In traditional networking, infrastructure such as cables, RAN equipment and core applications deliver connectivity at one performance level and to one set of KPIs.
But network slicing offers something different. Because network hardware can be separated from the software that controls it, network operators can create different ‘slices,’ or portions of the network, and offer these to customers according to how they want them to work.
Each slice is a standalone, independent, and end-to-end mini-network that operates as a single service – with its own network architecture, performance parameters, security policies, and service level agreements. But these different slices can share common network resources. For example, one slice can share RAN, transport, and core resources between customers, while another may use a dedicated network core for a single customer. This makes network slicing an efficient way for operators to sell network capacity, and an essential part of 5G services.
How does network slicing work?
Imagine two customers of a telecommunications network. One wants to provide connectivity to a 100,000-person sporting event that requires huge data availability over a defined area. The other is a drone operator that prioritises ultra-low latency. With network slicing, these two customers – each with very different requirements - can be served over a single network infrastructure by creating separate network slices for each one, and then assigning network resources as and when they are needed.
This diagram from 3GPP, who have been integral in developing network slicing standards, explains how network slices can work across existing physical and virtual network functions:
Credit: 3GPP
Within the context of 5G services, network slicing use cases fall into three main categories:
Enhanced Mobile Broadband (eMBB). This focuses on delivering high connection speed and capacity, for example to deliver alternate reality or virtual reality remotely.
Massive Machine Type Communications (mMTC). This involves connecting lots of devices or objects in a defined area – for example, farm machinery within smart agriculture
Ultra-Reliable Low Latency Communication (URLLC). This prioritises ‘five nines’ (99.999%) lag-free connection – think drone operation, remote surgery, or self-driving cars.
Why is network slicing important?
Dividing networks in this way means a single physical network can be used to deliver multiple services. This benefits both the operators and the users of networks. Operators can use their network resources more efficiently and create new revenue streams, and customers get a flexible and programmable network that is tailored to their needs.
For this reason, network slicing is an integral part of 5G network architecture, which requires exactly this level of flexibility and network programmability.
Network slicing is an important part of telco offerings to enterprises, who want to be in control of how their networks operate and perform. Operators can also offer ‘network slice as a service’ – carving out standalone, end-to-end networks for customers to run themselves.
How can telcos and network operators deliver network slicing?
For network slicing to work, networks need to be virtualised, efficiently managed, and automated as much as possible. This means virtualised network functions (VNFs), a smooth billing process, as much north-south network automation as possible, and software-defined networking (SDN) to manage traffic flows through each slice.
What are some real-world network slicing use cases?
Some network slicing examples include:
In September 2023, Vodafone and Ericsson trialled a dedicated network slice for mobile gaming, which delivered a 270% increase in download performance, a 25% decrease in latency, and 57% less jitter.
In the same month, Telia Norway demonstrated a standalone network slice for the Norwegian Armed Forces, which separated traffic entirely from the rest of the network to prioritise security and flexibility for traffic routing.
