Study on net-neutrality regulation | 18 Concerns over the Xbox case declined as the issue turned out to have limited commercial impact. The Apple FaceTime case was complex. A trade press article at the time reported that “Apple’s FaceTime, which allows live video conversations between users of Apple devices, has worked [to date] only over Wi-Fi. But Apple is changing that, opening the Skype-like service to function over cellular connections. The change comes when Apple’s newest mobile-phone operating system debuts Wednesday and will spread even wider once the new iPhone 5 starts landing in hands Friday. AT&T says it will make the video-chat service available on its cellular network for those with generally more expensive, shared data plans, which the company unveiled last month. […] Among other things, the company says that it is simply a business decision to use FaceTime as a hostage to move recalcitrant customers to a new plan.” 50 The FCC held discussions with AT&T, an FCC advisory group published an analysis (without 51 reaching consensus as to how to interpret the issue), but no formal enforcement procedures were initiated. AT&T had been slow to respond, but eventually claimed that the practice had been an unintentional “technical glitch”, that it had already resolved the issue for customers who had complained, and that it intended to resolve it for all customers.           52 Recent developments on zero rating A major recent initiative related to the implementation of zero-rated plans by all US mobile network operators (MNOs). The Open Internet Order of 2015 did not impose a flat prohibition on zero rating, but rather committed the FCC to case-by-case analysis. The FCC noted that, in 2016, “T-Mobile significantly expanded the number of participating, zero-rated edge providers in Binge On, an offering introduced in November 2015 that zero-rated standard definition video […] Sprint has also introduced unlimited data through its Unlimited Freedom and Unlimited Freedom Premium plans, and experimented with zero-rating of the 2016 Copa America soccer tournament. AT&T and Verizon did not introduce stand-alone unlimited data plans, but have eliminated overage fees in some instances for customers who exceed their data caps and launched their own zero-rating and sponsored data programs. For example, AT&T Mobility offers a “Data Free TV” feature on its DIRECTV app that enables its broadband consumers who also subscribe to direct broadcast satellite service from AT&T’s wholly-owned affiliate, DIRECTV, to view unlimited DIRECTV video content with no impact on the user’s mobile data monthly allotment. […] Verizon launched its FreeBee Data and FreeBee Data 360 sponsored data programs in January 2016.” 53 This rapid expansion of zero-rated offers inspired the FCC’s WTB to develop policy guidance, issued in a report late in 2016.53 In parallel with the report, it appears that enforcement investigations 50     David Kravets (2012), “Net Neutrality Groups Challenge AT&T FaceTime Blocking”, Wired, 18 September 2012; see https://www.wired.com/2012/09/factime-fcc-flap/ (viewed 6 May 2017). See also Open Internet Advisory Committee (OIAC) to the FCC (2013), AT&T/FaceTime Case Study, 20 August 2013. 51     Mobile Broadband Working Group / Open Internet Advisory Committee / Federal Communications Commission (2013), AT&T/FaceTime Case Study, 20 August 2013. 52     See http://www.fiercewireless.com/wireless/at-t-facetime-problems-gophone-technical-nature 53     US FCC Wireless Telecommunications Bureau (2016), “Policy Review of Mobile Broadband Operators’ Sponsored Data Offerings for Zero-Rated Content and Services”, pp.2–3. Ref: 2009152-254                                                     .

Study on net-neutrality regulation | 19 were initiated against AT&T Mobility and Verizon, which were both identified in the WTB report as using zero rating in ways that might violate net-neutrality principles. The concerns raised in the WTB policy guidance were, first, that the AT&T and Verizon plans seemed to favour their own content or affiliated content over unaffiliated content (thus posing an economic risk of vertical foreclosure), and second that the market shares of AT&T and Verizon were large enough that the practice could raise competition concerns. In WTB’s preliminary judgment, these factors did not appear to be present for the T-Mobile or Sprint plans. The policy guidance did not reach final conclusions, but rather noted that neither AT&T nor Verizon had yet provided an explanation that the WTB found adequate. Among the very first actions undertaken when Ajit Pai became FCC chairman in January 2017 were (1) rescission of the WTB report on zero rating (which had no operative effect in any case), and (2) termination of any investigations of AT&T and Verizon (and T-Mobile) for their zero-rating practices. 54 Thus it is not clear how the FCC would have finally ruled in these cases, based on its interpretation of the Open Internet Order of 2015. 3.3 Transparency obligations on ISPs in relation to practices which may affect net neutrality In this subsection we consider the steps taken to ensure that ISPs in benchmark countries are transparent about the services they offer, including details of traffic-management procedures. 3.3.1 Chile The net-neutrality regulation in Chile gives specific guidance on the information which ISPs must provide to meet the requirements of the law. Of note are the following articles from the regulation: •    Article 3 requires ISPs to measure technical QoS indicators on a quarterly basis, in line with ETSI standard EG 202 057-4 V1.2.1 (2008-07) 55 (Section 5) •    Article 4 requires ISPs to measure Internet access service replacement times on a quarterly basis, in line with ETSI standard EG 202 057-1 V1.2.1 (2005-10) 56 (Section 5.5) •    Article 5 sets out the specific information that must be reported by ISPs, which includes: — Commercial characteristics of the Internet service, including upload/download speed, download limits and guarantees of service — Contention ratios — Technical indicators (user access time, data transmission rate, proportion of failed data transmissions, proportion of successful user accesses, delay) 54     US FCC WTB (2017), “In the Matter of Wireless Telecommunications Bureau Report: Policy Review of Mobile Broadband Operators’ Sponsored Data Offerings for Zero Rated Content and Services”, DA 17-127, 3 February 2017. “Today, the Bureau sent letters to AT&T Mobility, T-Mobile, and Verizon Wireless closing the inquiries into each company’s sponsored data and zero-rating offerings, taking no further action. By this Order […] the Bureau now sets aside and rescinds the Policy Review Report and any and all guidance, determinations, and conclusions contained therein, including the document’s draft framework.” 55     See http://www.itu.int/itu-t/workprog/wp_a5_out.aspx?isn=6051 56     See http://www.etsi.org/deliver/etsi_eg/202000_202099/20205702/01.02.01_60/eg_20205702v010201p.pdf Ref: 2009152-254                                                     .

