Open Access Infrastructure and the Cost of Mistrust
Steve Song, Many Possibilities writes: Recently I've been writing about the relationship of openness and trust in open movements which has led me to think more about the strategic role of trust in general. Now I find that almost every problem I try to analyse seems to have a trust component that I hadn't previously considered. Apparently this is known as the Baader-Meinhoff Phenomenon in which a concept or thing you just found out about suddenly seems to crop up everywhere. Be that as it may, trust is proving to be a useful lens to look at old problems in new ways.
Most recently that happened in looking at the issue of Open Access Infrastructure in Ghana and Uganda. The best definition of Open Access I have come across is from an Infodev, which states:
Open Access is about creating competition in all layers of the [Internet] IP network allowing a wide variety of physical networks and applications to interact in an open architecture. Put plainly, anyone can connect to anyone in a technology-neutral framework that encourages innovative, low-cost delivery to users. It encourages market entry from smaller, local companies and seeks to ensure that no one entity can take a position of dominant market power. It requires transparency to ensure fair trading within and between the layers based on clear, comparative information on market prices and services. It seeks to build on the characteristics of the [Internet] IP network to allow devolved local solutions rather than centralised ones.
You can see in this definition that the word "open” has two meanings: 1) open to anyone to participate; and, 2) open in terms of transparency in pricing and terms of operation. Since the release of this report, many African communication infrastructure initiatives, both public and private, have claimed the mantle of Open Access and yet few offer any real transparency on their operation, especially when it comes to pricing. Perhaps this is not surprising as pricing is negotiated around many factors. For example, an operator may wish to offer a discounted rate early on to an anchor tenant in recognition of the steady income and commitment they represent. They may not wish to offer that same rate later on when the economics of their business is more secure. Whatever the case, it is extremely rare for an operator to publish rates.
This highlights a problem with the concept of Open Access in that any push for transparency is likely to be met with opposition from operators. This conundrum hasn't been solved since this paper was published. A more interesting approach might be to look at the problem from the perspective of trust. Interestingly, trust is something the authors of the paper go into in some detail. They clearly state that:
one of the major issues in an Open Access business model is that it requires trust in parties outside of your own organisation
and they further identify the means of establishing trust:
the two keystone business practices for establishing trust therefore are nondiscriminatory pricing and no competition from the infrastructure supplier itself
I was in Accra, Ghana two weeks ago and had the opportunity to speak with network operators and to hear from the horse's mouth what the challenges are. If you look at the map below of fibre infrastructure in Ghana, you can see a great deal of duplication in network infrastructure. From Accra to Kumasi all the way up to Tamale, there is fibre infrastructure maintained by Vodafone, MTN, and GLO1. Three sets of excavations, three cables, and three independently maintained fibre networks. Perhaps you are thinking, well, yes it's bad but it's already done, the investment is already made, and hey, it provides network redundancy. It is worth keeping mind that ongoing fibre network maintenance is a serious cost for operators. According to one operator, there are over 500 fibre network interruptions per month in Ghana. And in terms of providing redundancy, if the fibre is all trenched along the same highway, the odds are pretty good that an interruption such as roadworks would affect all three .
So, what's the problem? Looking at it through the lens of open leads us to conclude that policies that regulate open access infrastructure are needed to ensure that there is infrastructure that could be shared by all players, lowering costs. This has often been the analysis in many broadband strategies and yet there are very few demonstrably open access networks to be found on the continent. But what if open were actually the symptom and not the cause of sharing? What if we look at the problem through the lens of trust? Indeed, trust appears to be the number one obstacle to bringing down broadband costs in Ghana where, similar to many other African countries, it can cost more to bring data to the coast via fibre than to get it the rest of the way to Europe or the US.
When I had the opportunity to speak to the person responsible for terrestrial fibre infrastructure at one of the larger operators, he was quick to identify trust as the key barrier to cooperation. He said, "why would I let another fox manage my hens?” He went on to say that the problem of mistrust was not just among operators but also between operators and government. Operators feel that are always being approached cap-in-hand by government and government are frustrated that operators are not bringing costs down.
In the midst of digesting this problem, I was fortunate enough to meet one of the architects of Project Link in Kampala, Uganda. Project Link is an initiative by Google to build a comprehensive metro fibre network in Kampala to offer wholesale access to high-speed broadband. Launched in 2013, Project Link has sought to be a neutral player and trusted third-party in the access market in Kampala. Project Link now has over 800km of fibre in and around Kampala. That's a lot of metro fibre in quite a short period.
Viewed through the lens of open, Project Link doesn't look very good. Their website is pure PR with no substance. Nowhere on the net can you discover who actually owns Project Link (Google does) or who does the build-out and maintenance of the network (Soliton-Telmec does, apparently). Project Link is not transparent on pricing either. They negotiate prices like all other operators do, in private. What they are transparent about is the fact that a) they only sell wholesale; and b) they sell capacity not traffic i.e. there are no caps on traffic.
Historically, Uganda has had many of the same problems as Ghana in terms of a lack of trust among operators. And indeed, the initial reception to Project Link from incumbent operators, by all accounts, was pretty chilly. Operators' saw their own fibre infrastructure as their competitive edge and showed no interest in buying capacity from Project Link. What happened next was interesting. Smaller ISPs began to buy capacity from Project Link in order to offer competitive services to business clients in Kampala, undercutting rates from the incumbents. Business clients are the most lucrative clients for operators and it is not hard to see how this might quickly begin to sting for the incumbents.
Within a year, incumbent operators began to shift their views. They realised that over time, fibre was unlikely to be a competitive edge and, like towers before them, fibre networks might be better outsourced to a third party. Now you could see this as simply a case of a new competitor offering lower prices, bringing more competition to the market.
Viewed through the lens of trust, what I see happening is this. Google's entry into the market raised the cost of mistrust. Prior to Project Link, operators could afford to mistrust each other and build/maintain completely independent infrastructure. They could afford not to trust each other because they could charge enough to maintain their mistrust. By enabling smaller players to get into the broadband market, Project Link ended up making mistrust so expensive that operators were forced to shift their strategy. I found this interesting because one often thinks of trust from the point of view of a gentle, nurturing process but in a situation where trust is badly broken, I can see how a strategy of simply increasing the cost of mistrust might be more effective.
Finally, the issue of enabling smaller players is another element of Open Access that is often not emphasised enough. Healthy markets have always had room for small players to nip at the heels of larger players. That is something that has been missing from the telecom markets in general in sub-Saharan Africa and needs as much emphasis as openness.
The good news for Ghana is that there are rumours that Accra is likely to be the next city to benefit from Project Link. Stay tuned.
Source: Steve Song 27 February 2015