Radio Access Networks: Trends and opportunities for towercos

Enda Hardiman shares his views

Traditional telco structures are evolving rapidly. Evolution cuts across traditional domains of demarcation within individual telcos and between competing telcos. Competitive advantage can be secured by accurate identification of trends. This holds across the TMT sector overall, and is particularly manifest in the case of towercos. The advent of 5G will have far reaching effects. The imminence of ubiquitous 5G is however often overstated. 4G deployment will continue for some time. Deployment will however be in a context that allows rapid transit to 5G as 5G matures. Enda Hardiman of consultancy Hardiman Telecommunications considers issues and opportunities that present.

Evolution and technical innovation

There was a time, in the not too distant past, when the towers sector looked simple. MNOs sold towers to independent organisations, and leased back space. Towercos sold additional available space to other MNOs. New market entrants abounded, notably in emerging markets. The quality of the said entrants varied considerably. That was on the operational side. On the investment side there was a degree of naivety. Investors were attracted by opportunities for organic growth and by high valuation multiples. Sometimes they were surprised by risks and challenges that emerged from due diligence assessments. Matters however resolved. Professional market entrants brought stability, and the sector acquired a degree of operational maturity.

Operational maturity, however, is not synonymous with stasis. Technology developments, current and in prospect, define both risk and opportunity. There will probably always be a small market segment defined solely by real estate, but it will most certainly not be mainstream. Radical architectural initiatives are in prospect in networking. Business models of operators are evolving rapidly in response. It is in the interest of towercos to respond in suitable manner.

Cost pressures on mobile network operators

It is a fact of the telecommunications industry that customers expect, year-on-year, ever-higher data speeds. It is also a fact that customers are unwilling to pay significantly for the expected increments. Thus operators are constrained to find and deliver efficiencies, and most specifically to ensure that capital expenditure is allocated to the most efficient mechanisms of producing high speed and high capacity services.

This has led to something of a dichotomy. There are clearly many advanced, technologically enabled tele-borne services in future prospect. There is accelerating demand for the operator basic product, which is bandwidth. However, constraints on end-user spending have translated back to performance. In recent years mobile network operator EBITDA margins have fallen on average 16%. Operating free cash flow margins have almost halved. Shareholder return is barely above that of commodities. Operational efficiency thus defines a crucial, perhaps existentially determining competitive differentiator.

Disaggregation of network functions Allows capex and opex savings

Cost pressures on mobile network operators have come about at a time when technological developments have called the traditional business and operating models of all telcos in to question. Conceptual boundary lines, notably between retail and wholesale, have been under consideration since deregulation commenced 30 years ago. Whether or not vertical integration across network design, network operation and network commercialisation is efficient and necessary has exercised regulators, and experiments of varying degrees of complexity were conducted. The short answer is that, against previous models of technology, and discounting some broadband nuance for illustrative simplicity, the advantages of vertical integration outweighed the drawbacks. By and large, disaggregation that was attempted slowed key decision-making and inhibited innovation to unacceptable degrees.

It is no longer the case that disaggregation is unattractive. New generation network equipment features software configurability to a level several orders of magnitude more flexible than that of previous generations. The possibilities thus opened are broad reaching, and have been particularly capitalised on in the definition of 5G radio networking standards. Sharing of network facilities, as opposed to definition of strict demarcation of operational functions, is possible, realistic and gaining momentum. Operators thus anticipate significant CapEx and OpEx savings by elimination of replication in areas in which it is economic to cooperate.

Spectrum is key, but concerted management and allocation is needed

While perhaps a nostrum, it is nevertheless a fact that regulatory policy is often made without adequate consideration of the economic effects. Government machinery has a tendency to move slowly. Regulatory decisions are often made against outdated information and outmoded operating frameworks. Regulators are also notoriously susceptible to lobbying and special pleading, at national as at international level.

Of the areas that merit urgent regulatory attention at this time, perhaps spectrum allocation is the most pressing.

From a perspective of basic governing physics, there are two ways that the capacity of a radio base station may be increased. One is to increase power. That is not usually realistic, in that it acts against the fundamental objective of cell-based radio systems, which is to allow re-use of radio frequencies in non-adjacent cells. The other way is to increase the amount of spectrum that the base station can use. Harmonisation in spectrum allocation is urgently required.

Approaches to spectrum allocation are fragmented across countries. That which will be made available is unclear and often nationally specific. Manufacturers are constrained to design for a plethora of possible implementations. This increases chip and software complexity, prolongs development and testing times, and militates against achievement of economies of scale. Transnational initiatives are required. In the context of 5G, it is to be hoped that the ITU in particular will take a stronger line, and seek to avoid the patchwork that characterised 4G.

Spectrum allocation for 5G

Two distinct cases for 5G may be considered. The first case is 5G implemented in bands below 6 GHz. In order to avail fully of 5G speed benefits, contiguous 100 MHz allocations are required. The second case is 5G implemented in millimetre wave bands, at this time generally those in the 26-28 GHz region. Here, 400 MHz allocations are required. Harmonisation is somewhat better in the millimetre bands than in the sub 6 GHz bands.

Many countries have allocated spectrum in the 3.5 – 3.8 GHz band to 5G. Japanese Chinese regulators are considering 4.5 – 5 GHz. Other countries are considering 3.8 – 4.2 GHz. There is also discussion around 2.3 GHz and 2.5 – 2.6 GHz. Matters below 1 GHz are equally idiosyncratic with 400 Hz, 700 MHz and 800 MHz all under consideration. One might argue each of these on its own merits, but the risks of fragmentation remain. Lest the matter be thought simple, we set out in Figure one a compendium of currently evolving trends. The legacy of previously fragmented allocations will, it is trusted, be obvious.

