Tear up the rulebook for 4G in SA

SA needs a new approach when it comes to licensing the spectrum that will be used for next-generation wireless broadband networks. By Dirk de Vos.

The recent piece on TechCentral by BMI-Techknowledge senior consultant Martyn Roetter, in response to a Convergence Partners research report on wholesale open-access wireless networks, itself deserves some reply.

It goes without saying that the Convergence Partners’ research report speaks to that company’s own strategy, set out in several articles carried on TechCentral and elsewhere.

One could take the view that in the absence of sound regulatory policy in telecommunications in general and spectrum allocation in particular, Convergence Partners has simply forged ahead in anticipation of regulatory approval for its network plans on the basis that either it can secure the desired spectrum licences or, failing that, that it will be the main infrastructure provider to any new licensees.

The return on its investments would probably not be that dissimilar, either way.

Background
The current shape of SA’s telecoms sector is a result of an interplay of past policies, politics, the way our private sector traditionally operates, our small corporate sector, our limited skills, telecoms product development, technological change and, of course, the special case of our incumbent fixed-line operator, Telkom.

Policy or regulation is just one factor in a total mix of factors that shape our industry. That being said, good policy is a necessary ingredient in the efforts to achieve SA’s stated broadband objectives. So, the criteria against which we should judge policy is whether it is clear (a virtue on its own), is consistent with what is technically appropriate, whether it promotes investment in the sector and, obviously, whether it promotes or at least does not stand in the way of faster, better, cheaper and universal broadband delivery.

The central question to be asked is whether the telecoms industry, as presently structured, can realise these objectives. My view is that the answer has to be “no”. Instead, a policy that introduces wholesale, open-access wireless networks provides a far better platform for achieving the country’s objectives.

More fibre
Let’s turn to Roetter’s contribution, then.

The licensing of spectrum, especially digital-dividend spectrum, is mainly for fourth-generation (4G) long-term evolution (LTE) services at the last mile and not for backhaul and transport purposes.

That was the point of the Convergence Partners’ report. Fibre will be needed to link base stations. With large data payloads, fibre is the only answer. This is not to say that fibre to homes and businesses should not happen — a dedicated resource is always superior to a shared resource — but in large parts of our country, where there are low contention ratios, accessing fast broadband wirelessly is the only answer. Licensing LTE services would be a spur to more fibre investment, not less.

Spectrum auctions
The thrust of whether spectrum auctions are desirable or not, as Roetter appears to concede, is not whether these can be an effective way of allocating scarce spectrum. Rather, it is the question of what happens to the proceeds of an auction.

It is widely accepted that the 3G auctions that took place in Europe at the top of the telecoms bubble around the turn of the century, and which generated unprecedented revenues for European governments, delayed the roll-out of these networks.

The funds that could have been spent on the networks were absorbed by the fiscus.

In SA, the temptation to divert the proceeds of a future spectrum auction to the national treasury will be hard to resist. However, that money is sorely needed in the sector, even if applied to subsidise universal service.

This is the main problem with auctioning spectrum.

On the question of the desirability of wholesale open-access wireless networks, the most important issue is that the discussion is really about 4G/LTE. What separates LTE from previous generations of technology is that it is fully based on the Internet protocol (IP).There is no traditional or legacy “circuit switching” in LTE.

The consequence of this is that full-IP telecoms networks can more easily be analysed (and regulated) horizontally rather than attempting to regulate as if our current vertically integrated, full-service telecoms businesses were a given.

The technical model for doing so already exists in the format of the OSI 7 layer stack standard. This standard defines the different stages that data must go through to travel from one device to another over a network.

Of significance is that adopting this framework would allow investors with different risk profiles to invest in different layers without having to invest in a vertically integrated business that combines or integrates low-risk infrastructure elements which are mixed in with high-risk business services and retail application elements.

Open-access networks
This is already happening to some degree. Most undersea cables operate on an open-access basis and the investors in those include those uninterested in investing in other parts of the telecoms sector.

As we have seen, 80% of the capital needed to lay fibre is in the civil works — the trenching and so on. The regulatory regime governing this activity is not located at the telecoms regulator, the Independent Communications Authority of SA, at all. Instead, it is located within local authorities or the various national, regional or local roads agencies.

A class of telecoms investors is emerging that targets infrastructure but which has no interest in whether Google, Skype, Amazon Cloud Services or some budding start-up might prove a competitive threat.

