Christoph Wagner explores the net neutrality debate in Germany focusing on the notion of paid prioritisation, network management, and discrimination, drawing parallels with the German magazine distribution platform Presse-Grosso, as well as the two-sided market mechanism used by cable operators.

A comparison with magazine distribution and cable TV in Germany

Everyone who uses the Internet is familiar with the hourglass that appears when the computer screen freezes and nothing happens for a while. Depending on how long it stays there, this hourglass first of all evokes a feeling of powerlessness, then anger and finally provokes the user to desperately hit keys at random in the hope that this will actually make something happen. The hourglass, an anachronistic symbol of timekeeping, is supposed to prevent users from losing their patience. In the hourglass, every grain of sand is driven by the force of gravity and flows according to the principle of equality, reaching its destination after having passed through the narrow neck of the glass. Every grain of sand reaches its destination, there is no special treatment for individual grains.

Traditionally, data packages are transmitted on the Internet in a similar way. Regardless of their origin, content or the addressee, these data packages are routed through Internet bottlenecks according to the so-called "best-effort" principle. This means that all content and service providers have equal chances of reaching their destination irrespective of whether the data being transferred is text, a moving image, a file-sharing platform or a real-time service, such as computer games. As the number of such data-intensive services on offer increases, so does the likelihood of capacity bottlenecks and network congestion. This results in users being confronted with the hourglass more frequently, which in turn leads to an increase in the number of angry users contacting their access providers. Having signed up for flat rate price plans promising them high data rates, they then discover that, at times, they cannot surf the Internet any faster than in the old days of ISDN. Providers and users alike are no longer willing to accept this and are calling for higher speeds and possibly even special treatment and priority rules in order to reduce the hourglass-gazing interludes.

The access provider then faces the question of whether it is able to remedy the situation. Today, network technology allows data packages to be attributed to certain senders and the respective service category to be identified. It also allows certain contents/services to be filtered out and data to be transferred at different speeds ("service prioritisation"). Just as if the grains in the hourglass were pre-sorted so as to reach their destination through different-sized openings. Such data traffic management includes priority rules and special "lanes" for certain content and services so that users can obtain faster results and do not have to endure waiting time spent staring at the hourglass. All other services either remain in the data "queue" or are transferred at snail's pace.

Whatever is technically possible and meets the customers' requests will be implemented unless it is prohibited. The question of if and how this should be regulated is the subject matter of network neutrality. Neutrality means impartiality and parties are therefore not legally or economically linked to the content or service providers, like, for example, most of the companies belonging to Presse-Grosso, the Association of German Book, Newspaper &

Magazine Wholesalers, which deliver all types of magazines to kiosks irrespective of their origin. This neutrality is extolled as an important achievement in the context of a free and varied media landscape. However, due to the growing vertical integration of network operators and access providers, which increasingly offer their own content, in contrast to Presse-Grosso and for structural reasons alone, it cannot be assumed that the net will necessarily remain neutral. The network operators themselves have a substantial interest in traffic regulation, i.e. in promoting their own services and contents and obstructing those offered by competitors. And of course they also want to improve the small margins generated by their data transport services, whether through toll gates for content providers or special prices for higher speeds. Such conduct is in fact generally permissible in competition unless the network operator holds a dominant position in the market.

There are already plenty of examples of non-neutral network conduct; prominent examples are T-Mobile's blocking of Skype services or Comcast's obstruction of forums critical of its business practices. While in these cases it is obvious that the network operator was pursuing its own interests, data traffic management may well be in users' interests and then appear to be more "neutral". Comcast's detrimental treatment of data-intensive file-sharing platforms, for example, is in the interests of all other users as this permits faster use of less data-intensive services. In this case, the equal treatment of all users could even be used as an argument for not treating all providers equally, as all users should have equal access to a minimum data volume and not be impaired by individual heavy users of online games or file-sharing platforms.

Just as the new network technology allows for services to be slowed down, it also permits accelerated transport of services. Currently, each content provider has equal priority in accessing the net, all it has to do is to pay its hosting provider who in turn will purchase capacity to connect to the Internet. Large content providers will pay content delivery networks such as Akamai or Limelight to avoid many of the traffic jams of the "best efforts" Internet. The actual transfer of data over the Internet is not paid for by the content provider itself but rather by the network operators, which then pass on the costs to their customers. As flat rate price plans now prevail on the market, the network operators' margins have narrowed, fuelling a desire for additional sources of revenues. This could create a so-called two-sided market, as is the case with the cable networks, i.e. the network operators collect money not only from the end customers but also from providers that attach value to special treatment of their data. A preferential or particularly safe transfer (Quality of Service [QoS] or premium service) could, for example, be guaranteed in return for payment of an appropriate fee by the upstream content provider.

