The industrial field scenarios, such as metal processing, complex equipment assembly, and logistics bring a great challenge to timely and reliable acquisition of industrial equipment data like control and experience. Stable and reliable operation of field networks is very important for manufacturing enterprises to ensure stable production and improve the efficiency. Therefore, 5G networks should be capable of running continuously with high reliability for a long time after entering the OT domain of vertical industries. Most intelligent manufacturing facilities require the field network to operate steadily 24 hours a day and have a certain QoS guarantee, because any interruption will seriously affect normal operation of the enterprise.
To ensure end-to-end service determinism of 5G field network, a heterogeneous network can be deployed with redundant data transmission to achieve end-to-end high service reliability. Currently, there are two ways to enhance the reliability of data transmission: packet data convergence protocol (PDCP), and frame replication and elimination for reliability (FRER).
PDCP is used for redundant transmission of industrial data on different radio carriers. It supports carrier aggregation and dual-connection mode to reduce packet loss on the air interface caused by radio interference and ensure data integrity and timeliness.
FRER supports two modes: single-terminal dual-PDU sessions and dual-terminal FRER. Data is sent in redundancy mode through two independent user-plane bearer channels. The sender replicates data, and the receiver de-duplicates and reassembles data. This reduces service damage caused by packet loss in the transmission and core network and improves the reliability of end-to-end data transmission. Based on the reliability of existing terminals, ZTE uses its self-developed SmartEdge (SE) and NodeEngine (NE) to propose the SE+NE FRER solution. Through two independent radio links, FRER achieves optimal data receiving and forwarding at the convergence side. This not only avoids retransmission and even packet loss caused by occasional interference from the radio link on the air interface, but also avoids service interruption caused by sudden abnormality of a single UE. This solution can improve end-to-end service reliability.
FRER is applicable to control business in steel, port and mine with complex radio environments and high network reliability requirements. According to customer requirements and deployment patterns, ZTE has proposed three modes for FRER networking.
In the SE+NE FRER mode, uplink data is replicated through the gateway SE9102, transmitted separately through two 5G modules, and then converged, selective-received and forwarded at the NE side. After being duplicated on the NE side, downlink data is received, de-duplicated and forwarded by two modules of the gateway SE9102, as shown in Fig. 1.
In the unmanned driving project of a steel hot rolling factory, multiple dedicated PADs are deployed to build an independent inter-frequency physical dual-network in the 3.5 GHz and 2.1 GHz bands. The SE gateway is deployed for each travelling crane, and access control (one channel of PLC packets) and video data (two channels of HD IP cameras) are used together with NodeEngine to establish inter-frequency FRER channels. With the dedicated slices configured for control data and the EdgeQoS guarantee of NodeEngine, the delay effect of control data on the FRER channels reaches 20 ms@99.99%. Compared with a single-frequency network, the reliability of the inter-frequency dual-network is enhanced from 99.99% to 99.9999%, which helps enterprises achieve unmanned on-site driving.
In the SE+MEP FRER mode, the FRER gateway function establishes tunnels with UPFs through internal ebridge modules of the MEC platform (MEP). The uplink data is duplicated through the gateway SE9102 and sent respectively through the two 5G modules. Convergence and selective forwarding are implemented at the MEP side. After being duplicated on the MEP side, downlink data is received, de-duplicated and forwarded by the two modules of the industrial gateway SE9102.
ZTE has participated in building a 5G smart factory for a large steel company in China. To extend coverage of the 5G private network, control-plane and user-plane NEs of the core network i5GC such as AMF, SMF, UDM and
UPF are moved down to the branch area of the steel company. This meets the needs of 5G network application of the production control system such as smart molten iron transport, unmanned transport of finished hot-rolling products, driving positioning of travelling crane and intelligent warehouse management. In addition, a private dual-band wireless network (2.6 GHz+4.9 GHz) is built in combination with the SE+MEP FRER function to ensure that user services can satisfy network needs of high reliability and low latency.
In the above two networking modes, NEs or MEC servers need to be deployed on the network side, which increases the deployment cost to some extent. Based on existing 5G network nodes, FRER can be built in UPF in the future, and the UPF+hot standby function will be implemented. When the active UPF fails, traffic will be smoothly switched over to the hot standby UPF, and UEs do not need to be offline. This solution is highly integrated, reduces external dependency, and decreases deployment costs. It also reduces the number of network nodes, simplifies the network and improves network reliability, which is applicable to the scenario where UPFs are moved to edge in the campus network.
FRER is one of the important solutions to ensure service reliability in the field of production control. It can be implemented in actual networks based on heterogeneous network architectures such as 5G+WiFi, 5G+wired, and 5G dual-band. The FRER redundancy transmission mechanism improves the success rate of end-to-end traffic data transmission and ensures the reliability of accurate data transmission and reception. This provides deterministic service guarantees such as delay, reliability and availability in ToB industrial application scenarios. FRER is therefore a complete service-based high-reliability solution.