In recent years as regulators on both sides of the Atlantic have grappled with questions about net neutrality, there is at least one commitment they have made in common: to monitor market developments for signs that discriminatory or anti-competitive conduct may be chipping away at the openness of the Internet. While there has been minimal news out of the FCC on this front, European regulators have sprung into action. Recently, Europe’s consortium of telecom regulators, the Body of European Regulators for Electronic Communications (BEREC), released findings from a survey of 381 network operators (266 fixed and 115 mobile) concerning how they manage traffic on their networks. For those in the US who are accustomed to largely unfettered Internet access options, the BEREC findings may come as something of a shock, although the gaps in BEREC’s data make it difficult to draw completely firm conclusions.
The bulk of BEREC’s report analyzes the extent to which European operators restrict (“block/throttle,” as it is described in the survey) or provide some preferential treatment to particular types of Internet traffic, including peer-to-peer and VoIP traffic. On the fixed side, peer-to-peer traffic is the most common target – 21% of fixed subscribers in the EU face peer-to-peer traffic restrictions. Because of how p2p-restricting operators are distributed across Europe, 25% of European fixed line subscribers – 37 million users – live in countries where the majority of the ISP market is comprised of operators that restrict peer-to-peer traffic on every package they offer.
While mobile operators also target peer-to-peer traffic in some cases, their more common target is VoIP, with at least 21% of subscribers having VoIP-restricted packages. BEREC did not collect precise figures from operators that restrict VoIP on some Internet packages but not others, which means that in reality this figure could be 30% or higher. The majority (58%) of Europe’s mobile Internet users live in countries where the mobile market is dominated by operators that offer a mix of unrestricted and VoIP-restricted packages.
BEREC’s data reveal that these practices are not limited to small fringe ISPs. The average subscribership of ISPs that restrict peer-to-peer or VoIP applications is larger than the overall average of all European operators. For example, mobile operators that restrict VoIP on all their packages average 11 million subscribers each, compared to an overall European average of 1.85 million subscribers per mobile operator. Large operators are clearly an important part of the constituency of ISPs that restrict application usage.
As in the US, some European operators are offering their own specialized voice or video services, some of which share network resources with the operators’ Internet offerings. For 35% of fixed operators, these specialized services are architected in such a way that they could potentially affect the Internet services that share the same pipe, although the report does not go into further detail about the nature or circumstances of such effects.
Gaps in the Data
In some ways the BEREC report is more notable for what it does not say than for what it does say. The survey provided to operators seeks many of the same details that regulators – including Digital Agenda Commissioner Neelie Kroes — have been asking for operators to disclose. But operator-specific information is nowhere to be found in the report. The report likewise omits country-level data, instead presenting aggregations of national markets that reveal little about the prevalence of application restrictions in larger or smaller countries, or those where national regulatory authorities have been more or less active, for example. For all their talk about transparency, it seems at least a bit odd for the European regulatory authorities to be sitting on the very information whose disclosure they have been persistently calling for.
Furthermore, the survey leaves out a number of facts that are critical to understanding Europeans’ actual Internet experiences and choices: the differences between blocking and throttling, the consequences of contractual (as opposed to technical) constraints, and the differences in price or other factors between restricted and unrestricted packages.
The BEREC survey asks operators about whether certain applications are “blocked/throttled.” But there are obviously significant differences in user experience between having an application blocked versus throttled, and even the term “throttled” is highly ambiguous. An Internet service offering on which it is impossible to use peer-to-peer applications is arguably very different than a 2 Mbps package on which peer-to-peer traffic is restricted to 1 Mbps at peak times. Lumping these kinds of practices together makes it difficult to determine how restricted offerings in Europe actually are. BEREC also mentions (on four occasions) that restrictions may have less of an impact than the numbers would otherwise suggest since they are often only applied at peak times, but the report gives no indication of how often that is the case or the duration of operator-defined peak periods.
The report also emphasizes whether application restrictions are enforced technically or enforced by contract. But there is no information provided about the consequences to violators in cases where restrictions are enforced by contract. If these contracts are consistently enforced, what is the value of presenting the restrictions separately in the data, since the consequences to users are similar to those who experience technical restrictions? If contractual restrictions are not enforced, then why do operators have them in the first place? BEREC provides no explanation for its emphasis on this distinction.
Finally, the report discusses cases where individual operators restrict applications on some packages but not others, but it is difficult to draw conclusions about the extent to which these choices are meaningful without knowing more details about the offerings. For example, if unrestricted packages are significantly more expensive, require longer contracts, or are only available with specific mobile handsets, then subscribers may feel that they have less of a choice between restricted and unrestricted offerings than the mere number of packages of each type implies. Given how many operators provide a mix of offerings, particularly with respect to mobile VoIP, these details are vital to conducting a full analysis.
Although it is not clear if or when BEREC will further refine or expand its survey and report back again, the kinds of details discussed above will hopefully form part of any analyses that individual national regulators may conduct. In the UK, for example, Ofcom has committed to regularly report on traffic management practices, perhaps publishing its first report as soon as this summer.
Although discussions of transparency tend to focus on the information that operators publish themselves, there is no reason why the systematic collection of traffic management data by regulatory authorities should not be considered as an important complementary information source to operator-published policies and independent technical measurements. Each of these sources provides different kinds of granularity and standardization of traffic management data. Together, they should help provide an informed basis for regulators to determine whether and when their intervention is necessary.