With the upcoming 5G technology on our doorstep, there is a large variety of new use cases that can deliver a wide range of 5G applications. Each 5G service, according to the characteristics and purpose served, may have specific demands regarding capacity, latency, and reliability.
Although most people relate 5G to super high speeds and massive connectivity among various devices, many 5G use cases require different criteria to serve a large variety of applications that 5G technology has to offer.
5G enhanced Mobile Broadband (eMMB) services need high data rates to deliver superfast speeds to the end user. 5G Massive Machine type communications (MMTC) will create the smart cities of the future, with the evolution of the Internet of Things (IoT) and its ability to provide simultaneous connectivity to billions of devices around the world.
On the other hand, Ultra-reliable, Low-latency Communications (URLLC) applications are the 5G service area with the most demanding requirements on latency and reliability.
5G latency
End-to-end latency is the amount of time that takes a source device to communicate and reach its destination and the response time back, as perceived by the end user.
Latency in a communications system is affected by the processing delay at physical ports, switches, and routers, the signal propagation over the various transmission media such as fiber and radio links and the queuing delay under network congestion.
Typical latencies of a 3G network are in the range of 100msec, while for a 4G network, in the range of 40 up to 60 msec.
The 5G systems should be able to provide end-to-end latency as low as 1 ms for applications that require ultra-low response times.
5G low-latency services
5G Ultra-reliable, Low-latency Communications need very low latency, high-reliability network to support critical services such as e-Health, remote Surgery, Tactile Internet and Autonomous driving.
It is obvious, that if a car is moving at high speed and another vehicle is approaching towards the same direction, the response time has to be ultra fast in order to avoid an accident. The time that is needed for the system to identify the incoming object and take control of the vehicle needs to be able to meet the levels of human response time.
Another application that requires super-low latency is remote surgery. 5G end-to-end latency has to be less than 1.5msec for the surgeons to be able to perform real-time remote robotic-surgery in the years to come.
Other low-latency application examples are online gaming and advanced AR/VR services. Researches show that people do not experience any feeling of dizziness when the delay between the picture projected and the eye movement is less than 20msec.
Furthermore, studies show that in order to support vehicle-mounted communications in existing 4G networks, < 50 msec latency is required. The latency requirements for assisted driving, in 4.5G Vehicle-to-everything (V2X) according to 3GPP TR 23.785, is < 20 msec. Finally,
according to 3GPP TR 22.886, 5G enhanced V2X for self-driving, need latency levels of < 5 msec.
How is 5G super-low latency achieved?
The answers of 5G super-low latencies lie in the 5G network architecture, and the way traffic is routed between the devices and the end users.
One way to deliver low latency in a 5G network is to shorten the actual end-to-end path. In a 5G standalone system, the communication to the 5G core network will take place in a flexible and software configurable manner.
In a flexible 5G network architecture, the content can be placed closer to the end user, and the application server could be located even within a radio node.
Furthermore, the end-to-end path can be shortened when direct communication between two 5G radio nodes is established. Via a virtual X2 interface, the information, instead of being transported via the core network, could form a direct communication channel between the two radio nodes, or the radio node and another end user terminal (device-to-device communication), thus reducing delay times.
5G Mobile Network Architecture
Network slicing is a network architecture that will enable the 5G network to define virtual routes of forwarding traffic, according to each specific service needs. Software Defined Networks (SDN) together with Network Function Virtualization (NFV) are also techniques to assist this purpose and provide the required flexibility on the 5G networks.
NFV is an architecture where virtual machines share a common platform of physical entities (e.g., servers and nodes) to create virtual functions or building blocks. These blocks, when properly connected, form a specific customized path, according to each service’s unique requirements.
SDN is a software-defined network architecture, that is used to separate the data plane (network elements) from the control plane (central controller). The centralized network controller will be able to route the services over the appropriate service blocks, thus, providing flexibility and scalability in each 5G service configuration and enhanced end-to-end network performance.
5G latency compensation techniques
Mesh network topology: Whenever possible, Mobile Operators should focus on developing a mesh transport network in order to be able to create alternative paths for service routing. The use of an SDN controller could assist the selection of routes with the shortest fiber and radio lengths. The SDN controller, together with MPLS segment routing can form individual paths to differentiate the traversing flows of Low Latency and eMBB services, according to the network peculiarities.
Synchronization: Tight synchronization is needed to achieve high accuracy in 5G networks. A GPS receiver could be used in every network node, but this constitutes an expensive practice and is usually not applicable in a real system environment. In the transport area, where GPS signals are not available, a synchronization protocol for the synchronous communication between the elements is the preferred solution. This kind of protocol is defined by the IEEE 1588v2 standard, based on precision time protocol (ptp), where phase and time synchronization should be supported by all transport devices of the network chain. The ITU-T G.8275.1 is an enhanced ptp telecommunications profile for phase and time synchronization, that provides a uniform solution for all vendor devices across the network.
Quality of Service: Quality of Service and Hierarchical-QoS techniques are needed to ensure the forwarding of critical and high priority services even in cases of network congestion. Traffic policing can also be used to minimize the risk that unexpected traffic patterns will compromise the latency of high-priority traffic.
L3 to the edge: L3 cell site aggregator can be used to create L3 end-to-end services with more flexible and dynamic routing capabilities. Physical interfaces: Optical interfaces should be used for better synchronization, while bidirectional SFPs could minimize instability between the transmitter and receiver path.