Realizing the promise of 5G – Transport Networks

5G promises to deliver data with peak speeds that are 20 times faster than the current 4G networks, with lower latency and increased capacity. The following table shows how mobile networks have evolved from 1G to 5G:

In addition to higher peak rates and low latency, 5G will support more people and devices connecting at the same time without overloading the network.

ITU has defined three categories for potential 5G use cases:

  • Enhances mobile broadband (eMBB) – eMBB will bring high-speed mobile broadband service to users. It is also being considered for last mile connectivity to provide broadband connection to homes.
  • Massive machine-type communications (mMTC) – 5G will allow connection of millions of low-power sensors, thereby driving the evolution of the Internet of Things (IoT) – Smart Homes/cities/grids, transforming manufacturing processes.
  • Ultra-reliable and low-latency communications (uRRLC) – 5G will enable a new class of applications like autonomous vehicles, robotics etc., that demand very low latency and high reliability connectivity to cloud.

Impact on Transport Networks

5G service requirements place immense demands on Radio Access Networks (RANs). Current transport network architectures and bandwidth are not enough to support 5G rollouts. Several industry-standards groups and carriers have been rethinking existing network architectures to meet requirements of 5G. The objective is to provide the flexibility required to realize the 5G performance targets for different usage scenarios. Here we will cover some key aspects that have an impact on transport networks design in 5G:

  • Programmability of transport networks
  • Higher bandwidth
  • Network slicing
  • RAN functional decomposition
  • Synchronization requirements in RAN

Programmability of transport networks is essential to supporting highly variable bandwidth requirements of future RANs. SDN principles could be exploited to reconfigure switching hardware or VNFs on commodity hardware platforms. This offers flexibility to configure aspects of the network on the fly, like managing slices or managing bandwidth etc., to support end-to-end service goals.

Support for higher bandwidth is one of the first aspects being addressed in preparation for 5G networks. The 1Gbps interfaces in existing cell site routers are no longer enough to support 5G cells that could have capacity as high as 20Gbps. New white boxes for cell site routers have 10Gbps interfaces to support 5G data rates and are in the process of being deployed in existing 4G networks.

The intent of network slicing is to partition network and equipment into sub-networks with different properties. Resources for each sub-network are obtained by slicing the resources from the physical network. Though this topic is still the subject of several studies and research, there are some technologies – FlexE, OTN slicing, wavelength slicing, and VPNs – that are being considered for slicing in the transport domain. The final choice will depend on the services being targeted to be served by the network and the type of transport network being installed.

Functional decomposition of the RAN splits RAN functionality into centralized and distributed locations. The split architecture provides flexibility in designing a RAN network that meets the requirements for different services in 5G. While several different functional splits are possible, there are various trade-offs such as performance, cost, complexity, and transport demands that need to be considered before deciding on what kind of decomposition will be deployed. When possible, configurable splits offer the potential to adapt a network architecture to fit particular use cases and requirements, so that it will deliver the desired speed, latency, and throughput. The following figure shows the high-level architecture of a 5G transport network.

The split architecture has created new transport requirements for Fronthaul and Midhaul networks. Initial 5G deployments may roll out with a Distributed RAN (D-RAN) architecture where the Radio Unit (RU), the Distributed Unit (DU), and the Centralized Unit (CU) are co-located, but to support the whole range of services promised by 5G, operators will have to start planning for split architecture (C-RAN or Hybrid). Fronthaul and midhaul networks are the subject of studies being carried out by various groups in the industry, and some key technologies that are at the forefront are:

  • TSN: Time-Sensitive Networking for Fronthaul – IEEE 802.1 CM
  • eCPRI
  • FlexE
  • Radio over Ethernet (RoE) – IEEE 1914.3
  • OTN
  • PON for fronthaul

Cellular base stations require sub-microsecond time accuracy to function, and split architectures and advanced radio features require accurate time synchronization in the RAN transport network. Also, several technologies, like TSN, depend on time synchronization to operate accurately. Providing a timing source at every node may not be a feasible solution due to the cost of installation and increased complexity. Precision Time Protocol (PTP) can transfer timing and synchronization information using L3/L4 packets, which makes it very effective for 5G RAN. Master clock/time reference can be installed closer to the backhaul, and timing/sync information from it can be distributed to the whole network. ITU has specified a telecom profile for PTP in the ITU G.8271.1/G.8275.1/G.8275.2 standards. These standards cover several different use cases where a new network is being designed or transfer timing/sync is being added onto an existing network that does not support PTP.

With our custom software development expertise, IP Infusion Innovations enables clients to realize the promise of 5G networks by increasing speed, reducing latency and increasing capacity in the Radio Access Network.