In the waves of hype and counter-hype that characterise debates about new technologies, often the reality is lost. There are those that argue that 5G will immediately usher in a broad revolution of all wireless-enabled areas. Conversely, there are those who say that there is no use case that 5G can solve that 4G cannot.
The reality is something in between, but while there may be some over-hype on the side of 5G supporters, it would simply be wrong to label 5G as “LTE on steroids”. There are some things 5G does that 4G simply cannot do. One of the most crucial is network slicing –the gateway to some exciting opportunities.
Custom network configuration -- known as network slicing -- is one of the most unique characteristics of 5G. In fact, it may be thought of as the most significant difference between 4G and 5G. In previous generations of wireless, the networks were configured universally, and each customer, each network operator had few options for customisation or optimisation of their network configuration. As the technology has evolved, certain aspects of it are more well-suited to certain types of vertical applications, each requiring different levels of performance. Put simply, network slicing allows operators to offer customised resources for the different types of applications running on their network.
A recent research paper published by Heavy Reading entitled ‘5G Deployment & Service Strategies’ found that network operators are generally optimistic about the value of network slicing. In fact, more than 70 per cent of network operators report plans to deploy network slicing to some degree within three years of commercial launch of their 5G service.
What will it look like?
It's important to think about network slicing as a truly end-to-end concept. It applies to the entire network: from the user equipment, to the radio network, to the core network, and beyond that to the transport network, the cloud and application servers. Yet, it's the 5G core network that makes network slicing possible, which would have otherwise been impossible with LTE.
There are tremendous efficiencies to be gained with network slicing, both on the side of the network operator and on the side of the end-user. If we look at the differences between an autonomous vehicle application, a healthcare application, or a massive IoT application, all require different types of resources. Some are very resource intensive or bandwidth intensive, while some require high levels of resilience, and others require a certain level of quality of service (QoS) or quality of experience (QoE). For each of these different use cases, the network operator will be able to use network slicing to customise the network and in effect, make it as efficient as possible.
For instance, a healthcare provider might require some very high reliability and low latency type of resources in their networked applications, and some redundancies when applications like remote surgery or heart monitoring run on the network. The healthcare provider needs very high reliability on these applications because the consequences of low reliability can be devastating. For this type of network slice, the operator will make sure there's enough redundant resources on the network to ensure this ultra-high level of reliability.
In the case of an IoT application, the needs are different. The main type of IoT data flowing over the network will be small data from things like sensors. In that case, reliability isn't nearly as much of a factor as management of a very high number of connections. So those particular resources can be sliced differently for this use case. In addition to this, things like mobility management might not be particularly important for a large IoT network, because many of the connected devices might not move around very much, if at all. Without the need to provision mobility for a network of millions of devices, the operator can save on capital expenditure, and savings on a significant amount of network resources can be realised as well. This kind of network would be sliced to optimise for a high volume of connections, rather than the highly redundant, low latency, high bandwidth network slice we imagined in the healthcare scenario.
These different scenarios (along with the many others, all with their own specific requirements) all provide significant benefit to the operators in particular, who can now offer more customised (and appropriately priced) solutions to their customers, as opposed to the older models where an operator had to try their best to offer a one-size-fits-all solution. With this type of virtualised architecture, customers will be able to finally deploy networked solutions that are tailored to fit their specific needs.
Networking slicing and 5G
The Heavy Reading research paper also indicated that, while overall sentiment is positive about network slicing, roughly 30 per cent of operators indicated they'd be making limited or no use of slicing in the early years of 5G commercialisation. This is likely because in the early years of 5G, the majority of use cases will be focused around providing enhanced mobile broadband (eMBB), which will look more like an extension of what we already see today with LTE use cases. In fact, the early 5G mobile broadband use cases may not require much network slicing at all.
But as time goes by, as 5G infrastructure is built out, and as we move closer to fully standalone 5G, more slices will emerge. These will be enabled by 5G core network deployments to cater to the many different use cases that 5G will open up the doors too.
Quite simply: 5G-enabled network slicing will enable far more applications and use cases than we can possibly think up. The capabilities of network slicing will make possible new scenarios: for example, autonomous vehicles, and IoT services.
These types of network configurations and these many different types of connections even aren’t possible with 4G, and it will be interesting to see how these scenarios and many others evolve in the coming years.
Saad Ahmad, Senior Manager of R&D, InterDigital