Figure one: Spectrum allocation planning forward

Without straying in to excessive detail, it may be apparent that TowerCo planning needs to consider MNO plans for implementation mixes, single band vs. multi-band antennae, associated wind loading and, most particularly, MNO objectives of how best to achieve coverage and capacity objectives. Matters as they are evolving are thus several orders of magnitude more complex than the original considerations of 900 / 1800 MHz and 850 / 1900 MHz, and, indeed, incrementally more complex than was the case with respect to 4G. This is in some measure because 4G and 5G will need to coexist.

5G and 4G will coexist, small cell implementation will continue

5G implementations define opportunities and discontinuities forward. However, they do not define the summary vista as it immediately presents. Operators have made major investments in 4G, and adequate returns on those investments must be realised. Equally, 4G equipment is ubiquitously available and standards are, by and large, common across manufacturers. That cannot as yet be said of 5G. 4G will continue to be deployed. Meeting exponential growth in capacity demand is being addressed by installation of small cells, which, at this time, are predominantly 4G. Small cell deployment metrics are presented in figure two below.

Figure two: Small cell shipments and deployments

Small cell capacities and design considerations are illustrated in Figure three below. It will be noted that the governing, as opposed to implementation specifics are not confined to 4G.

Figure three: Small cell design considerations

Macro cells, of course, represent a significant element of towerco business, as do small cells of the micro type. We have designated small cells of the pico type as a possible area of interest for towercos. For some towercos, those willing to extend capabilities into radio engineering and elements of active equipment management, the pico sector most certainly an area of opportunity. We are less sure that small cells of the femto type, which essentially define a consumer mass market, will be suited to the ambitions of any but the largest towercos.

Femto cells gaining in importance

It should however be stressed that the matter of femto cells is not to be dismissed lightly. Femto deployments may yet be a core element of 5G ecosystems. This is because of the propensity of regulators to assign, and MNOs to deploy 5G in the 3.5 GHz band. 3.5 GHz signals do not propagate well to indoor locations. Thus, at the very least, in many instances an external or window-based antenna will be necessary, attached to a 5G transceiver / router. The femto cell in that case will likely define a WiFi hotspot, but of a somewhat more sophisticated type than now commonly used in domestic deployments. One concept of small cell management involves automatic collation of equipment information to allow self-configuration of networks. These matters are even more accentuated in the case of millimetre technology. Millimetre waves do not penetrate buildings to an appreciable extent.

Mixed 4G / 5G implementations will stimulate sharing

In parallel with the deployment of 4G cells, operators will consider deployment of 5G facilities, both in the absolute, as a cost-efficient way to meet demand, and to ensure that migration to 5G can in time be achieved in as seamless a manner as possible. This engages the concept of sharing of facilities across 4G and 5G, which also opens paths to facilities sharing between operators. Concepts of already-familiar Multi-Operator RAN (“MORAN”) and Multi-Operator Core Network (“MOCN”), sharing are relevant. Regulatory frameworks in this area differ widely. Some regulators encourage active network sharing, and some specifically prohibit it. The latter is unlikely to be durable. Economic cases for sharing are usually compelling, and are usually favoured by operators.

Figure four below illustrates the four possible cases of 4G / 5G sharing facilitated by MORAN and MOCN

Figure four: Coexistence of 4G and 5G in shared networks

For simplicity of illustration, we have not depicted shared antennae, but such is possible, notably in the case of shared spectrum. Clearly, this is an area that TowerCos will need to address. Leaseup rates will be affected, as will some aspects of landlord real-estate agreements. Some of these latter engage rental increases if multiple operators are hosted. As is already the case of leading towercos, understanding of radio planning principles, and, indeed, specific implementation options will bring competitive advantage.

Summary impact on towercos

New generation networks, including 5G, are inherently more flexible than were previous generations. Logical demarcations between operators have progressively lesser technical significance. Software definition of shared facilities is gaining rapid momentum. RAN sharing and spectrum sharing are increasingly common.

Practical considerations dictate that 4G deployment will continue. This, in the absolute defines opportunity for towercos in more or less traditional manner. It should not, however, be considered in isolation. The purpose of cell densification is to increase network capacity. 5G implementations will be undertaken selectively along with 4G, under the aegis of a common defining architecture. In-building services will gain prominence. It will be in the interest of towercos to possess skills in that area in order to engage meaningfully with operator objectives.

Infrastructure sharing, including spectrum sharing, is gaining prominence. That it will become the norm over the period of migration from mixed 4G / 5G implementations to “Pure” 5G is not in doubt. Base station rationalisation will occur, as, indeed, will antenna rationalisation on individual shared base station sites. Overbuild will become progressively less common, and ultimately be eliminated.

“Grass and steel” as an operating model for TowerCos is redundant. As already exemplified by the more forward looking towercos, expertise in capacity planning, radio planning and systems integration will define competitive advantage. Disaggregation sets an economic and operational case for TowerCos to assume roles, functions and new business lines that are currently embedded in operator RAN entities.

_

Enda Hardiman (hardiman@telecoms.net) is Managing Partner of Hardiman Telecommunications Ltd., which he established in 1994. From offices in London, Hong Kong, and Singapore, Hardiman Telecommunications serves TowerCos, Telcos, Mobile Operators, PE funds and Debt Providers worldwide. Hardiman holds a degree in Mathematical Sciences from Trinity College Dublin, and is a Chartered Engineer, a Fellow of the Institution of Engineering and Technology and a Fellow of the Institution of Mechanical Engineers. Prior to forming Hardiman Telecommunications, he worked with Raytheon Corporation, Westinghouse Electric Corporation, Siemens and Ericsson in engineering and in business development roles.

Leave a Reply