The proof of the pudding is in the eating

10 Feb 2014 by Nico van Eijk on Net neutrality

The Netherlands is among the few countries that have put specific net neutrality standards in place. It was the first country to do so in the European Union. It remains to be seen if these rules are working or if they are just another example of symbolic regulation: the proof of the pudding is in the eating.

A net neutrality law with teeth

Contrary to the European Union approach, which lacks a material implementation of net neutrality principles, the Dutch parliament decided to take a firmer position. By amending the Telecommunications Act, parliamentarians introduced a quite detailed regime on net neutrality. In Article 7.4 of the act, it is stated that providers of public electronic communications networks via which Internet access services are delivered and providers of Internet access services shall not hinder or slow down applications or services on the Internet.

There is a limited group of exceptions to this rule. Hindering and slowing down Internet traffic is allowed: a. to minimize the effects of congestion, whereby equal types of traffic must be treated equally; b. to preserve the integrity and security of the network and service of the provider in question or the end-user’s terminal; c. to restrict the transmission of unsolicited communication (spam) to end-users, provided that the end-users have given their prior consent for this to be done; and d. to implement a legislative provision or court order. Another very important net neutrality principle was based on incidents of blocked applications such as Skype and on the announcement by mobile operators that they would start charging for applications. The Dutch net neutrality article also forbids providers of internet access services to charge for internet access services dependent on the services and applications which are offered or used via these services.

The impact of the Dutch law

The no-blocking/no-charging restriction had an immediate effect on the market, in particular the mobile one. Originally, the mobile providers intended to block or to charge for specific services (Skype, WhatsApp), but they had to abandon the idea due to the new net neutrality rules. This led to a new subscription structure, with a substantially increased emphasis on data traffic. Data bundles are priced more specifically, and existing packages with unlimited data access have been replaced by packages with a specific size (data caps) and speeds. In fact, voice is no longer a dominant factor in the pricing models.

But how did these changes affect the consumer? The no-blocking/no-charging rule more or less killed traditional texting (SMS), but it is too early to tell whether net neutrality has had an effect on the overall costs for mobile broadband. A recent study commissioned by the Ministry of Economic Affairs seems to indicate that the overall price levels and options in the Dutch market are in line with the prices in other European countries. We might assume that voice income has been substituted by data income without having a substantial positive or negative effect on the monthly subscription fees.

The new neutrality rules had no effect on the fixed market. Internet service providers on cabled networks have no history of blocking traffic. Only one incident with the slow-down of traffic was reported but turned out to be a ‘misunderstanding’. One should keep in mind that the Dutch fixed broadband market is very competitive with the incumbent operator offering high-speed DSL or fibre and the cable television network operators offering high-speed broadband via their coaxial networks. The Netherlands belongs to the top broadband countries in the world.

A precedence case

In a very recent case, the regulator in charge - the Authority for Consumers and Markets - took a first decision on applying the new rules. Let me give you the facts first. Passengers on most Dutch trains have free internet access. The service, called T-Mobile HotSpot in de trein, is provided by T-Mobile, based on a contract with the Dutch Railways NS. In order to get the signal into the moving trains, T-Mobile uses its 2G/3G mobile network. The architecture of this network is focused on voice services and on keeping them available under varying circumstances. The rest of the capacity is used for data traffic. Now, the data service (needed for internet access) fluctuates strongly on board, due to the rapid velocity of trains. To counter this, T-Mobile has decided to block all peer-to-peer and streaming services (YouTube, Netflix) and to slow down file transfer. Without these measures there would be congestion and a lack of capacity. Users, having to share the available connection and capacity, would not be able to use the connection in a practical manner.

In a two-page letter, the regulator accepts the argument that the blocking and slow-down is necessary to remedy the consequences of congestion. The information provided by T-Mobile also led them to assume that all services are treated equally (no discrimination, everything is blocked).

The decision leaves many questions unanswered. Maybe not the most important ones, but certainly relevant enough for determining jurisdiction: does providing internet access in a train mean that a public communications network is involved, or is this mainly a private service offered by using (parts of) a public network? Can we compare the situation with offering internet access in a hotel, in this case a ‘moving hotel’? In an earlier case, the regulator determined that in most instances offering internet access in hotels must be considered a private service between the hotel and its guests. Without a public network/service involved, the rules simply do not apply. This is also a major problem in the US, where the court has limited the power of the Federal Communications Commission to regulate net neutrality1.

More interesting still, is what I would call the short-cut conclusion about congestion. T-Mobile uses its slow 2G/3G network on trains, saying that its new 4G network is too costly, requiring modifications on board the trains. Consequently, could one argue that the congestion is ‘self-inflicted’? Is this a relevant circumstance? Are operators under an obligation to keep their infrastructure ‘state of the art’?

Finally, the measure taken by the operator is drastic: all peer-to-peer and streaming traffic is blocked. Is such a measure sufficiently proportional? What about trains with just a few passengers or with passengers who mainly use voice services, and what about areas with sufficient coverage?

The devil is in the pudding

It is my impression that the regulator either did not think about these issues or did not want to deal with them. However, this is what net neutrality — more specifically network neutrality — is about. What do we mean with vague notions such as ‘congestion’ or giving dedicated space to specific services (the ‘managed’ versus the ‘unmanaged’ internet, the debate about allowed forms of discrimination)? A decision like this raises more questions than it answers.

In a study, we concluded that the future of the net neutrality debate will be dominated by video. Today, it is video that requires substantial amounts of capacity and therefore creates the most interesting challenges (also from a commercial perspective, as the video consumption/exploitation model seems to change). This T-Mobile case is a typical example.

On fixed networks, we will see the same kind of problems. Netflix and YouTube together make up 50% of peak-time internet traffic in the United States. In The Netherlands, Netflix was introduced in December 2013. Cable operator UPC reported that already 5-10% of its traffic consists of Netflix binge viewers. New ‘congestion’ issues will be on the table soon.

Of course, it is not about congestion exclusively. This is why I prefer to use the broader term ‘net neutrality’. ‘Network’ neutrality is only one element in a complex value chain that regulators have to deal with. We will have to wait and see what the future brings: the proof of the pudding is in the eating.

Add new comment