Your browser is not supported. Please update it.

11 January 2022

Neutral host model gets into gear with Open RAN

The neutral host model gained ground during 2021, largely because it is being seen as a good fit for a number of 5G deployment situations, urban and rural, indoors and outside.

The emergence of Open RAN has also assisted by encouraging shared virtualized infrastructures that allow greater flexibility and freedom of choice, while avoiding contention over selection of equipment vendors when multiple MNOs or enterprises are involved.

They are not a panacea though and it is important to distinguish between situations where they are desirable or feasible, and those where they are not. It is certainly true that their stock has risen with 5G, because the higher frequencies call for greater density, which increases the economic and logistical appeal of infrastructure sharing. Nevertheless, the neutral host model will gain less favor in areas of intermediate density, including suburban and more populous rural areas where MNOs will prefer to build and maintain their own separate networks of macro cells as they compete over price, performance and coverage.

One of the potentially most lucrative and game changing use of the neutral host (NH) model is in private networking by giving enterprises a flexible infrastructure they can segment for different applications or user groups, including customers and partners, with scope even for deriving money through sale of excess capacity to MNOs or others.

Yet, the NH model first evolved almost purely to serve MNOs in more remote areas where the cost of base station deployment would otherwise have been excessive or not justifiable for single operators.

This has evolved into three categories in terms of architecture or configuration. First, there are the spectrum-based NHs where the host owns its own RF space, typically still serving multiple MNOs on some kind of rental basis.

Secondly is the newer category focused on multi-operator small cell-as-a-service (SCaaS). Without its own spectrum, this model can rest on multiple clusters of small cells, each allocated to a given MNO or other entity, with shared backhaul. Alternatively, an SCaas can be a single virtualized cell supporting multiple frequency bands for different MNOs, and segmentable that way.

The final category is the private enterprise NH network that can be segmented for different applications or services and user groups. A hospital, for instance, could deploy such a network, segmented for high bandwidth internal applications such as transmission of medical images, while also providing indoor coverage. MNOs can ride on these systems to reach their subscribers as in- or out-patients, in return for a fee. Similarly, MNOs may want to hitch a ride over enterprise NH networks in other sectors, such as retail, to reach customers.

Indoor coverage for large sites such as offices, stadiums, medical campuses and hotels, has been the primary use case for the NH model in the past. Another is to serve rural or remote areas where MNOs would look to pool their infrastructure to serve a relatively sparse base of users. The NH is a convenient and cost-effective model for doing this that avoids the need for direct bilateral or multilateral arrangements among operators.

The third major category relates to urban areas requiring 5G or 4G densification through small cells, where the need to be close to users for capacity or latency again makes it prohibitively expensive for all operators to deploy separate equipment. There are also aesthetic, planning or logistical constraints over deployment of too many base stations, so again the NH model can be a good compromise.

There are various specific categories within these three broad use cases, such as coverage alongside roads and railways, where again it does not make sense for MNOs to run separate network lines. Another rapidly emerging category covers temporary sites on various timescales, ranging from construction sites requiring coverage for up to a few years, to music festivals where it might be just a few days at a time. In either case, the NH model can economically enable permanent infrastructure to be recruited or allocated temporarily.

Activity in most of these areas accelerated during 2021, especially in private enterprises. As we reported in our last issue of 2021, this can happen when the commercial case is not strong enough for MNOs to deploy infrastructure themselves. In many cases, this is because subscribers of all MNOs need to access the private network, but not in sufficient numbers to justify individual deployment, except in the case of very large enterprise sites. In that case a NH capable of supporting specialist service providers will often make better economic sense.

We reported on the acquisition of two young players in the private wireless space –  NH provider DenseAir, which was bought by Google-backed Sidewalk Infrastructure Partners (SIP), and enterprise core vendor Quortus, which has become part of Ericsson’s Cradlepoint unit.

There are, of course, hurdles to be overcome deploying NH networks in any of these categories. Billing can be one of them, as can SIM provisioning, which can be handled by the NH or participating MNOs. For smaller enterprises, the NH model may be deployed alongside the existing public networks to simplify SIM provisioning and billing. For example, such an NH could implement passthrough to whichever operator a given user subscribed to, which would circumvent complex billing issues.

Another challenge, which could arise for enterprises, NH providers or MNOs, is orchestrating end-to-end service level agreements (SLAs), given the division of responsibilities. Some technology vendors in the field, such as Celona, have implemented micro-slicing policies across their systems, capable of providing relevant metrics on latency and packet error rates to all parties, including enterprise applications running on private devices, and also relating to traffic flows broken down by operator.

Such technology vendors have been reporting that growth in NH deployments has had the side-effect of boosting demand for Open RAN products and technologies. As one example, Mavenir has reported that such deployments often require its MAVAir software serving virtualized RANs, deployed last year by two UK NH providers, Ontix and Vilicom.