A draft version of the EU regulators’ guidelines on net neutrality has been leaked in a French magazine and here (PDF, 20 MB).
First off, a quick overview of the regulatory process: the EU enacted the Net Neutrality Regulation in November 2015. However, that text is full of ambiguities and potential loopholes. The Body of European Regulators of Electronic Communications (BEREC) was therefore tasked with creating guidelines to clarify the meaning of this law. BEREC has until the 30th of August to adopt the final guidelines. On Monday, BEREC will publish the draft version of its guidelines and start a public consultation that runs from the 6th of June to the 18th of July. This leak is a draft version of BEREC’s Guidelines as of mid-May 2016.
Right off the bat, we can say that it’s not all bad. BEREC seems to have covered the fundamental aspects of net neutrality. The Guidelines make it clear that the worst violations of net neutrality are prohibited, such as arbitrarily blocking certain websites. That being said, the Guidelines also contain major loopholes on three main issues: zero-rating, traffic management and ’specialised services‘.
Zero-rating: guidance towards a wild west anarchy
Zero-rating is a practice, currently common throughout Europe, where traffic from certain sources does not count towards people’s monthly data cap. For example, you could have a 2GB limit for ’normal‘ internet usage, but Spotify would not count against that limit. Applications that do not count against people’s data caps have a major competitive advantage. Thus, zero-rating allows ISPs to make certain applications more attractive than others, thereby picking winners and losers online. That’s a huge net neutrality problem.
Ideally, the EU would follow the examples from India and the Netherlands, and outlaw zero-rating entirely. The Regulation allows BEREC to do so: commercial practices like zero-rating that restrict the freedom of end-users to use and offer services can be prohibited. However, the leaked Guidelines fail in this regard. Instead of outlawing zero-rating, they are moving towards a ‚case-by-case‘ approach where it is always left to the telecom regulators to decide about each zero-rating deal individually. Some would be prohibited, but others would be allowed – and these outcomes would difficult to predict.
This is a missed opportunity. Not only because it will permit certain net neutrality violations, but also because it creates complete legal uncertainty. ISPs will be tempted to push the limits of these vague rules, while regulators will have a hard time monitoring, assessing and enforcing the rules in this grey area (which they themselves have created!). While ISPs and regulators fight over the details, users will be pushed towards applications of the ISPs’ choosing. Content providers like start-ups will be discriminated against, because they have no chance against zero-rated competitors. Neither will know whether they will be protected until the regulators have made a decision on each specific zero-rating deal in question.
Instead of a predictable digital single market that attracts investment, the EU ends up with an uncertain wild west situation where the strongest players can push the limits for years to come and every country will find their own solutions. BEREC’s case-by-case approach reeks of compromise, and leaves everyone worse off – except perhaps litigation lawyers. (paragraph 28–45)
Traffic Management: The death of the Best Effort Internet
The Guidelines also contain weaknesses when it comes to ‚traffic management‘. These rules affect how ISPs manage their networks. This can be done in a neutral, application agnostic way, or by discriminating against certain forms of traffic; for example by throttling video content or voice chat. To prevent such discrimination, traffic management needs to be as application agnostic as possible, and application-specific measures should only be taken under exceptional circumstances. In general, the leaked guidelines support these principles – with one big exception: the Guidelines would allow ISPs to move towards a system where different applications are separated into different quality classes at all times even in the absence of congestion. Such a class-based system leaves far too much wiggle room for ISPs to discriminate on a day-to-day basis. (paragraph 63)
Such a class-based Internet allows ISPs to distort competition by offering priority to some classes of applications but not others; it harms applications that are misclassified, whether deliberately or inadvertently; it harms start-ups by making it harder for them to get the type of service they need; it risks discriminating against encrypted and anonymized traffic; it harms users whose needs for a certain class of service differ from the ISPs’ assumptions; it is less transparent and creates uncertainty about the performance of particular applications in any particular network; and finally, as with zero-rating, the complexity and ambiguity of this approach makes it more difficult for regulators enforce.
