5G Network Slicing Technology on the Edge: It’s All about QoS on Shared Resources
Network slicing is one of the most important features of 5G networks, as it is the one with which communication service providers have the most business-related expectations. This is because 5G network slicing technology is expected to enable new business models, where customers can order a network slice as a service.
5G network slicing for vertical industries allows the emergence of network tenants
Being able to purchase network slices means that customers will become “network tenants”, who can have their own “piece” of network that is designed per specific industry vertical. This leads to an evolution of new, very interesting business models, which support the roles of infrastructure, network slice and service providers.
This concept of complex business models is not a new one, but network slicing in 5G will enable the creation of slices on demand, which should lead to a dynamic creation of new partners, where each one has their own slice. The phrase “on demand” is in fact the key for network slicing technology concept to be able to became a real new business enabler.
Owning “personal” networks
Having a network slice has the effect of giving the tenant “their own” a network, and that it is tuned for their specific business needs. From the network perspective, it means that such a slice has pecific quality of service requirements (QoS), which must not be affected by the fact that there could be more than one slice /tenant competing for the same network resources.
Providing this kind of isolation can be really challenging, especially with on-demand creation of slices, which requires not just best efforts, but an ultra reliable low latency (URLLC) service quality.
Low latency support in 5G edge computing
Low latency is yet another key aspect of 5G technology. One of the technological architecture assumptions was that it would support the concept of the edge and 5G edge computing. With simplification at the edge of the network, network functions and applications (for example, real-time factory robot coordination applications), can be collocated at the edge, avoiding the problem of multi-hop communication resulting in unaccepted latency.
Although the introduction of the edge concept to 5G is one change which can be perceived as challenging compared to 4G networks, it may, from the network slicing perspective, make life easier in some cases.
Low latency, edge computing and 5G: real life examples
For example, for slices created for two tenants (let’s say - car producers, who have their own factories in different locations, utilizing the concept means that both of these tenants can use different edge nodes. So, even if both slices have stringent requirements (URLLC) and may need to operate at the same time, they can be served by physical resources which are delivered by different radio sites and edge data centers. In this case, it is relatively easy to ensure isolation between the two tenants, and traffic generated by one of them will not impact the other. Some may even argue that this is not “real” slicing, as the slices don’t share the same resources.
The situation is more difficult when the business application has very stringent requirements related to low latency, and at the same time is all about “not at the edge”. An example of such a case is remote surgery, in which the goal is to enable a doctor to perform the operation with the patient far away. Having low latency enables the doctor to “feel” the body of the patient. In this case, ensuring QoS and isolation is literally a matter of life and death, and yet the connection must operate via resources of the entire network so will definitely not be based on dedicated hardware. It may also be the case that different surgical procedures take place at the same time, and none of them should be impacted by other services implemented on the same network.
Network slicing is certainly a very interesting concept, especially from the perspective of enabling a great many varied business cases in support of different industries and verticals. We must remember, though, that the ability of the network to provide the expected QoS is essential for the success of network slicing technology as an enabler for 5G digitization of almost all aspects of our life.