Dear Sir / Madam,
Please take this Stakeholder comment regarding the BEREC net neutrality guidelines creation into consideration.
The diversity and innovative capacity of the Internet comes from the low cost of innovation and low barriers to entry. These principles ensure that every established business, start-up or non-commercial service — regardless of their size — has an equal opportunity to communicate with a global audience in a manner equal to their competitors. This driving force for the prosperity and diversity of the online economy can only be ensured by an open, neutral and non-discriminatory Internet. When internet providers are allowed to interfere with the decisions of their customers by economic or technical discrimination, this freedom is lost. Recital 1 of the EU Regulation on net neutrality says that legislation has to be interpreted in a way that ensures our freedom to access and distribute information and that protects the Internet as an engine for innovation.
The current BEREC guidelines create a solid foundation for the protection of these principles. The enormous task BEREC was left with by the legislator has been fulfilled in a balanced and careful manner that ensures the protection of the rights of consumers and businesses guaranteed by the regulation. The guidelines provide much needed clarification to the text, but need to be further specified in a few points.
“Zero-rating” is a commercial practice imposed by internet providers. It allows unlimited access to certain sites and services, but imposes a payment for accessing the rest of the internet.
It is good that there are a number of clear restrictions on zero-rating in BEREC’s draft guidelines. However, if most of the current forms of zero-rating are going to be banned or severely restricted, why not ban zero-rating altogether? That would cut five pages from the guidelines and make the job simpler for the National Regulatory Authorities whose job it is to implement the guidelines.
There are forms of commercial practices that interfere with users’ rights protected under Article 3(1) of the EU Regulation to access and, in particular, to distribute information freely. When a commercial practice of an ISP discriminates between providers of content, applications and services by making them unequally accessible (for example, if you have to pay to access some sites/services, but get “free” access to others), this constitutes an arbitrary interference of users’ rights established under Article 3(1) of the Regulation and should be prohibited according to Article 3(2).
Recital 7 of the Regulation defines the types of commercial practices that require national regulators to intervene. However, the language of this recital that “National regulatory and other competent authorities should be empowered to intervene” and “should be required, as part of their monitoring and enforcement function, to intervene” only provides the minimum floor for regulatory intervention and not a maximum ceiling of the scope of this regulation. National Regulatory Authorities have a strict mandate to implement the restriction on harmful commercial practices of Article 3(2) of the Regulation. This means that a slower (and resource-intensive) case-by-case approach is not an appropriate implementation of the legislation.
Application-specific zero-rating (i.e. zero-rating of individual applications or whole classes of applications) and zero-rating for a fee (i.e. where application providers pay to have their data zero-rated) are commercial practices that systematically — regardless of their scale and the market position of the players involved — interfere with the end-users’ right of Article 3(1) to impart information, and therefore materially reduce end-users’ choice in practice. If people have to pay to access YOUR information and get access to other information for free, this is quite obviously a restriction on the right to distribute information, as described in Recital 1 of the legislation. It is therefore logical that such practices be banned under the provisions of Article 3(2).
In addition, National Regulatory Authorities have to ensure clarity and predictability of authorised business models in the digital single market to fulfil the goal of this Regulation to “guarantee the continued functioning of the Internet ecosystem as an engine of innovation”. BEREC’s mandate pursuant to Article 5(3) of the EU Regulation is to contribute to the “consistent application of this Regulation” by issuing clear rules. A case-by-case approach falls short, since the legality of each zero-rating offer will have to be assessed individually by 31 enforcement bodies and radically different patterns of what is permitted and prohibited in each country will accumulate over time, as a direct result of these case-by-case decisons. This legal uncertainty discourages long-term planning and innovation, and is therefore detrimental to investment in the European start-up economy.
Finally, most of the commercial practices highlighted by BEREC have a harmful effect on the fundamental rights of end-users protected under the Charter of Fundamental Rights of the European Union. By making certain services unequally accessible, zero-rating infringes on the media freedom and pluralism (Article 11(2) of the Charter). Zero-rating also constitutes a discrimination against the right to provide services in EU Member states and the freedom to conduct a business for every competitor of the services or applications that are being zero-rated. (see Articles 15(2)), and 16 of the Charter of Fundamental Rights).
