Measuring Latency & Other QoS Parameters Across 5G Networks

Author : Andy Cole, Anritsu

21 June 2023

Figure 1: RTD figures can differ based on the asymmetric path lengths that data traffic passes through
Figure 1: RTD figures can differ based on the asymmetric path lengths that data traffic passes through

As critical applications such as robotic control, self-driving vehicles and remote medical procedures become more reliant on mobile communications for connection, mobile network operators and system designers must guarantee that the electronic decisions are communicated quickly and reliably. Failure to achieve the ultra-reliable low-latency communications (URLLC) promised by 5G could result in catastrophic events and cause injury or worse.

5G mobile communications networks need to support time-critical applications, where machines send information to other machines that must be acted upon immediately. Emerging technologies that use Internet of Things (IoT) mobile connectivity to communicate vital decisions are developing rapidly.

  

The automotive industry is moving towards autonomous driving, where electronic systems take complete responsibility for ensuring a safe journey.  This requires that detailed information is gathered from onboard sensors that operate in a 360o arc around the vehicle, and that information about external conditions, such as road layouts, traffic bottlenecks, roadworks and pedestrians, from roadside video and other sources, is communicated and subsequently responded to in the shortest possible timeframe.  The speed of data transfer to and from the vehicle is therefore of vital importance.

This means that communication networks must be both robust and reliable, plus able to prioritise applications based on how critical they are - network operators need to make certain of ongoing URLLC operation throughout the installation.

As well as ensuring that data is successfully delivered throughout the network, they must also be confident that it will arrive in good time. Quality of Service (QoS) measurements, such as throughput, utilisation, latency, jitter and packet loss, all need to be given proper consideration. However, it must be acknowledged that latency often poses a particularly difficult challenge. 

Figure 2: An Anritsu MT1000A unit measuring simultaneous up-link and down-link latency during a drive test
Figure 2: An Anritsu MT1000A unit measuring simultaneous up-link and down-link latency during a drive test

Latency measurement involves using two test instruments, both of which need to reference the exact same time with a high degree of precision, even when separated by several km.  In the case of a moving vehicle, latency needs to be measured while the test instrument is travelling at high speeds.

Defining delay
Latency is another term for delay.  In telecommunications, it refers to the time it takes for a packet of data to travel from the source where it originates to its assigned destination.  This is the one-way delay, not round-trip delay (RTD), and since the latencies in a network are rarely the same on the forward and return paths, it cannot be assumed that latency is the same as RTD divided by two. 

RTD is often used as an alternative measure to latency.  However, RTD refers to how long it takes for a packet of data to travel through the network, and to be returned to the source by means of some kind of loopback mechanism at the far end. 

 

A ping test is also sometimes used as an alternative, as it is relatively easy and cheap to achieve.  However, ping is a RTD measurement and usually performed without the use of test grade instruments.  While mobile phone apps and PC-based ping results can provide a rough estimation of RTD, this is unreliable and doesn’t consider the critical point that data can take considerably different amounts of time to travel in each direction.  
 
If perfect symmetry of the transmission paths in each direction can be guaranteed, such that delays are the same, then latency can be assumed to be half RTD. However, for many reasons, this is rarely the case. The path lengths can vary if each one-way data traffic is routed differently.  It also needs to be considered that electronic processing in the network and system equipment takes time, and that networks can become congested - thereby causing buffering to occur. These, along with a number of other reasons, can lead to network delays, all of which can affect the latency budget.

Figure 3: Measurement of QoS parameters with regard to the communication link established between network infrastructure and a moving vehicle
Figure 3: Measurement of QoS parameters with regard to the communication link established between network infrastructure and a moving vehicle

Measuring latency 
To measure latency between two possibly widely separated locations requires two measuring instruments, with time clocks that are aligned and synchronised via means of a global navigation satellite system (GNSS), such as GPS.  The accurate time is recorded in each test data packet when it leaves the transmitting test instrument.  The receiving instrument compares this to the precise time of arrival and calculates how long it took for the packet to arrive at its destination - this is the latency of the signal.

As defined by the 3GPP, 5G is a mobile wireless access technology that needs to support time-critical applications. Mobile devices, sometimes referred to as user equipment (UE), establish connections wirelessly with 5G base stations, which in turn provide connections to the wider telecommunications networks. If the UE is travelling, it will move in and out of range of the base station, and connection will be handed over to an adjacent base station. Since the size of each 5G cell may be quite small, this handover process can occur frequently.

This poses the question of how to measure latency between two locations, especially when at least one is moving. Firstly, we must ensure that the reference clocks at both ends of the measurement are synchronised with one another.  The transmission protocol must support Ethernet test frames containing accurate time stamps to be transmitted in both directions. Then, the test equipment needs to be able to record the result alongside its geographic location, in order for its operative to fully understand what latency performance has been achieved.

Figure 4: The down-link latencies in relation to different MNOs
Figure 4: The down-link latencies in relation to different MNOs

An MNO must guarantee reliable geographic service coverage to its subscribers. One way to achieve this is by partnering with other MNOs. Then, if network coverage cannot be achieved from the primary MNO, there is the opportunity to handover to an alternative network and maintain QoS. Therefore, it is important to measure the QoS not just in relation to the primary MNO, but also simultaneously examine that of the supporting MNO. 

Use case example
Successful simultaneous up-link and down-link latency measurement, referenced to GPS locations, can be achieved by using Anritsu test equipment.  

In the example shown above, the two test instruments measure the latency, jitter and all other QoS parameters, between a travelling vehicle and a fixed location in the network. The mobile routers have dedicated test SIMs with access point names (APNs) that support fixed IP addresses.

The clocks of both test instruments have been synchronised to GPS. The instruments also log the GPS position so that the exact location of the moving vehicle is known when the QoS measurements are recorded. The instrumentation is measuring continuously and the maximum, the minimum and calculated average values for each parameter are recorded at 1s intervals. The results files compiled are stored for tester replay and can be exported in .csv format.

  

Correlation of QoS parameter values with the respective GPS locations allows the creation of .kml files that can be viewed in standard mapping software, and values can be associated with colour coding to produce a heat map to easily identify where issues occur. In this example, the test instrument was hand carried while walking around the FIRA Exhibition Centre, Barcelona, in January 2023.  The resulting map shown in the images below displays the results for down-link latency.

Figure 5: Down-link latency mapping
Figure 5: Down-link latency mapping

In addition, the very accurate and extensive set of data acquired in these tests can have other uses. Through it, cumulative distribution function (CDF) and distribution curves can also be created.

Conclusion
To summarise, due to the nature of mission-critical applications that rely on fast and accurate communication of information with a need to act on resulting decisions, latency is fast becoming one of the most important quality measurement parameters of any data communications network. The increasing prevalence of emerging applications, such as autonomous driving and next generation machine-to-machine (M2M) communication, are driving this. To ensure URLLC is maintained, fully understanding latency in 5G networks is essential. It is not sufficient to rely on unqualified RTD measurements with basic ping-related tools, not simply because of the doubtful accuracy and uncertainty associated with ping, but also because networks and other devices incur time delays that vary in each direction. Latency and jitter measurement can only be assured using proper test-grade instrumentation across the network.


Contact Details and Archive...

Print this page | E-mail this page






This website uses cookies primarily for visitor analytics. Certain pages will ask you to fill in contact details to receive additional information. On these pages you have the option of having the site log your details for future visits. Indicating you want the site to remember your details will place a cookie on your device. To view our full cookie policy, please click here. You can also view it at any time by going to our Contact Us page.