessly combining virtual reality (VR), Magos haptic gloves of Quanta & Qualia. The MMT project aims to overcome significant limitations faced by current VR haptic gloves and other VR applications. These limitations include challenges in accurately and consistently tracking finger motions in complex and fast-paced scenarios, and to interact via voice for guidance. The primary objective of the MMT project is to provide a more immersive training environment.
Magos gloves, renowned for their high-precision finger tracking, physics-based interactions, and haptic feedback capabilities, combining with 6G/5G connectivity will significantly enhance the collaboration, realism and effectiveness of VR training environments.
The expected result of the MMT extension is to create a highly immersive and realistic training solution that significantly improves trainee skill acquisition and engagement. The solution will be scalable, replicable and adaptable for various professional training scenarios, including healthcare, aviation, industry 4.0, robotics, and gaming.
Targeted Use Cases
- Health, Education and Smart City verticals are accepted.
- All use case topics are accepted. Testbed prefers use cases targeting one or more of the following topics:
- XR, haptics, and other low-latency applications leveraging the MMT extension for immersive media experiences;
- Service differentiation and resource isolation through network slicing and QoD APIs (e.g., CAMARA, NEF);
- Security- and privacy-focused use cases for data protection and trusted communications.