The EU Regulation on net neutrality allows specialised services (“services other than internet access services”) under strict safeguards. Article 3(5) and Recital 16 require the optimisation of specialised services to be objectively necessary for the functionality of key features of the service. This would not be the case with services that could also function on the open, best-effort Internet. Furthermore, Recital 16 prevents specialised services from being used to circumvent general net neutrality traffic management rules. Any deviation from these safeguards that would widen the applicability of the concept of specialised services would increase market entry barriers and thus weaken the innovative potential of the Internet as a whole.
If ISPs are allowed to charge online services for preferential treatment, they have an incentive to stop investing in network capacity for the “normal” Internet and reduce their data caps, in order to encourage their customers to use specialised services. This effect would be detrimental for minorities, disadvantaged people, not-for-profit services and start-ups that cannot afford special access to all networks. This would also be detrimental to the development of the free, open and innovative Internet ecosystem.
Specialised services allowed under the EU Regulation must come with their own capacity, separate from the Internet access service. They cannot undercut the average maximum bandwidth that the EU Regulation guarantees.
Paragraph 118 of the draft BEREC guidelines suggests that the delivery of specialised services could limit an individual end-user's Internet access service capacity. This is not in line with Article 3(5) of the EU Regulation. It also contradicts the requirements for the provision of specialised services in paragraphs 113 and 117 of the draft guidelines.
Furthermore, the legislator clearly established its intention to ensure that end-user’s Internet access service capacity remains unaffected by the delivery of specialised services by modifying the wording accordingly during the negotiations. In these final negotiations on 6. July 2015, the legislator decided to delete the word “other” before “end-users” in Article 3(5). That final version of that article now establishes that specialised services cannot be usable or offered to the “detriment of the availability or general quality of Internet access services for end-users.”
Finally, paragraph 118 of the draft guidelines is not in line with Article 4(1)(d) of the Regulation (on transparency) nor with paragraphs 142 and 144 of the draft guidelines, as the average and maximum bandwidth agreed between the ISP and the end-user are no longer met.
The EU Regulation has very clear rules on what constitutes reasonable traffic management. According to its Article 3(3), all traffic management should be as application agnostic as possible. Every deviation from this rule, such as class-based traffic management, could harm competition by offering priority to some classes of applications, but not others, for example.
Class-based traffic management also harms applications and services that are misclassified, whether deliberately or not. This is a particular risk for traffic from small businesses or start-ups. Class-based traffic management also risks discriminating against encrypted and anonymised traffic, which could be throttled by ISPs. It also harms users whose needs for accessing a certain class of service differ from the ISP’s assumptions. Finally, the lack of transparency around this practice creates uncertainty about the performance of particular applications in any particular network. As with zero-rating, the complexity and ambiguity of this approach makes it more difficult for regulators to enforce it. Therefore, applying class-based traffic management instead of application agnostic traffic management is unnecessary, disproportionate, discriminatory and hinders transparency.
In light of these harms, paragraphs 54, 55, 57 and 63 of the draft guidelines are not yet fully in line with the EU Regulation. Article 3(3) subparagraph 2 clearly requires traffic management to be transparent, non-discriminatory and proportionate, in order to be deemed reasonable. Those conditions have to be read in the context of Article 3, to distinguish reasonable and unreasonable forms of traffic management measures.
Paragraph 63 of BEREC’s draft guidelines interpret “reasonable traffic management” in a way which is inconsistent with the legislator's intentions. The draft guidelines would allow far too broad class-based traffic management measures. These would be based on the functionality of the service and protocol used, but it seems clear from recital 9 and the structure of Article 3(3) of the Regulation that the legislator only intended “reasonable measures” to be based on the Quality of Service requirements of traffic (classes for sensitivity to latency, jitter, packet loss and bandwidth).
According to the proportionality principle and paragraph 58 of the draft guidelines, all forms of user-controlled quality of service and consumption based traffic management should be applied and exhausted before more intrusive measures are taken. Therefore, BEREC should bring paragraphs 54, 55, 57 and 63 more clearly into line with the EU Regulation.