Poland is latest country to raise issue of 5G RAN sharing

The issue of network sharing tends to move into the spotlight whenever a new generation of technology is on the horizon. It is likely to be the focus of particularly intense debate with the advent of 5G, and the idea of a single national network is being discussed again – not just in emerging economies, as happened in 4G, but in European countries like Poland and The Netherlands.

Many operators are struggling to find a strong business case for deploying 5G. They have plenty of capacity left in 4G and the services which would really require the capabilities of 5G seem, to many, futuristic or uncertain in the potential for monetization.

Yet governments, and some industrial sectors, are impatient for 5G – a situation which seems to lend itself to a co-investment approach, to accelerate roll-out of the new networks while keeping costs affordable for MNOs. The requirement becomes even more urgent when densification is considered, with a large number of operators expecting to deploy 5G initially to serve targeted zones of coverage and capacity around city centers, business parks or industrial facilities.

There are various approaches to sharing the cost of deploying and maintaining a network. Passive infrastructure sharing is increasingly common as operators offload their towers and look towards dark fiber to support their next generation backhaul and fronthaul requirements. In the small cell environment, sharing of sites and backhaul will be almost inevitable in highly dense scenarios, and there are already specialist neutral host providers, as well as tower and fiber owners, which are looking to extend their model into small cell-as-a-service (SCaaS).

The logical extension of this will be to share active equipment too – in calculations Rethink has carried out related to different small cell roll-out models, a neutral host approach, in which a third party acquires the sites but also implements and manages multi-operator access points, can reduce TCO per operator at least three-fold. Neutral host specialists like Dense Air (which is even acquiring licensed spectrum to support its build-outs) as well as towercos like Crown Castle in the USA, are showing the way forward and such initiatives are set to accelerate once densification becomes a mainstream strategy.

Active RAN sharing in the macro network has been a more controversial topic. Operators fear losing control of their ability to differentiate based on network quality or optimization, while regulators are wary of reduced choice and competition. Amid the high capacity demands of 5G, there are also concerns that a shared RAN would quickly run out of steam – in the UK, Vodafone and O2 have pulled back from their rare active RAN sharing partnership, Project Beacon, in areas such as London where very high capacity is expected to be needed. Elsewhere, however, Vodafone is still one of the industry’s strongest proponents of sharing – it recently signed a RAN sharing deal with TIM in Italy.

Other approaches to sharing include a wholesale, neutral host macro network, supporting all the MNOs and potentially other service providers; a variation on that, which would see one of the operators, rather than a third party, deploy the network, on a utility basis; and a network co-funded by the operators and key stakeholders.

The last of these options is likely to become more prevalent when 5G Release 16 is commercialized and it becomes possible to deploy low latency, high availability networks, optimized for the needs of industrial IoT or critical communications. At that stage, key industries will be keen to see the 5G networks planned and implemented to suit their requirements – and one way to ensure that is to share the cost burden with the MNOs, in return for a say over the network priorities and KPIs.

This has already been seen mainly in advanced Asian economies like Japan and China. In the former, Rakuten is partnering with several public utilities to reduce the cost of its greenfield roll-out, while railway operators, device makers and manufacturers all invest in networks via partnerships with the three established MNOs. In China, a group of Internet giants such as Alibaba, and industrial companies, have taken stakes in China Unicom, and made it clear they will expect their investment to be rewarded by a 5G roll-out that is well tuned to their requirements.

The wholesale national network is the most contentious of all the options because it often comes at the behest of government, and therefore accompanied by concerns that operators will have their freedom of action – in technical and commercial decisions – restricted.

In 4G, efforts to create a shared, wholesale nationwide network were seen in Russia, Kenya and South Africa, but all of the failed amid MNO hostility and failure to agree contractual terms. A similar effort in Mexico seems, so far, to be progressing better, despite a fraught and much-delayed early phase.

