This week, we attended another seminar hosted by the Westminster eForum, on this occasion focused on priorities for 5G. Unlike many of these events, while there were sponsors, it wasn’t a vendor love-fest, but rather a collection of thoughtful speakers and some interesting perspectives. All in all, a very worthwhile morning and which provided some nourishing food for thought.
Among the many interesting topics, that of Network Slicing was much discussed. So much so that, during the concluding panel, Caroline Gabriel, of Rethink Research remarked that network slicing is fundamental to the success of 5G. We couldn’t agree more, but we would go further and suggest that it’s specifically the combination of network and RAN slicing that is really the key. Here’s why.
While early iterations of 5G will essentially be focused on enhanced Mobile Broadband (eMBB), future investment will depend on attracting business partners that really need new capabilities, such as ultra-low latency, in order to support specialised capabilities or to access applications and services in the cloud. In addition, there is the potential to deliver capacity and coverage to where it is needed, when it is required.
Network slicing is fundamental to this flexibility, as it allows dedicated virtual networks to be delivered across the same physical infrastructure. Thus, an MNO can allocate specific resources to an MVNO, or to another kind of business partner. In theory, any number of such virtual networks can be provisioned, although in practice there are likely to be limitations, as we saw at the recent Ericsson analyst day in London, during which Maria Cuevas from BT outlined the results of a number of early-stage trials of the technology.
Notwithstanding this, network slicing holds out much promise. A few years ago, Peter Dykes, then at Ovum, suggested that policy control had the potential to be the “best thing since sliced bread”. However, while the promise was clear, it hasn’t quite delivered what might have been expected. Network slicing has the potential to, finally, allow policy control to fulfil its potential.
However, there is a caveat. To be really effective, slicing must extend end-to-end, which means into the RAN, as our friends at Argela have argued (you can see a demonstration of what they mean at MWC - Hall 1, 1C40). But slicing is also where policy control will start to come into its own. The policy control function (PFC) will be the main agent through which different policies are delivered, in a way that is similar to the existing PCRF / PCEF architecture. In the future, this may be supplemented by device level policies, effected by a function resident in user equipment (GoS Networks was way ahead of its time in this respect and we anticipate renewed interest in such approaches).
Each slice will need dedicated QoS and service parameters and the PCF will be integral to the delivery of the required service and performance levels. So, with that in mind, why is this technical capability so important to the commercial success of 5G and, crucially, what implications could this have for the way in which the 5G ecosystem takes shape?
Network slicing will enable any business or potential service provider to obtain the performance they need from a holder of licensed spectrum. However, while standalone 5G is coming, it’s going to be a number of years before anything like national coverage is available. As one of the speakers noted, it will follow the people – so, coverage will be focused around the cities with the highest population density.
This is because 5G is going to be very, very expensive. In order to obtain the promised performance, there will be a huge increase in the number of cells required. According to Professor Dohler from King’s College, London, up to 500,000 access points will be required in the UK alone. As he remarked, no MNO is going to commit to this level of investment nationally. Therefore, different delivery models will be required – and new partnerships.
5G enables new capabilities and is expected to unlock a flood of new innovation. Indeed, many of the use cases simply aren’t known at the present time – but are expected to emerge once people begin to explore the technology. However, a focus on the most populated areas will restrict such innovation to those that are well-served in terms of coverage. Consider a business in a rural area that may benefit from 5G – we already know that high-speed broadband has been a key factor in enabling hubs of innovation, but what about industries that might benefit from 5G, such as mining or others?
What’s needed is a way to obtain network coverage wherever its required and that is unlikely to result from the efforts of MNOs alone, who have to secure ROI for their network spend. What would be more helpful is if businesses can either directly or through partnerships build their own local micro network points and then connect into the macro cellular network via backhaul. It would be great to imagine that MNOs will support this, but it’s a bit of a stretch. There needs to be a way through which localised, 5G hotspots can be build, either by communities, local authorities or businesses, acting in concert or unilaterally.
That’s why a different approach for 5G may well be needed. The award of national licenses and the usual pattern of deployments that slowly creep across the country is almost inimical to the innovation that 5G is supposed to unlock. I still fail to get consistent connectivity in London, let alone down in the depths of Dorset. And yet, places like Dorset must be connected to the new network, so that new civic or other services can be enabled.
There are several ways in which this can happen. For example, MNOs could be compelled to share their spectrum assets if they do not deploy in given locations within a specified timeframe. Alternatively, network sharing and national roaming would enable more infrastructure to be delivered, faster. Similarly, allowing companies to build their own micro cellular footprint, while connecting to the wider network could also accelerate deployments – and network slicing provides the means to ensure that each obtains the service level required.
The real issue is that MNOs will not rollout out 5G at a sufficiently fast rate for the innovation potential to be realised. If deployment cannot be accelerated, then it risks being a costly failure. So, the priority for 5G must be to enable licensing and regulatory conditions to ensure that rollout is not the sole responsibility of MNOs. That’s why slicing is so important – it’s the technology that can enable businesses and other service providers to secure the services they need, wherever they are.
If we can achieve a more open regulatory framework in which spectrum and infrastructure can be shared, backed by national roaming, then there will be a much greater chance of securing the benefits from 5G. If we leave it to the MNOs, then deployment will be limited to clusters around key cities. While that’s necessary, it won’t be sufficient to secure the possible advantages. To thrive, 5G needs to spread rapidly across the country, which means that other stakeholders need to be given the chance to either build islands to support their needs or to engage with others to do so.
While 5G offers some benefits to individual users, the real benefit is for business use cases. But, the coverage requirements and associated investments in infrastructure are simply too large to leave in the hands of MNOs. The old model of essentially offering monopolistic licenses isn’t going to be tenable. At this stage in the evolution of 5G, regulators must act to ensure that there is a wider ecosystem, so that islands of 5G connectivity can be created alongside national rollout plans.