There are thus three basic patterns of deviation from network neutrality. Firstly, the blocking or obstruction of certain competing services or unfavourable content by the network operator, which is obviously problematic and should be inadmissible on the basis of applicable law alone; secondly, the preferential treatment of certain services from content providers who that are related to the network operator and pay extra to the telecom operator, and thirdly, the obstruction or preferential treatment of certain data-intensive services or offers aimed at achieving "neutral" data traffic management in order to avoid bottlenecks.

Before looking at these categories, the primary question arises as to whether it is acceptable to deviate from treating all offers and contents equally. The Internet has not only become a medium and factor for public opinion-forming processes but is also now central to the freedom of information. Against this stands the network operator's constitutionally guaranteed right of ownership allowing it to deal with its network as it sees fit. There is no doubt that statutory interventions in these freedoms always require justification. This raises the specific question of whether network neutrality must be safeguarded by the state in order to protect the constitutionally guaranteed fundamental rights of communication on the Internet. If this is the case, the socially-beneficial nature of Internet property rights should justify corresponding reasonable restrictions, similar to the must-carry regulations in the cable industry.

In this respect it is first of all important to recall that the net's non-discrimination of all data without exception is no guaranteed fundamental right. Even in terms of the concept, network neutrality is not to be confused with the principle of treating all data in the net equally. A communications order that satisfies fundamental rights may still be upheld even if certain content is prioritised, as long as each provider always has access to every user and vice versa. It is doubtful whether fundamental rights are violated merely by the transfer of individual services or contents at different speeds, at least where the user still has access of acceptable quality at an acceptable speed. Neither has a strict principle of equal treatment been recognised for transmitting other media offers. The mustcarry requirements in the cable industry go some way along these lines but are by no means open to all providers and even the must-carry cable capacity varies in quality and reach. Smaller providers or newcomers usually have no must-carry status.

With regard to the press, the German "Presse-Grosso" is a distribution system which is largely provider-neutral, is available to all providers, and ensures the delivery of print products with limited circulation to all newspaper kiosks. For the purposes of freedom of the press and press diversity, this service is very welcome and also adequately justifies its exemption from the otherwise widespread prohibition of distribution cartels. But Presse-Grosso is not the only permissible distribution model for press products. Exclusive distribution by publishers is permissible and has recently come back into fashion. What is more, neutrality of press distribution is not based on statutory regulations but is rather the result of all interested publishers and distribution partners successfully organising themselves. Presse-Grosso cannot therefore be used as an example for a network neutrality rule decreed by the state. It may rather serve as a call to all network operators and access providers to agree on a code of conduct establishing a neutral and non-discriminatory distribution system. As is the case with the publishing industry, state regulation could foster this development by easing distribution cartel law accordingly, a move which would be justified in the interest of securing fundamental rights.

With the exception of the FCC's more proactive approach, regulators generally take a wait-and-see attitude. The EU Commission and Ofcom recently carried out similar consultations which reflect a complex picture of different interests and possible solutions. For the rational observer, however, there is no imminent danger requiring state action. Those in support of governmental intervention often use the fundamental, socio-political argument of the freedom and openness of the net which might result in a two-tier society if stronger players were given priority, and thus impair the constitutionally guaranteed fundamental rights of communication of the weaker. This has to be taken seriously, although it is easily forgotten in this respect that large parts of the population are still unable to access the Internet via broadband as there is no DSL access in rural areas. If the issue at hand were truly fundamental in terms of social politics, shouldn't the government have put more effort into ensuring nationwide broadband coverage?