If the politics of creating a platform which the MNOs share equally prove too daunting, some markets may eventually go the same way as electricity, railways and even wireline broadband has in many regions – towards a utility model, in which a single operator builds the infrastructure and supports all the others.

As for a third party running that network, there have been a few attempts to create a national neutral host platform, but these have often faltered on regulatory concerns (LightSquared, now Ligado, in the US is still struggling to get its wholesale-only network off the ground); or because, if the MNOs will not sign up, there is insufficient revenue to tap. The latter barrier should be lowered in the 5G era as more and more companies look to offer optimized services that require greater control over the network than pure over-the-top or MVNO arrangements. Content providers like Netflix, web giants like AWS and industrial or private cellular operators are all increasingly wanting to make deep connections with the network itself to enhance the QoS of their offerings.

In 5G, the intensified need to reduce TCO may prompt operators and governments to look again at sharing models. Last year, consultancy McKinsey published a report arguing that network sharing would be essential to “reduce cost and make 5G deployments feasible”.

In Poland, state-owned fiber network operator Exatel is calling for a consortium of stakeholders to implement a shared 5G network, which would reduce costs for all, and particularly for those which want to enter the wireless space from other sectors such as wireline or cloud.

A shared approach would make it easier for a company like Exatel to narrow the gap with the big MNOs – T-Mobile Poland has already started to build out 5G, using Huawei equipment, and will offer full commercial services from early 2020.

Exatel’s CEO Nikodem Bończa-Tomaszewski told Business Insider Polska that a consortium of public and private companies, co-investing in the new network (financially or by supplying resources like fiber backhaul), would be the “safest and cheapest way” to implement 5G. He said the emergence of “softawareization” in 5G networks would make it easier for operators to secure their own virtualized share while still being able to differentiate their services and QoS.

Other benefits, said Bończa-Tomaszewski, would include an improved ability for the government to oversee network security as cybersecurity concerns, stemming from US-China tensions, start to overshadow 5G plans in Europe and elsewhere.

However, no sharing plan on a national scale is likely to get government approval without the support of most of the MNOs. In Poland, TMO looks reluctant to support initiatives that might help its rivals, new and old, to reduce its 5G headstart, while Orange has also said it would prefer to deploy alone. However, Plus – the MNO owned by Polkomtel – is reported to be interested in the sharing idea.

In the Netherlands, there has been considerable public debate about the issue, sparked by an article in the business newspaper, Het Financieele Dagblad, which held up the Vodafone Italia/TIM arrangement as an example of how 5G roll-outs would be afforded.  “I think it is untenable to build three more networks next to each other,” Remko de Bruijn, partner at consultancy AT Kearney, said in the article. “Financially, the telecom companies have to start working together more.”

However, the three Dutch MNOs are not rushing to comply and all gave wary comments to the newspaper.

Two factors that may slowly change operators’ attitudes to sharing are the evolution of network slicing, and the integration of the edge cloud with 5G connectivity. In the latter case, operators are keen to add edge compute capabilities to their own locations – central offices and even cell sites – to support enhanced experiences for their core user base (for instance, low latency augmented reality gaming). But when it comes to the growing enterprise requirement for combined edge computing and connectivity, usually within a building or site, MNOs will need to work with other asset owners, including the cloud giants, if they are to tap into that opportunity.

As Exatel pointed out, network slicing removes many of the fears that a shared network would reduce the ability for each MNO to differentiate the quality of its connectivity. Ultimately, the vision for slicing is that a single orchestrator would be able to draw on resources from many asset owners – wireless, wireline, cloud – and different combinations of elements would be called up automatically on-demand, to support a particular service or user.

That is a long way off, but even before full slicing there needs to be a platform that will enable the wide diversity of service providers and use cases – the diversity that is the main justification for 5G. The rising popularity of private LTE networks in enterprise settings is a step towards a model in which the network is optimized for many different purposes, and that will drive the development of full end-to-end slicing.