5G, a converged network for multiple application scenarios, will revolutionize the internet of things, bringing unprecedented experience and new business modes. It will enable enhanced mobile broadband (eMBB), ultra-reliable low latency communication (uRLLC) and massive machine-type communications (mMTC) applications. It is predicted that 5G will grow mobile data traffic by 500–1000 fold and typical user data rates by 10–100 fold, increase peak transport rates to 10 Gbps or above, shorten end-to-end latency by 5–10 fold, and improve network efficiency by 1000 fold when compared to 4G. The high performance requirements of 5G are driving reconfiguration and innovation in RAN, core and transport networks.
In the architecture of 5G transport network (Fig. 1), cloud-based core network, C/U separation, and data-plane slicing for distributed deployment will make the network more flat. The RAN architecture will also be reconfigured into into three functional entities: CU, DU and AAU. The transport network is divided into three parts: fronthaul, midhaul and backhaul. 5G will place higher requirements on the transport network in terms of bandwidth, latency, connectivity, reliability and SDN capability openness. The major challenges are:
● Fronthaul and midhaul: 5G fronthaul and midhaul will have high latency requirements for the transport network. Currently, the fronthaul latency budget does not exceed 30 μs, and the midhaul latency budget does not exceed 150 μs. In a traditional packet transport network, the store-and-forward queue scheduling mechanism is adopted and the non-congestion forwarding latency for a single node is 30–100 µs, which can hardly meet the requirements. It is therefore necessary to introduce a new low-latency forwarding technology.
● Backhaul: The bandwidth increases over 10 fold, and the traffic model changes from aggregation into full-mesh. The dual connectivity of 4G and 5G convergence, inter-base station coordination, as well as load balance and multi-homed backup for cloud-based core network deployment make traffic more complicated and dynamic. This requires the transport network provide scalable bandwidth and flexible mesh connection.
● Ultra-high-precision time synchronization: Ultra-short frames, carrier aggregation and coordinated multipoint (CoMP) are introduced in 5G to improve time synchronization accuracy by an order of magnitude from ±1.5 μs in 4G to ±130 ns.
● Network slicing: The core network and RAN adopt SDN/NFV-based cloud slicing architecture. Network slicing is based on different application scenarios, and different functional slices have different requirements for bandwidth, latency, network functions and reliability. The 5G transport network is a part of the 5G end-to-end service path and must meet the needs of different services in multiple scenarios. Moreover, 5G is an open network that can meet application requirements of vertical industries and leasing services. Therefore, the transport network is required to support service separation and independent O&M of sliced 5G networks. Different transport network slices are allocated to different service types, with each transport network slice serving as an independent physical network.
To meet the transport needs, ZTE has innovatively developed Flexhaul—an end-to-end 5G transport solution that integrates L0-to-L3 network functions to provide scalable large bandwidth, low latency and flexible L3-to-edge service addressing. The solution has the following attractions:
● It uses the same equipment for 5G fronthaul, midhaul and backhaul. An end-to-end transport solution can be offered, and the equipment can be flexibly deployed to meet network needs.
● It supports the IP and flat optical network architecture, and also supports FlexE over DWDM to enable multi-wavelength multi-link bandwidth binding for capacity expansion. This can greatly enhance bandwidth scalability and reduce initial network construction costs.
● It adopts innovative FlexE tunnel to expand networking capabilities from the interface level to the network level. FlexE timeslot switching, OAM and fast protection switching are supported for creating virtual slicing networks on the Ethernet. Slicing links are similar to circuit pipes that have ultra-low latency and ultra-low jitter. The bandwidth of slicing links can be flexibly configured according to 5G granularity, and services are strictly separated between different slices.
● It uses segment routing in combination with SDN intelligent traffic engineering to address the needs for ubiquitous flexible connectivity brought by cloud-based 5G core network and base stations and to meet the ubiquitous mesh networking requirement caused by L3 down-shifting to the base station. Segment routing decouples service instances from the network, greatly enhancing network capability and scalability of supporting ubiquitous mesh connectivity. Segment routing can be easily integrated with the SDN technology that can calculate an optimal forwarding path to meet service needs based on network traffic and topology resource conditions. The routing information can be delivered to the source node without controlling other nodes in the forwarding path or exchanging signaling. This greatly improves network control performance.
● It supports the slicing of forwarding plane, control plane and management plane. The forwarding plane uses FlexE tunnel for slicing, with each slice having its own independent topology. According to service requirements, different L2/L3 network protocols can be selected such as segment routing and MPLS -TP. Different slices have their own control and management planes. Through the coordination of slices in the wireless, bearer, and core networks, an end-to-end 5G slicing solution can be provided to meet the multi-scenario multi-tenant application needs of 5G vertical industries.
● It provides end-to-end high-precision time synchronization technologies, including high-precision time sources based on common mode and common view, transport devices supporting high-precision timestamp and phase detection, and network control technology based on intelligent clock time. These technologies can meet application requirements of 5G base stations for new air interfaces, inter-site coordination, and location-based services.
ZTE has made great progress in the R&D of 5G Flexhaul products. At the MWC Shanghai 2017, ZTE released ZXCTN 609—a 5G Flexhaul pre-commercial product, and offered a live demonstration of the industry’s first FlexE tunnel technology. Through the FlexE tunnel, different network slices were created, with services being strictly isolated and having no impact on each other. The minimum forwarding latency for a single node is less than 0.5 μs, and the fast protection switching time is less than 1 ms. These performance indices set new records for Ethernet, meeting the requirements of 5G fronthaul, midhaul and backhaul services for ultra-high-reliability and low-latency transport.
ZTE is promoting the R&D and trials of 5G transport solutions in an all-round manner. As a leader in the 5G era, ZTE will continue to innovate in 5G transport technologies, solutions and equipment development and provide operators with competitive and cost-effective solutions. ZTE has been well prepared for 5G transport innovations.
5G transport, challenges and trends, eMBB, uRLLC, mMTC