Many ISPs are eager to “optimize their network” with this class-based system because it offers a short-term relief for a structural problem: what European ISPs should really be doing, is investing in their network. We will end-up in a situation where best-effort Internet is a thing of the past and the quality of each and every application is decided by your ISP. Maybe this is the Internet which Commissior Günther Oettinger wants. (the Commission was part of all negotiations within BEREC)
Specialised services may eat your internet away
‚Specialised Services‘ is a new concept in the net neutrality debate. It was first coined by the European Commission to describe an indeterminate group of new services which could not possibly function over the ‚regular‘ Internet and would therefore require special treatment. This group would be exempted from net neutrality rules, creating major loopholes if defined too broadly.
Thankfully, BEREC appears to have interpreted ’specialised services‘ in a sensible way: According to the leaked Guidelines, ‚specialised services‘ cannot create paid fast-lanes for old internet applications. They have to be new services that cannot function over the open Internet. Otherwise, all net neutrality rules could simply be circumvented. The Guidelines require that specialized service offerings are not to the detriment of regular internet access services and may only be offered if the ISP has sufficient network capacity to offer them in addition to any internet access service. The regulation and the guidelines are very clear on this point. But in one of the last negotiations in April text was included that completely contradicts this safeguard.
BEREC has decided that specialised services may take away bandwidth from the users’ Internet connection, so long they do not affect the quality of other users’ Internet access. Specialised services may now cannibalise one’s own internet access service up to the point where only the minimum speed is not undercut. Normally you would expect that when you buy an ‘extra quality’ service from an access provider, you would get this service with bandwidth on top of what you already have. For some inconceivable reason BEREC decided to do the exact opposite, although the regulation and the guidelines are very clear on this point. If these inconsistencies are not clarified in favor of net neutrality, specialised services could create major headache for regular internet users and “eat away” the bandwidth of the normal internet. (paragraph 118) [see details below]
In addition, there is a worrisome loophole about the determination if an application qualifies as specialised service’ . The guidelines seem to suggest that the provider of a service can set the quality requirements of its service at their own discretion instead of an objective assessment based on the real quality requirements of the application. Those requirements are vital for the decision if something qualifies as specialised service. That contradicts the clear text of the regulation and could allow ISPs to use specialized services to circumvent net neutrality protections like the ban on fast lanes. (paragraph 102 & 104)
Next Steps
The leaked Guidelines should give rise to careful optimism, but not to complacency. To get foolproof net neutrality safeguards, there is much room for improvement. There is also much to lose. The hard-fought progress on net neutrality could be undone entirely if the Guidelines are watered down any further.
It’s certain that industry lobbyists will be working around the clock to sway the consultation in their favour. In the face of this onslaught, we need a strong response from the countless stakeholders who benefit from an open internet, from average users, gamers and musicians to start-ups and government agencies. This is why www.savetheinternet.eu is running an online tool to facilitate participation in the consultation. Every response brings us a step closer to net neutrality in Europe.
On the 2nd and 3rd of June, the BEREC plenary in Vienna will decide on the draft that will enter the official public consultation. A demonstration in Vienna will bring the demand for real net neutrality to the streets. Sources close to the negotiators say there is a coordinated effort by some countries to introduce new loopholes into the Guidelines at the last minute. This is particularly worrisome since, in early May, the BEREC drafters (technologists and economists) handed their work to the international relations department (mostly lobbyists).. For one month, the political battle within BEREC ensued before the public had even a chance to look at the Guidelines. It will be interesting to see which Guidelines will be presented on 6th June at the BEREC press conference in Brussels.
**explanation on the double selling of bandwidth**
This ties into an improvement on transparency. Users will no longer get only advertised “up to” speeds. Additionally they will get accurate minimum, average and maximum speeds about their internet connection assured.Imagine your ISP sells you a 10 Mbit/s Internet connection. With the new rules your contract ensures that you always get maximum 12 Mbit/s, average 7 Mbit/s and minimum 4 Mbit/s. Then the same ISP sells you a 5Mbit/s IPTV specialized service. Instead of having now 15 Mbit/s in total the ISP sells you 5 Mbit/s twice.