Two key requirements for many industrial use cases are transmission time (also referred to as latency) and message reliability. As described in the subtopic “Performance Testing Parameters” {link to subtopic “Performance testing parameters” for topic “Performance testing…” here}, transmission time refers to the time interval between the transmission of the first byte of a message at the source and the arrival of the last byte at the target. Packet reliability refers to the percentage of messages whose transmission time falls within a specified limit.
The arrival of packets transmitted or received through the 5G system is not deterministic but follows a statistical distribution and may include outliers, as explained in the subtopic “Transmission Time Problems and Outliers in 5G Networks: Contributions, Outliers, and Considerations for End Users” {insert link to subtopic “Transmission Time Problems and Outliers in 5G Networks: Contributions, Outliers, and Considerations for End Users ” here}. An example of such a distribution is shown in the figure below. Therefore, stating transmission time as a single fixed value is only accurate if it accounts for 100% of all packets, including all outliers. For this reason, transmission time should instead be expressed as the time within which a certain percentage of packets arrive. Typically, the closer this percentage is to 100%, the higher the corresponding transmission time. For example, 99.9% of packets might arrive within 6 ms, whereas 99.999% might arrive within 10 ms.
This relationship between transmission time and reliability must be measured for each specific combination of a 5G system and user equipment (UE), as well as for specific packet sizes in both uplink and downlink directions. Conversely, if an application requires a certain level of reliability, this relationship can help determine the maximum latency that can be guaranteed for that packet reliability.
Figure 1: Two examples of transmission time shown as a histogram with respect to the number of messages, where the 95th percentile (P95) is indicated, i.e., 95% of the messages have a transmission time below this value. In both a) and b), UDP packets are used for testing. Scenario a) shows uplink transmission of 76-byte packets, while scenario b) shows downlink transmission of 66-byte packets. Different 5G systems were used in each case. (Source: ifak)
As illustrated in the figure above, it is important to emphasize that transmission time always includes some degree of jitter. Even if the distribution is bounded by a value, e.g., 10 ms, this does not mean every message arrives exactly at 10 ms. Instead, arrival times may range anywhere between, for instance, 3 ms and 10 ms. This variation can pose challenges if the application in question is sensitive to jitter.
