After a two-year trialogue, the European Commission, the European Parliament, and the Council of the European Union presented a consolidated draft of the Digital Single Market Regulation (DSMR) on 30 June 2015. The DSMR, among other things, aims at creating a single legal framework for net neutrality in Europe and is expected to be adopted in 2016. The most prominent and controversial feature of the DSMR is that it divides the Internet into “regular traffic” and “fast lane access”, thereby allowing exemptions from the so called principle of net neutrality, that is, the “one lane Internet”.

I. The pros and cons of the “one lane Internet”

Net neutrality means that all data packets are directed through telecommunication networks at equal quality or speed irrespective of their content, purpose, origin, or recipient. According to this best-effort principle, end-users of data-intensive services have no quality guarantee. Consequently, Internet Access Providers (IAPs) are not allowed to require fees from service providers for connecting them with end-users or offer them prioritised transmission of content and application data in return for a fee.

However, unlimited net neutrality can stifle the development of innovative online services, which rely on a minimum quality of data transmission (e.g. connected car technology, virtual reality applications or advanced telemedicine). Giving IAPs the possibility to prioritise the transmission of data streams to particular end-users, thus privileging certain content and applications, might solve this problem to a certain extent. Equally, agreements between IAPs and service providers could ensure that a service could be used with a minimum quality of service (QoS) by implementing smart network management systems or reserving network capacity.

Aside from competition concerns, one of the major arguments against exemptions from net neutrality is that granting separate network capacity for certain services may slow down the best-effort channel, thereby adversely affecting the quality of the (regular) best-effort-based internet. Exemptions from net neutrality could also impede the rise of start-up companies, which, at least in their initial set-up phase, will often lack the funds to purchase privileged network access or QoS.

In an attempt to resolve this conflict, the US Federal Communications Commission (FCC) adopted the Open Internet Rules and introduced a two-tier solution in Section 706 of the Telecommunications Act 1996 and Title II of the Communications Act in early 2015. The European legislator has opted for a similar compromise: Under Art. 3 (3) DSMR, discrimination between, restriction of and interference with data traffic are strictly prohibited in order to ensure an “open internet”. Data streams have to be treated equally unless unequal treatment is objectively justified. According to the European Commission, any kind of paid prioritisation of data traffic shall be prohibited. However, the DSMR allows for three controversial exemptions from net neutrality: traffic management measures, “zero rating”, and specialised services.

II. Traffic management

Traffic management measures are technical measures that enable IAPs to selectively manage data streams on their network. Such measures allow “fast lanes” for certain categories of data. Traffic management can also be used to reserve network capacity for certain users or block particular types of content.

However, under Art. 3 (3) DSMR, IAPs are only permitted to take reasonable traffic management measures. Such measures are supposed to contribute to a more efficient use of network resources and improve the general quality of data transmission in support of end-users. In order to be deemed reasonable, traffic management must be transparent, non-discriminatory, proportionate with regards the objective of general improvements in quality, and based not on commercial considerations but rather on objectively different quality requirements for certain traffic categories.

Furthermore, network management measures are also permissible on an exceptional basis in order to protect certain public interests, e.g. facilitate criminal law enforcement, maintain the security and integrity of the network, and reduce network overload, for instance during peak watching hours of football games.

III. Zero rating

Zero rating describes the practice, seen predominantly in mobile telephone markets, where certain content or applications are not included in end customers’ data usage. For instance, several IAPs in Europe have offered Spotify in bundled contracts for the provision of mobile phone services with limited data volumes and unrestricted use of Spotify.

The provision of zero rating services is allowed under Art. 3 (2) DSMR, since end-users are free to agree on a certain speed for internet access and data-volume tariffs with IAPs. Such agreements and commercial practices are only prohibited if they infringe on the end-user’s right to use the internet access of their choice, access and distribute information and content without hindrance, and use and offer applications and services without hindrance.

As a rule, zero rating offers are unlikely to constitute such a hindrance in a competitive IAP market, since end-users are still able to use the zero rating services via other IAPs. However, if a certain service is available exclusively in a bundle with a particular IAP and this curtails the end-user’s right to freely choose internet access, say because of a monopoly, this may constitute a violation of Art. 3 (2) DSMR. In such cases, antitrust remedies might also be worth considering.

IV. Specialised services

According to Art. 3 (5) DSMR, IAPs are allowed to grant QoS to specialised services or offer specialised services themselves. Specialised services are “internet access services […], which are optimised for the specific quality requirements of specific content, applications or services, or a combination thereof, where the optimisation is necessary in order to meet the requirements of the content, applications or services for a specific level of quality”. The assessment whether the conditions of this definition are met is likely to be done primarily on the basis of the relevant technical requirements. The European Commission, for example, classifies IPTV services, high-definition video conferencing, automated driving and e-health services (e.g. remote surgery) as specialised services. The common feature of these services is that they are particularly quality-sensitive and cannot be used as expected by either service providers or end-users without guaranteed fast and stable data-transfer rates.

