5G New Radio (NR) originates with a vision of pervasive connectivity, extreme data rates, and low-latency yet highly reliable networks. The international telecommunications union (ITU), working with the international mobile telecommunications (IMT), created the IMT-2020 vision that identifies three primary use cases for 5G NR:

  • Enhanced mobile broadband (eMBB)
  • Ultra-reliable low-latency communications (URLLC)
  • Massive machine-type communications (mMTC)

5G NR Release 15 was frozen in the summer of 2018, but the standards continue to evolve with Release 16 planned the end of 2019. Release 16 further optimizes 5G NR to support new use cases and types of services. Study items include enhancements to the physical layer to support ultra-reliable low latency communications (URLLC) for industrial IoT, extending frequency up to 114 GHz, enhancements on multi-user MIMO, access to unlicensed spectrum, integrated access and backhaul, cellular-vehicle-to-everything (V2X), and user equipment (UE) positioning and power efficiency.

Mobile network operators (MNOs) need to consider a network that can handle the many advances expected in 5G NR. To deliver the lower-latency, higher throughput, and higher network availability promised by 5G, MNOs are utilizing radio access network (RAN) and core architectures based on software-defined networking (SDN) and network function virtualization (NFV). SDNs centralize the network control by decoupling the forward process of the network--the data plane--from the routing part of the network--the control plane enabling centralized network programming. NFV complements SDN by decoupling the hardware from software functions creating even more flexibility.

Network virtualization is critical for 5G NR
NFV is a key enabler to flexible, highly scalable, and cost-efficient networks needed for 5G. It decouples software and hardware functionality instead of cyclically adding or upgrading purpose-built hardware. A virtualized core provides a smaller footprint, reduced costs, support for many features, improved scalability, and dynamic resource allocation. NFV also enables a 5G-ready platform that can implement new architectures such as mobile edge computing (MEC) and network slicing.

Network slicing partitions the network
With NFV, some of the RAN and core network physical infrastructure is replaced with software-based virtual machines (VMs) to segment the physical network into multiple virtual networks that are partitioned into network slices, each with its own capabilities. This allows MNOs to partition the network for specific service requirements. Network slices can be optimized for latency, throughput, security, or other attributes that align to different subscriber requirements. Take, for example, a utility company that needs a service to support thousands of sensors, each sending small amounts of data infrequently; or a hospital performing robotic surgery, which requires high reliability with guaranteed packet delivery.

The partitioning of multiple network "slices" can take place in both the core network and the RAN (Figure 1).

network slicingFigure 1. Network slicing partitions the network to deliver different levels of services.

Virtualized radio access network centralizes data processing. Equally important will be the ability to build and validate a high-performance RAN. 5G NR relies on key technologies such as the use of wider channel bandwidths, carrier aggregation, massive MIMO, and modulation schemes up to 256 QAM to achieve higher throughput in a given cell. This can equate to massive amounts of data that need to be managed cost-effectively.

A virtualized RAN helps manage and optimize the processing of huge amounts of data. It uses a distributed topology where the remote radio units (RRUs) are placed at the tower and baseband signals are transmitted over long distances to the base band units (BBUs) using a low-cost technology such as common public radio interface (CPRI), e-CRPI or O-RAN. The BBU have been moved away from the base station and into a centralized area in the data centers where resource can be optimized (Figure 2).

Distributed RANFigure 2. A distributed RAN centralizes baseband units in data centers.

Some service providers are getting a head start on implementation. For example, Korean service provider SK Telecom (SKT) has implemented a cloud RAN model with Nokia, which is claimed to be the world’s first commercial deployment. This approach scales traffic more effectively and provisions network resources in the cloud according to demand.

Virtualized RANs require higher-bandwidth and lower-latency interfaces from the base stations to the data centers. There's an interesting report by ITU that explains the need. Both the RAN and the core need to be functionally tested and load tested to determine how the infrastructure handles the massive data expected with 5G.

Mobile edge computing enables low-latency services
In contrast, MEC is an emerging architecture that moves the processing, storage, and management to the base station or small cell, putting the computation closer to the RAN's edge for support of low-latency and on-demand services. It positions the computing resources closer to the edge of the network for multiple cell-site baseband processing. The physical proximity of computing resources results in reduced transport latency, which enables smart infrastructure services with sub-millisecond response times for applications that require low-latency and high-reliability communications such as autonomous driving or automated drone traffic control. A centralized network could not support these use cases due to longer round-trip transit times.

In initial 5G rollouts, operators need to balance centralized processing and MEC, selecting the architecture that meets their need to balance ultra-low latency verses cost. There is a trade-off between cost and performance; locating processing and storage in a single central office is significantly less expensive than locating in the field.

Ensure your network is 5G ready
5G is much more than higher speeds. The standard involves new technologies that will require drastic changes in your mobile network. New network features such as SDN and NFV enable networks to adapt more effectively. As an MNO, you can prepare for 5G NR by refreshing or replacing network equipment that can upgrade to gigabit performance and low-latency. Migrating to a cloud-based, virtualized platform also allows a more efficient deployment of network resources and an easier path to future advances in MEC and network slicing. This means depending more on data center functionality and less on purpose-built hardware.

Delivering these highly flexible network architectures requires extensive validation and test of the core infrastructure and RAN. Performing load tests with realistic subscriber traffic to verify data plane and control plane performance, individual network function performance, and overall quality of service (QoS) will be essential. Also, end-to-end testing can be used to validate network performance across a virtualized RAN and into the data centers.

To deliver the ultimate subscriber experience, be first or fast to market with new services, while lowering your network operating costs, ensure you are working with a first-to-5G partner with expertise in end-to-end telecommunications design, test, and measurement.

Sheri DeTomasi is the 5G New Radio Solutions Lead at Keysight Technologies.