Previously, we investigated a recent paper presented to the 2024 Joint European Conference on Networks and Communications (EuCNC), Experimentation-as-a-Service for Validating 5G Use Cases in a Large-Scale 5G Platform.

In addition to discussing EaaS itself, the paper also highlighted a second important consideration for operators, Quality of Service (QoS), presenting a test use case (UC) to underline its importance. That’s what we’ll consider in this blog. Read on…

In 5G Networks, the ability to support variable Quality of Service is an essential requirement for market success. In this blog, we’ll investigate why that’s the case, why operators need to be able to change QoS on a granular basis, and as a result the importance of having access to a framework that enables testing of desired scenarios.

Why is this so important? Because to meet the evolving service demands of their customers (both enterprise and consumer), QoS must be assured. The recently presented EuCNC paper on Validating 5G Use Cases in a Large-Scale 5G Platform makes this point using a specific (if hypothetical) use case which we’ll now investigate.

Background: What’s driving the QoS issue?

The use case in question is drawn from the media industry, one in which use cases typically demand the provision of high bandwidth, ultra-low latency capabilities, and access to massive connectivity. With access to these, advanced use cases are arguably somewhat more common than they are in other domains, for example those involving real-time camera control, remote and live production, use of high-definition cameras and drones, as well as supporting immersive experiences.

Common requirements

Such use cases share fundamental underlying support requirements, perhaps foremost among which is high uplink (UL) throughput in 5G networks. Live camera feeds, immersive VR/AR content and high-resolution drone footage are all reliant on high upload speeds to ensure real-time transmission without compromising quality or experiencing lag, so the right levels of QoS must be assured.

Enabling the necessary UL throughput isn’t, however, straightforward. It demands diverse and complex approaches, for example involving Time Division Duplexing (TDD) within a frame structure. TDD is necessary because it enables the flexible allocation of resources for both uplink and downlink – for instance by prioritizing the former while maintaining the latter.

UL is also challenging because media production typically involves multiple cameras, some of which are on air at any given time and some of which may not be. Since UL capacity is likely to be limited, the ability to prioritize the quality of specific streams (most obviously for on-air cameras as opposed to those in preview) is vital. Doing so requires the network element to expose QoS to an application function with control of the cameras.

Testing QoS

To address and assure Quality of Service in such a use case, the EuCNC paper proposes using a testing environment in which, to meet the necessary (high) UL throughput requirements, two key parameters will be configured:

• TDD frame structure
• Multiple-Input Multiple-Output (MIMO) layers

In its test, DL:UL ratios are configured based on the use case’s specific requirements, with the network tailored to prioritize the crucial UL data flow needed by the applications involved in service delivery.

The results are collected in a physical gNB connected to the multi-vendor SA CN.  The test was able to demonstrate that TDD frame structure heavily impacts throughput performance, from around 130 Mbps up to 175 Mbps. So QoS control directly impacts measurable outcomes by enabling smoother operations and higher-quality content transmission.

The air interface requirements will correspond to requirements in the core network user plane for UPF performance, for both uplink and downlink — which can be verified with Emblasoft Evolver. In this use case, it is critical that the UPF can load balance the GTP traffic (UDP streams) in both uplink and downlink flows. Typically, the UPF is optimised mainly for downlink performance, but here it is critical to verify that of the uplink too.

This UC is, of course, just one example in one domain that demonstrates the importance of supporting variable Quality of Service but it’s nevertheless instructive in demonstrating the impact of the function in a rapidly involving services landscape. Think, even casually, about the direction in which media workflows alone are evolving – for instance towards more immersive and dynamic remote production experiences – and it’s easy to see why QoS and resource allocation are becoming ever more central to enabling the demands of a range of other, emergent use cases.

QoS Control through NEF

A second aspect of testing outlined in the paper focused on Quality-of-Service control realised through the network exposure function (NEF). In this test, to prioritize the quality of on-air camera streams for a hypothetical remote Media production, an NEF was added to an experimental network slice, after which functional testing was run.

In the testing scenario, the NEF exposed 5G network services and specific events (e.g., UE location, analytics on UE behaviour, connection status, etc.) to the required 3rd party Application Functions.

The latter would then leverage the resulting information to trigger actions including changing the QoS of a session or redirecting the traffic so, for example, an AF could change the QoS of an active PDU session on the fly.  After the successful completion of simulations, tests were made using a physical UE and gNB, in which the actual QoS changes could be observed.

Emblasoft Evolver can automate this kind of integration test by executing a performance test before the QoS/PDU session modification is executed. Evolver then sends an API call to the NEF, requesting PDU session modification. A similar performance test is carried out after the QoS change to verify that the performance has changed according to the requested modification.

This can be achieved from a single Evolver test scenario, thanks to the multi-protocol support for N1/N2, N3 (NAS/NGAP, GTP), as well as SBI HTTP/2 Json. In such a scenario Evolver simulates both the UE CP and UP traffic, as well as the AF SBI traffic, greatly simplifying test case automation.

QoS: a broad necessity

For more detailed insights into the technical aspects of the testing undertaken, it’s worth reading the full EuCNC paper which you can find here. Meanwhile, it’s important to grasp that it’s the ability to test Quality of Service that verifies the adaptability of the 5G network to cater to the diverse and evolving needs of (in this case) the Media industry but also others that you seek to serve.

Quality of Service and Emblasoft

The underlying component used in the QoS testing described was Evolver from Emblasoft, a service assurance, validation, and testing solution for any mobile network services (both 5G and legacy).

Evolver provides automated testing and monitoring for both control and user plane applications in 5G networks, validating both services and quality assurance while supporting centralised control of all network testing and active monitoring programmes.

Emblasoft is a global provider of Service Enablement, Active Monitoring, Load and Functional Test solutions for, 5G, VoLTE, 4G, 3G and IMS infrastructure. With Emblasoft’s solutions, operators and equipment vendors can obtain and deliver new products and services that push the boundaries of technology.

If you’d like to discuss testing variable Quality of Service in more details and learn more about Evolver from Emblasoft, please get in touch with us.