Summary Bullets:
- As a concept for running diverse logical networks over a common physical infrastructure, network slicing has been linked closely to 5G network transformations and 5G’s aspirations of servicing the needs of consumers alongside myriad industries.
- As it gets put into practice, however, a number of questions around slicing still need to be resolved: How granular will slices be? Which networks and network elements will be sliced? How open will slicing be to third parties? What can network prioritization teach us? Most of these questions revolve around business considerations, not technology considerations.
Last week, the GSM Suppliers Association, along with a who’s who of telecom networking vendors and industry associations (Nokia, Huawei, ETSI, GSMA, NGMN, IEEE, 3GPP), hosted its 2017 Network Slicing Summit, billed as “the first virtual event on 5G technology to serve all industries.” We had the honor of delivering a presentation as part of the summit. You can view a copy of it here.
The basic concept of network slicing is well understood: a technology for leveraging a common physical network to serve the specific requirements of diverse use cases. And given 5G’s objective of helping service providers target a myriad of different verticals (with varying requirements), the perceived linkages between 5G and network slicing are easy to understand. Starting from here, however, our presentation argued that there remain a number of important questions around network slicing which need to be conclusively addressed.
- Granularity. How specific will slices be in meeting the demands of many different audiences? Will a small set of templates be offered up, or will slices be built on a custom basis as requested by customers?
- Network Diversity. End-to-end slicing suggests that many different networks and service components will need to be sliced: RAN, transport, core, diverse radio networks… even the user device. Will each of these be sliced? If so, what’s the value of doing so, and when will true ‘end-to-end’ slicing materialize?
- Openness. The larger slicing vision includes the ability for enterprises to order slices and have them provisioned, potentially transparently to the service provider. Is this realistic? What does it tell us that SD-WAN, offering some level of service programmability, has only recently materialized?
- Today’s Slicing. While not technically slicing, network prioritization is available today, delivering similar benefits. What does its usage (or lack thereof) tell us? And where slicing of 4G LTE networks is possible, is there any reason we shouldn’t expect to see it applied there, too?
- SLAs. If network slicing aims to set up SLAs delivered by a service provider, how will we define those SLAs given the potential performance variability inherent to wireless networks?
- Net Neutrality. Network slicing aims to cater network resources to the needs of specific use cases. Where forms of prioritization are prohibited by the precepts of net neutrality, does this mean network slicing will be off limits?
While these questions address different considerations in commercializing network slicing, they all have one thing in common: they all (mostly) treat business considerations, rather than technical considerations, around network slicing implementation. That means any operator looking to execute on the promise of network slicing will not only need to build out the requisite network capabilities, but also a new set of requisite business relationships and policies around sliced services. Luckily, because slicing can be applied to today’s 4G LTE networks, operators can begin working on both the technical and business considerations around slicing today.