Unlike earlier networks, which were designed as monolithic blocks, the 5G system supports “network slicing”. This network architecture enables the multiplexing of independent, virtualized logical networks, known as network slices, over the same physical infrastructure. Each network slice is an isolated, end-to-end network, specifically tailored to meet the requirements of a particular application, with defined capabilities and characteristics. A service provider can make slices available for use by different enterprises or different entities within the same enterprise. By default, a 5G network consists of a single network slice, which includes all 5G system functions and defined quality of service (QoS)  characteristics. One option for deploying a non-public network (NPN) is to use a network slice within a deployed public network. More about this can be found in the topic “5G NPN Deployment Options.”

Network slices are particularly beneficial in industrial ecosystems, allowing the separation of specific processes into different slices with services tailored to varying needs. For example, one network slice could be optimized for services used by smartphones, while another could be dedicated to IoT services supporting time-sensitive networking (TSN) or ultra-reliable low-latency communication (URLLC) features. Network slices can also be used to allocate a predefined amount of resources for specific use cases and services. Essentially, every network slice is capable of supporting all possible services.

The following figures illustrate scenarios where devices connect to different data networks via a multiplexed 5G network with a defined number of slices, packet data unit (PDU) sessions, QoS flows, radio bearers, and slice-specific QoS. According to 3GPP specifications, a particular PDU session is associated with exactly one specific network slice. If a UE needs to connect to more than one network simultaneously, it must establish at least one PDU session for each slice. A network slice can support multiple PDU sessions for each user equipment (UE), and conversely, a UE can support up to 16 concurrent PDU sessions. If sufficient resources are allocated, a network slice can support all of a network’s active PDU sessions.

Figure 1 shows a UE accessing a single network slice with one PDU session and multiple QoS flows. The network slice supports edge cloud-based services. Figure 2 shows a UE accessing a single network slice connected to two data networks—one supporting edge cloud-based services and the other supporting time-sensitive networking (TSN) services. In this case, the UE requests two PDU sessions, one for each service. The final example in Figure 3 shows a UE accessing two network slices concurrently—one supporting Edge Cloud-based services and the other supporting TSN-based services. This may occur if one network slice supports the 5G QoS identifiers (5QIs) required for the Edge Cloud use cases, while the other slice supports those needed for TSN.

For more information on establishing a PDU session within a network slice and a deeper dive into QoS, refer to the white paper “5G QoS in Industrial Automation.”