Consequently, whether a service shall be qualified as a specialised service could be determined on the grounds of a comparative test: the question at hand would be whether the relevant service may also function without a certain QoS or whether a certain QoS is of material significance to its proper operation. User expectations are likely to play a crucial role in the implementation of this test. In light of the open definition of the term “specialised services”, the future challenge for regulatory authorities and courts will be to assess user expectations and preferences in an (empirically) plausible manner in order to determine the necessity of prioritised data transfers. However, a service does not qualify as specialised if it also works without QoS and QoS has only been purchased in order to secure competitive advantages vis-à-vis competing service providers.

1. Borderline cases

There will be borderline cases when services that have operated without QoS in the past wish to use QoS in order to increase the quality of their service. Consider the case of search engines. Today, a search engine is defined more by the accuracy of its hits and less by its speed. Nevertheless, speed does play a role for providers and users. This is why Google prominently displays the time taken for a search query above the first hit. In this context, it is questionable whether a search engine operator could be classified as a specialised service and purchase QoS only because of its intention to speed up its service. It is open to debate whether this assessment would be different if the same search engine operator also used QoS to implement a data-intensive optimisation of its search algorithm – a service improvement that would not have been possible without QoS.

Another problem is to what extent the purpose of a service and the content of the transmitted data need to be considered when classifying a service as a specialised service. Consider networked driving or telemedicine services. The mere goal of these services provides a sensible argument for prioritising them. Yet networked driving and telemedicine might be services which, from a technical perspective, do not require QoS to be functional. If technical necessity was the only decisive criterion, the service could not benefit from prioritisation, whereas under a purpose-oriented or content-specific view, the service could be classified as specialised.

2. Maintaining sufficient network capacity

Regardless of what services are classified as specialised in the future, IAPs will not be allowed to exploit the exceptions from net neutrality to obtain improper advantages or curtail end-users’ liberties. Consequently, Art. 3 (5) DSMR provides that specialised services may only be offered if the network capacity is sufficient to provide them in addition to any “normal” internet access services. Furthermore, specialised “services shall not be usable or offered as a replacement for internet access services”. In practical terms, this means it would not be permissible to channel internet traffic through VPN services or proxy servers and have this traffic privileged under the specialised services exemption.

Ultimately, specialised services shall not be to the detriment of end-users in relation to the availability or quality of internet access. Whether the use of specialised services may have such an effect is to be measured primarily in accordance with minimum quality requirements for best-effort traffic – the minimum QoS. This minimum QoS will be established by the national regulatory authorities pursuant to Art. 4 (1) DSMR. The intention behind this provision is to ensure “levels of quality that reflect advances in technology”.

However, it is not clear how minimum QoS should be determined. Determining a high minimum QoS would reduce the network capacity available for specialised services. Assuming that QoS can be granted in return for a fee, this would likely entail an increase in the prices for QoS and keep the number of specialised services low. Particularly for specialised services that lack funds, such as start-ups, this would make market access more difficult. Conversely, establishing a low minimum QoS might lead to an increased range of specialised services. However, there is a risk that the general quality of the “open internet” would suffer due to the increased use of network capacity by specialised services.

These considerations also tie in with the question whether a minimum QoS should be established for Member States’ national territory or for individual, geographically defined areas. To date, there are considerable regional differences concerning the availability and quality of internet access. As such, the quality of internet access in cities and urban conurbations is generally high, whereas it is comparatively poor in many rural areas. Determining a minimum QoS on a national level and using urban regions as a basis for a minimum standard might make it impossible for IAPs to ensure the minimum QoS in rural areas. Conversely, using rural areas as a benchmark could result in a loss of quality in cities.

Finally, it is not clear which procedure will be followed if an IAP has already allocated its excess network capacity to particular specialised services and more capacity for other specialised services cannot be provided without harming the best-effort traffic. The intended purpose of the specialised service might also play a role in this context so that specialised services, which aim at protecting common goods or valuable legal interests might be granted privileged access vis-à-vis other specialised services in instances of increased demand.

V. Towards a fragmented European market?

The aforementioned examples show that there are different approaches on how to apply the exemptions from the principle of net neutrality. Therefore, although the DSMR will be directly applicable in the Member States, it is likely that national regulatory authorities will have wide discretionary powers to impose their own net neutrality regimes. There is a risk that the DSMR will not adequately guarantee harmonized standards and legal certainty within the EU, especially when it comes to contractual agreements between IAPs and specialised services. Against this background, the European legislator should clarify the scope of the exemptions under the current draft of the DSMR to make sure that the initial objective, namely to boost innovation of (international) high-end online services and create a level playing field, can ultimately be achieved.