backup
March 20, 2024
Comments 0
5G Ultra Reliable Low Latency Communications (URLLC) are designed to meet the demands of availability and latency-sensitive services. In order to support URLLC, 5G mobile networks must have low latency, minimal packet loss, and no packet arrival out of order. The user plane must have a one-way latency of 1 ms according to ITU-R.
We can define URLLC more precisely if we break down the acronym and examine the prerequisites:
With the first 5G release 15, 3GPP took the first steps toward URLLC. 99.999% reliability and 1 ms latency were specified for the New Radio (NR) interface. Nevertheless, the Non-Standalone Architecture (NSA) required the radio signaling and the core network to rely on LTE, which prevented them from meeting the URLLC end-to-end delay requirements.
A new end-to-end 5G architecture called Stand Alone (SA) is defined with the release of 3GPP version 16. It can now function without LTE thanks to its own 5G core, which also provides two significant features: network slicing and mobile edge computing (MEC).
The distance between the server and user equipment as well as network performance are the two main factors that affect end-to-end latency. They were both tailored to fit URLLC uses. Next, we examine 5G technology enablers:
5G New Radio
Flexible numerology, scheduling optimized for low latency, or uplink grant-free transmission have all been used to optimize the air interface for low latency. Enhancements to HARQ, robust control channels, and micro diversity are important for increasing reliability.
The subcarrier spacing can be adjusted from 15 kHz to 240 kHz with a new numerology; larger spacing leads to shorter symbol durations and subsequently shorter scheduling intervals. Because scheduling algorithms can schedule mini-slots, they are able to further reduce transmission latency. Use of uplink grant-fee transmission can prevent delays brought on by transmission resource requests.
By utilizing multiple antennas on both the transmitter and receiver sides, micro diversity prevents single link failure and creates distinct spatial signal propagation paths. Work was done to ensure robust Control Channels with low error rates in order to assure reliability; NR is implementing novel coding and low modulation and coding scheme (MCS) for their transmissions. By allocating resources in advance for retransmissions, the repetition mechanism HARQ is improved to lower latency and boost reliability.
Network Slicing
One essential 5G feature is network slicing, which enables resource allocation on demand for users with various service needs. A flexible partitioning and isolation of resources from the effects of other users results in the creation of an end-to-end logical channel. From the radio interface all the way up to the core network, the QoS demanded by the user slices is configured on demand.
For instance, 5G can produce a low latency slice for URLLC service for robotic control and a high capacity video streaming slice for enhanced Mobile Broadband (eMBB) services without strict latency constraints for the same user. Only the 5G core network’s standalone (SA) architecture is compatible with this feature.
Mobile Edge Computing
As a component of the Cloud-Radio Access Network (C-RAN), Mobile Edge Computing (MEC) hosts user applications at the “Edge-side,” significantly reducing latency and improving reliability. Thus, radio access is primarily to blame for the transmission delay. Reducing the number of nodes on the data path improves reliability and eliminates the need to go through the core network when hosting at the edge.
The end-to-end latency is increased by all network elements, including nodes that introduce processing or queuing delays and wired or wireless transmission pathways. While choosing appropriate components, it is crucial to understand each of these latencies; however, the end-to-end latency is the only one that demonstrates how these components function together. Therefore, it is crucial to measure the network’s end-to-end latency both before and after deployment.
Possible setups for end-to-end latency measurement
When using near-end measurement, packets are sent into the network being tested from one port on the instrument, routed back to the second port, and simultaneously reversed to test the opposite direction. Since there is no need to synchronize two instruments and the path ends in a single device, this setup is simpler. However, it is more difficult to distinguish between the uplink and downlink latency components and calculate their respective contributions to the total latency.
When conducting a far-end measurement, packets are routed from one instrument to the other from two locations apart. A high accuracy external clock, such as the timing signal included in a Global Navigation Satellite System (GNSS) signal, is required to synchronize the two instruments when they are being used. This offers more flexibility. For instance, the instruments can measure downlink and uplink directions precisely by placing one at the server and the other alongside a user device. Another example are two devices that can be mounted inside of moving cars; when paired with portable user gear, these instruments allow for the measurement of latency while the vehicle is in motion.
It is necessary to measure latency and reliability at various locations in order to comprehend the performance of a network. The performance, timestamp, and location data can then be combined to create a heat map, which is akin to the widely used wireless signal coverage maps. These latency heat maps can be used to assess the real network performance in a given area and identify areas that might be problematic.
As seen in Figure below, measurement devices are installed in cars traveling through an area of interest, and another is situated next to a MEC server. To connect to the 5G network, the cars are outfitted with GNSS receivers and a cellular user equipment (UE), like a smartphone or 5G dongle. For instance, we can describe the circumstances in which MEC applications must operate by assessing the quality of the link between the MEC server and the cars. Alternatively, we can find out if the connection quality is good enough for the vehicles to share sensor data by platooning the vehicles and measuring the link between them.
The all-in-one field tester, MT1000A, is useful for assessing the connection quality of developing 5G URLLC networks. It is capable of assessing quality metrics such as packet loss, jitter, latency, and pattern or sequence errors. With its modular design, the MT1000A supports a variety of technologies including combined 10G/100G Ethernet, OTN, SONET/SDH, OTDR, PTP, and CPRI through its various modules. The software package enables testing automation and is user-friendly and versatile.
The MT1000A employs a custom application protocol to test URLLC latency; end-to-end latency is determined by subtracting the timestamp at packet sending from the timestamp at packet receiving. As seen in Figure below, the precision of the latency measurement ranges from nanoseconds to microseconds, contingent upon the underlying transport technology. This can be compared to ping, the conventional approach that is only accurate for near-end measurements and provides a rough estimate of network latency.