The industry needs these types of investors, too.

The wholesale open-access model, used by newer subsea cable operators, has reduced dramatically the price of international bandwidth while vastly increasing the amount of available bandwidth.

SA uses less than 10% of the international bandwidth now theoretically available to it.

Open-access policies for land-based backbone fibre could render the same result as it has for subsea international fibre. If this is so, then it is surely incumbent on those opposing wholesale open-access wireless networks to give very good reason for their views.

Vertically integrated
Why is any of this important? Our incumbent telecoms networks are all tightly vertically integrated businesses and that is a problem if we want fast and ubiquitous broadband.

SA’s cellular network operators generate about 85% of their revenues from voice calls and only 15% from data services. More importantly, data services generate far less in terms of profit.

Is it then a good idea, then, to have these types of businesses, which are still struggling to monetise data services, roll out 4G broadband services?

My view is that we need a new approach for 4G. As a telecoms analyst said, “4G is not merely faster 3G”. The all-IP nature of LTE provides a strong foundation for seeing wholesale open-access wireless networks as simply an extension of an infrastructure type business. As such, serious consideration should be given to licensing this spectrum to operators who want to be wholesale access providers.

On a different level, the wholesale model also provides for the entry of more service providers at the edge that can access a decent connectivity solution and effectively compete with incumbent operators using proper wholesale pricing rather than the current “retail-minus” pricing structure available to them.

Not a panacea
None of this is to say that wholesale open-access networks are a panacea. There are legitimate concerns about competition and discriminatory pricing. Roetter points to some of these problems in his piece. Reduced infrastructure competition is probably the biggest concern.

In the early days of the cellphone industry, operators sold themselves on their superior nationwide coverage. The result was better coverage, even if separate Vodacom and MTN towers were co-located all over the country.

There is no point in obsessing about competition for competition’s sake. Some infrastructure, such as roads and bridges, do not submit easily to the economics of competition. Besides, up to now, it is not entirely clear that our incumbent operators have been entirely competitive. Some of us still remember the infamous “London agreement” between Vodacom and MTN, in which they allegedly colluded over prices in the early days of the industry.

Ensuring that there is proper competition between a handful of vertically integrated operators in a corporate sector, where the main decision makers know each other well, is tough to enforce. Standard economic theory demonstrates this.

Besides, the horizontal approach to regulation allows competition to be promoted in the layers where it should occur, while not forcing it on those layers where competition just produces a wasteful replication of infrastructure.

It is here where my biggest concern with Roetter’s approach lies. I would argue that his “pragmatic” or “nuanced” approach does not serve the country well. Instead, we need a clear conceptual framework as a departure point.

Wholesale open-access wireless networks for LTE, using digital-dividend spectrum, provides a superior starting point than fumbling through the various nuances Roetter mentions.

Skills
And what about skills? Well, all-IP broadband networks are increasingly modular, commoditised and, frankly, less “intelligent” – dumb pipes if you will. Moreover, as everyone who has looked at the CVs of the skills in the sector knows, the really skilled technical personnel have generally worked for a few of the telecoms providers at some time in their careers. They go where the money (and fun) is to be found.

Just look at Cell C CEO Alan Knott-Craig. I still have to rub my eyes in disbelief when I see him now argue the converse of high termination rates, the fees the operators charge each other to carry calls onto their networks. As every observer knows, this is exactly the opposite of what he argued when he was CEO of Vodacom.

Both arguments were delivered in a “hand-on-heart” fashion, pointing out the numerous benefits to the poor and disadvantaged sections of our community. The reversal of his position appears, to use industry terminology, “seamless”.

South Africans are famous for being a forgiving lot. I certainly am, and I welcome his reversal in approach. But the change of heart proves another point: policy makers should listen to but not pay too much attention to the views of the telecoms industry. They have a keen sense of what is good for them and their shareholders. They are also adept at producing arguments that endeavour to demonstrate that what is good for them is also good for the country.

The lost opportunities occur when incumbent operators capture telecoms policy. A lack of a discernible telecoms policy for digital-dividend spectrum or policies characterised as “pragmatic” or “nuanced” benefit only the incumbent operators. The objective of cheap, ubiquitous broadband services requires a different approach.

• Dirk de Vos is an independent commentator and head of QED Solutions