one fiber network is enough or… fiber as an utility


Today’s post is about the necessity to offer an open access layer 0 and layer 1 infrastructure… does it has sense to offer dark fiber so all the operators deploy their own networks?

In Switzerland they are following this approach, but… is this level of unbundling necessary?

My personal opinion is that it is not necessary from the cost point of view, although I will not enter into regulatory issues.

Swiss multi fiber approach is a good solution in a market where all the operators want to keep with their existing business models, but… are those business models future-proof? Is this the best business model for the end user?

We all know that incumbents are not happy sharing infrastructure, but now, all operators are private companies so they have their boards that dictate the deployment strategies based on business profitability. This means that what used to work in the past, it is unlikely to work now, as the copper networks that where deployed in the 20th century were built by public telecom companies that later on where privatized, most of them keeping the copper that was paid by all the taxpayers.

Now, the picture is quite different. In most of the countries we have the incumbent operator, which was the former telecom public company, and new operators that deploy their own CATV network or resell incumbent’s services. Others deploy fiber, but this is just a minority.

The cost of deploying a new access infrastructure is huge, so just very big operators can afford to deploy NGNs. Thus, regulatory bodies believe that it is good to deploy dark-fiber-based operators so other operators can hire their infrastructure and deploy their active equipment. This is the approach that will be followed in New Zealand, for instance.

Although this is a intermediate approach, I do not believe it is the best approach, because it is not optimal from the cost perspective point of view.

Following this operation model, each operator will have to deploy their own active equipment in every single home, take care of the maintenance… and this increase the cost, thus, the price of the service.

So… what would be my recommendation?

Keep things simple and deploy one single network, operated by a public or regulated company, offering ubiquitous and identical services to all of the network (layer 2 and above) operators. Following this model, we will reduce

  • deployment costs,
  • operational costs and
  • will be able to offer universal services.

So, in a single strategic decision, we would be able to sort out all the limitations that are stopping next generation networks to be developed. Also, this would be the cheapest way to deploy NGNs, which means that potentially, service costs would also be reduced.

We can summarize this concept in a single sentence: telecom infrastructures should be considered a utility.

We would not understand to deploy parallel power networks, right? Why then we want telecom operators to compete at the infrastructure layer? I think this has no sense…


9 thoughts on “one fiber network is enough or… fiber as an utility

  1. Your simplistic utility model does not take into consideration the lock-in effects of having a single layer 0/1 provider.

    If there is just a single layer 0/1 provider there is a real risk that the wrong technical solution is deployed. In a monopoly situation everybody bears the sunk cost and nobody gets lower costs. In a competetive environment, everybody gets to move on and only the party who made the wrong choice is left to hold the bag.

    Fiber technology is by no means a mature technology, so it is to early to call a one size fits all solution. Just think about it: P2P, PON, split ratios, splitter location, RF overlay, IPTV, WDM, aggregation strategy, vendor interoperability, CPE choice, remote power, emergency services, indoor/outdoor demarcation, battery replacement, …, the list just goes on and on. Just imagine getting one of those things wrong and forceing everybody to deal with it as a systematic limitation.

    Competition is good on all layers, both active and passive. Just look at how multifaceted the competition and technology is in a competetive unbundled market. You have your basic xDSL, SDSL, bonded xDSL, etc, each targeting different customer segments and different customer needs.

    The very same is needed in a competitive NGN environment.

  2. Hi Zed,

    Thanks for your comment.

    I know that my comment may be seen as simplistic, but the situation in Spain is so bad, that I can only think about simple solutions to develop NGNs.

    Competition has demonstrated not to be effective and the real picture here is that we have still one very dominant operator, Telefonica. The rest of the players can not do much because of the prevalent position the incumbent.

    Note that Telefonica is one of the largest operators in the world and Spain is a relatively small market so in practice, they have total control of the landline connections.

    Also, I believe that competition is good but not on all layers. Sorry to diverge, but in the same way that it has no sense to develop two highways one next to the other, I think it has no sense to develop to NGNs network in parallel.



  3. Let me put my position in another way: would you trust Telefonica to run a layer 0/1 network as an utility, open and equitable to all?

