Your browser is not supported. Please update it.

26 July 2022

Enterprise 5G solutions proliferate, threatening to make slicing redundant

Special Report: Industrial 5G

Network slicing has been held up as the ‘killer app’ for 5G since the first work on 5G standards began, yet it remains very challenging to define, let alone deploy. While operators grapple with the issues, in the meantime, enterprise players are moving ahead more quickly with other solutions, such as private cellular or WiFi 6/E.

There are implementations of technologies that could be called slicing – fixed network slices; static pieces of transport or sometimes wireless access networks that are carved out for a particular customer; enhanced forms of virtual private networks. But the real vision of 5G slicing supports dynamic, automated creation of slices on the fly, their capabilities defined by the application calling them. These slices would potentially be ephemeral, or certainly easily reconfigured, and they would reach from end-to-end (from transport to core to multiple access networks), and even involve international roaming.

This vision is a long way off becoming a reality. There are many trials, but most at a preliminary stage. Some early movers, such as the Chinese or Korean MNOs, have launched slices for specific industries or enterprises, but these use very customized architectures that require significant budget and skills to implement. The same is true of Rakuten Mobile and Dish Network, both of which have built large-scale Open RANs that will support slicing from day one. Rakuten has not yet detailed any commercial plan for this capability, though Dish is more ambitious, claiming it will run a wholesale-only business in which even its own mobile brands, such as Boost, will have their own slices, as will enterprise customers.

A more uniform, less hand-crafted approach would help a larger range of operators, and other service providers, to start to implement slicing, as would clearer ways to integrate existing networks with the new cloud-native ones. But several of the key enablers remain immature. One, of course, is the 5G core, which is essential to support full end-to-end slicing including the air interface. Another is the virtualized RAN, which is also essential if agile, reconfigurable slices are to be created in the cellular access as well as the transport and fixed networks. And many advanced slicing use cases also rely on edge compute infrastructure to support both the virtualized networks and the data processing.

The latest in a rising tide of slicing tests has been described by Telefónica and (see below), and this joins other interesting trials and proof-of-concept activities that have taken place over the past couple of years. These include Deutsche Telekom’s innovative showcase of international federated slicing, jointly with SK Telecom; and Vodafone’s announcement of the UK’s first fully automated end-to-end slice demonstration.

But these are baby steps towards a highly complex architecture for which the business and ROI cases are not clear. After all, the scale and nature of enterprise demand for 5G is becoming more visible, but how often will it require a dedicated slice, and how will that be monetized? Where enterprises do want 5G connectivity with very high quality and predictability, they are tending to opt for private networks, and in that market, the MNO may not always have a dominant role in the value chain. Today’s report highlights just two of the big industrial players, Bosch and China’s JD, that are building international businesses around private wireless, without the need for slicing at this stage.

Hybrid public/private connectivity, with increasingly flexible and software-defined handover, as well as the emerging approach of subnets – which are an interim solution between private networks and dynamic slicing – will all provide enterprises with a broadening set of options that should meet their connectivity and pricing requirements. Slicing is too good a concept to disappear altogether, but it does depend on a degree of network cloudification that will not be mainstream for most operators for many years, effectively making universal, standards-based dynamic slicing a ‘6G’ issue