The network neutrality hard-liners also fail to see that the requirement for service differentiation in the context of data traffic management is undoubtedly also connected to the equalising effect of the flat rate price plans on the part of the users. It is generally acknowledged that a substantial proportion of the data traffic is caused by a few so-called "heavy users", who benefit greatly from such flat rates. There is thus no corrective mechanism on the part of the user which suggests treating the scarce net resources economically. This encourages users to collect excessive amounts of data on giga hard disks, even if it is not possible to absorb or process such saved data sensibly. Reasonable flat rate caps and additional fees imposed on excessive data traffic would probably release quite a lot of network capacity, thus rendering service prioritisation largely superfluous. But the advent of flat rates was once also celebrated as a great achievement for the net and their abolition would certainly be highly unpopular. No one in Europe would want to miss out on the Internet radio stations of New York or Buenos Aires which are streamed to us at great expense via glass fibre cables on the seabed of the Atlantic Ocean.

From the network operator's perspective, data traffic management and the development of new revenue sources on the part of the providers would probably be easier to implement than flat rate restrictions. The network operators quite skilfully try to make groups, such as Google, accountable for the data traffic they generate. YouTube is said to generate approximately 25% of all data traffic in the USA; why not make them pay for it? It is certainly correct that companies like Google greatly benefit from network neutrality and the users paying for the transport of their services. This is the basis for the free-of-charge Internet business models which finance themselves via advertising revenues. By contrast, advertising-funded television and radio broadcasters have to finance the transport of their content via conventional distribution channels themselves; the same is true of magazine publishers.

A reversal of the existing model would doubtless contribute not only to a reduction in data traffic volumes, but also to a drastic reduction in Internet diversity, which is surely not in the interest of the users or of media regulation. The result would be the survival of the few large providers able to pay for the Internet transport of their content, but certainly not the small radio stations from overseas. Fortunately, such a scenario is currently not very realistic. It is more likely there will be faster service and QoS guarantees for individual providers that pay extra charges, while the remaining content is transferred at normal speed, with users possibly having to accept small waiting periods. But even this is not unproblematic as regards network accessibility and protection of diversity. Even if network operators dominating the market create special "lanes" for their own content offerings for example, this may be detrimental to other services when it comes to the network bottlenecks. The essential criterion when assessing the permissibility of this approach is the size of the remaining "free" and neutrally-managed part of the network capacity of an operator. A model for such allocation of an operator's network capacity might be provided by the German Interstate Treaty on Broadcasting and Telemedia [Rundfunkstaatsvertrag], in its clauses governing platform regulation, according to which only one third is at the platform operator's free disposal. Data traffic management and prioritisation would then only be permissible for part of the available capacity, whilst the remaining part would be subject to the best-effort principle, as before. The level of service quality outside the special lanes and premium services paid for by the providers should not fall below a guaranteed minimum.

Despite the fact that, according to the coalition agreement, the German Government was planning to speak out in support of guaranteeing network neutrality, nothing at all happened in this regard for a long time. It is only in the recently submitted ministerial draft bill for the amendment of the German Telecommunications Act [Telekommunikationsgesetz – TKG] that this issue has been taken up.

The draft bill for the TKG clearly assumes that such basic-level provision of network neutrality may also be effectively protected by means of consumer information and transparency in competition between network providers. Providing the consumers with appropriate information prior to the conclusion of a contract enables them to choose a provider which operates according to the principles of network neutrality. Information must, in particular, be provided on all procedures established by the network operator for measuring and controlling data traffic so that capacity utilisation or overload is avoided (network management techniques) and also on the possible effects of these procedures on service quality as well as on all restrictions imposed by the provider on the use of the terminal equipment it supplied. In the event of subsequent changes, a user's move to another network operator is to be facilitated.

Consistent with recently amended European directives, the draft bill grants the German Federal Network Agency [Bundesnetzagentur – BNetzA] the possibility of defining minimum requirements for service quality. This should prevent service deterioration and data-traffic obstructions or slow-downs outside of the premium services. It will certainly not guarantee equal treatment in data traffic but will prevent an unacceptable lowering of the besteffort standard if more and more QoS services are introduced.

The protagonists of network neutrality will undoubtedly be disappointed by this draft bill as it implicitly abandons the principle of data equality in the net and apart from that does little more than provide transparency rules and lip service to network neutrality. However, this is the compromise reached at a European level in the 2009 revision of the telecom directives. The German legislator's cautious step-by-step approach provides the industry with the opportunity to first of all seek a self-regulation model for itself which establishes neutral data traffic regulations, continues to reserve significant network capacities for the open best-effort principle and, above all, avoids government intervention.