  4. Actually, Zed, this is one of the approaches: functional separation.

    This is what has been done in UK with BT, so they separated infrastructures (BT Wholesale) from retail services.

    Also, this is the same principle that the Australian government is planning to do with Telstra, the incumbent.

  5. Well, Carlos, if you think functional separation is a viable solution then you are in for a world of hurt. Now, if you were proposing structural separation, then we would have something to work with.

    Functional separation is like having a wife and a girlfriend and pretending that you’re not cheating on either one because you take you ring of when you leave the house.

    You can ask any bitstream ISP in the UK about how well that functional separation is working out for them. Just don’t stand too close to them when they start talking about wholesale rates and the iPlayer. Notice at the same time how quiet BT’s retail arm is whenever these issues come up.

    As what comes to the Australians, they’ve always been a bit different down there. They have always been determined to learn from their former master’s mistakes and been adamant about repeating the same mistakes, but with more vigour and enthusiasm.

  6. Hi Zed,
    Maybe I am wrong, but functional and structural separations are synonyms for me. If you have different definitions, I would be happy to know.

    My concept is to split Telefonica into two different companies, so there is one company that offers wholesale services to all the service providers and another company that offers retail services.

    Telefonica offering both, wholesale and retail services, has been proven not to work, as here in Spain we have many ADSL resellers and prices are very expensive, with Telefonica dominating the market.

    In Spain, we have an example of an layer 2 open access network in Asturias ( which works quite well. There, end users can enjoy 100Mbps symmetric at 30EUR/month, for instance.

    The open access operator there is a public entity that offers bitstream services to all the Internet and video service providers.

    This would be my approach for the rest of the country.

  7. Functional and structural separation are two very, very different things.

    Functional separation is the means by which a single company offers retail and wholesale services through separate subsidiaries. Functional separation is basically all just a paper exersice.

    Structural separation is splitting a formerly single company into two different new companies where one company offers retail services and the other offers wholesale services. A structurally separated company may not have any cross-ownership and may not in any way cross-subsidise retail and wholesale services.

    Functional separation is just a sham. It’s like putting new wallpaper on a crumbling building. It might look good for a while, but all the time the landlord is screwing you and all the other tennants for every cent you have while the landlord is living it up, him and his buddies, who own and operate the building, in the luxurious penthouse.

    Structural separation is what real competition is made of. It removes any incentive to treat any single retail service provider preferrably or to maximize the profits of any single retail provider at the cost of the wholesale service provider.

    While it is obviously better to have a public entity as a monopoly provider, rather than a business, one should always remember that monopolies are bad for everybody else but the monopoly owner.

    Econ 101 teaches us that it is rational behaviour by a monopoly to maximize profits and to recover all costs. No matter how well intended, the subjects of a monopoly will be on the hook for every euro spent, be it a good or a bad investment. As it is the only game in town, you simply have no choice.

    Furthermore, as history has shown us, public entities have a tendency to mushroom out and to grow to any extent their budget will allow. This expansion policy combined with a monopoly income base is a receipe for disaster.

    It might look good now, but how about in 10 or 20 years after a few technology misses and a couple of harebrained projects? In a monopoly, you are on the hook for not only the current OPEX and CAPEX, but for every single mistake made, be they small matters or hughely expensive affairs.

    We do not need any new monopolies in the NGN, telecom has had enough of that for the past 100 years. All we need are the new fiber optic cables, the industry will take care of the rest. If there is a need for a layer 2 bitstream operator for cost savings, one will become available as a matter of business as usual.

  8. Zed, then we agree on structural separation, right?

    The problem that I see with telecom networks is that they tend to be natural monopolies, and this is very difficult to regulate.

    If we want telecom networks to be businesses, then we have the problem of universal access and the creation of digital divides. If we want them to be public owned, then we have optimization problems…

    In any case, telecom networks are strategic for the development of countries so we can not leave their development just to the economic interests of telecom companies.

    The perfect solution does not exist, but the most advanced countries in the NGN arena are following pseudo-public of favourable-regulated policies so I believe that a public intervention in the market is almost a requirement.

    Maybe combined models are the key, where we have open operators on non-economically-viable areas and direct competition in dense cities, but there, we will be wasting resources deploying parallel highways.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s