5G is poised to transform the connectivity landscape with advancements including multi-Gbps peak data speeds, ultra-low latency, enhanced reliability and scalable network capacity. However, these advancements come at a cost. In this scenario, network function virtualisation (NFV) network architecture that uses IT virtualisation technologies is the key. NFV aims to virtualise network node functions that interface with each other to support communication services.
Key NFV use cases in the 5G context
The foundational benefits of 5G such as architectural flexibility, programmability, reliability and scalability are realised through the successful implementation of NFV. NFV alone is not sufficient, but when it is coupled with cloud computing and software-defined networking (SDN), several use cases emerge that support 5G from a technology perspective as well as from the total cost of ownership (TCO) perspective.
Network slicing enables operators to separate physical network hardware into multiple virtual networks, referred to as network slices. Each network slice can be tailored to cater to end user requirements such as providing access to an MVNO or serving an enterprise customer. Furthermore, each slice can be allocated customised resources to achieve the required calibrated user experience. Use cases such as healthcare would rely more on reliability than throughput. Other use cases such as augmented reality and online gaming would rely on high throughput and low latency. This flexibility to be able to allocate available resources as per user experience requirements helps in using network resources more efficiently. Also, the ability to provide a wide range of user experiences creates several 5G commercial use cases.
As highlighted earlier, 5G networks will result in a large number of radio base stations, increasing network capital and operational costs. Telecom operators can address this issue to some extent by leveraging vRAN, which has deep roots in NFV. vRAN enables virtualisation of radio hardware, enabling operators to create a pool of virtualised baseband units that are hosted in consolidated hardware infrastructure as opposed to being fragmented at actual site locations, which is the case currently. This will help significantly reduce capital expenditure on radio base stations. It will also save operational costs with reduced energy consumption and rental expenses, and optimised preventive and corrective site visits.
Mobile Edge computing
Edge computing has opened up exciting possibilities in the mobile network space, providing storage and processing capabilities much closer to the end user, that is, at the radio layer. In a 5G scenario, large volumes of data will be created by end users and devices. The need for low latency means that we will not have the luxury of hauling this data to the core for processing anymore. NFV creates architectural flexibility, allowing operators to allocate storage and processing resources at the edge so that mobile edge computing can be accommodated at the radio layer.
5G networks will differ significantly from previous generations in terms of the volume of data that will need to be supported. Furthermore, the dynamic nature of traffic flow will be in contrast to the previous upstream and downstream directions of traffic flow for a given user session. This will require resource configuration to be agile and flexible. The dynamic flow of traffic is made possible in part due to SDN, which allows the creation of virtual networking devices (a router for example). In legacy network systems, these changes would be done manually, taking a considerable amount of time, but with SDN, these changes can be automated. SDN is also complementary to NFV, contributing significantly to the virtualisation of network functions and devices that run on commodity hardware.
One of the key upsides of NFV is the ability to automate operations through the management and orchestration layer. Auto healing can be used for recovering a service when there is a hardware fault. The impacted virtual device is recreated immediately on another host, making sure that the application and the service it supports continue to run, providing an uninterrupted user experience. Another key use case from an automation perspective is auto scaling. By defining thresholds in the system, the assigned capacity for an application can be increased automatically when the threshold has been reached, typically at busy hour or when there is an unexpected increase in traffic. These use cases improve reliability while keeping the TCO low by avoiding manual intervention.
Operators need to balance the monetisation potential of 5G against costs and NFV is an integral foundational block to support operators in capitalising the scale, data and automation that come with a thriving 5G ecosystem.