Study on net-neutrality regulation | 20 — Replacement time of the service — Quality and availability of the link — Traffic- and network-management practices, including their characteristics and effects on the service provided to users. The information includes the types of applications, services and protocols that are affected •    Article 6 sets out a requirement for the provision of information through SUBTEL’s information-gathering portal (STI) •    Article 7 reiterates the principles of the law. In addition, it acknowledges that traffic- management practices may be carried out, providing that the actions do not affect competition. Article 7 also requires that any traffic-management actions are described to users in a clear and intelligible publication. Of note to this study is that the information regarding traffic-management practices is submitted only in qualitative form: no numerical (measurement) data is required. Also, under the law, ISPs must submit the information via SUBTEL’s STI portal (discussed in more detail in Section 4.2). SUBTEL also monitors the commercial terms of operators’ broadband plans via the information published on their websites. There is no specific consideration of specialised services; the law applies in a general sense to all services on public (i.e. non-private) networks. 3.3.2 India 57 ISPs submit Performance Monitoring Reports to TRAI every quarter. At present, there are no net- neutrality-specific transparency requirements placed on ISPs (although TRAI included this issue in its recent net-neutrality consultation). The recent consultation has considered the following elements of transparency information in relation to net neutrality: •    price information and commercial terms (e.g. price, promotions, fair use and data caps) •    performance characteristics (e.g. advertised speed, actual speed, minimum speed, latency, packet loss, suitability for real-time applications) •    traffic-management practices (e.g. congestion management, bandwidth throttling, preferential treatment and blocking, and whether such practices are applied to a category/class of traffic or user, and the triggers and time periods) •    specialised services (including monitoring of their impact on delivery of Internet services). The consultation has also looked at the following key options for how information is disclosed: •    disclosures provided directly to consumers by an ISP •    disclosures to the regulator 57     See http://www.analytics.trai.gov.in:8001/trai/qos/method.php Ref: 2009152-254                                                     .

Study on net-neutrality regulation | 21 •   disclosures to the general public, or •   a combination of the above. 3.3.3 USA The FCC has provided a non-exhaustive list of what information should be disclosed by ISPs. These guidelines were originally developed in 2010 (and have remained in effect, since the transparency aspects of the 2010 Open Internet Order were never overturned by the courts), but were strengthened in 2015 and expanded to include pricing aspects. Under the Open Internet Order of 2015, ISPs are required to disclose speed, latency and packet loss. For speed, the Order talks about two distinct things: expected performance versus average performance (an average or median). ISPs that voluntarily participate in the Measuring Broadband America programme and make their results public are deemed to have complied with these requirements, and thus granted “safe harbour”. Following the release of the 2015 Order, the Chief Technologist, Office of General Counsel and Enforcement Bureau subsequently issued more detailed guidance on what should be disclosed, and how. 58 The detailed guidelines include, for instance, information on how “actual” performance compares with median or quantile speed measurements. The guidelines include both fixed and mobile broadband. Transparency requirements relating to traffic management are similar to those in Chile, in that they are qualitative. ISPs must provide: •   details of application-specific network-management practices •   details of subscriber-triggered network-management practices •   a pointer to a web page for more details. Regarding the requirement to describe network-management practices, ISPs are asked to “provide a brief description and a link to a full discussion that identifies application-specific [or subscriber- triggered] network management practices, when such practices are triggered, and the effect such practices could have on performance”. 59 There is no systematic, periodic review of the disclosures that ISPs make to actual or prospective customers; rather, the process is complaint-driven. The FCC may conduct spot checks to ensure that disclosures are clear. 58    US FCC (2016), “Guidance on Open Internet Transparency Rule Requirements”, GN Docket No. 14-28, DA 16-569, 19 May 2016. 59    See https://apps.fcc.gov/edocs_public/attachmatch/DA-16-357A1_Rcd.pdf Ref: 2009152-254                                                 .

Study on net-neutrality regulation | 22 3.4 Monitoring and supervision by NRAs In this subsection we consider the steps that NRAs in the benchmark countries have taken in the past, are currently taking, or intend to take to monitor/supervise ISPs for practices that may violate net neutrality. 3.4.1 Chile Following enactment of the net-neutrality law, SUBTEL began a programme of collaboration to help define the scope and process for ISPs to report the information required by the law. Two initiatives were enacted, both with the University of Chile: •    “Consultancy to elaborate, implement and monitor Quality Indicators for Internet Access Service in Chile”, which was awarded to the University’s Foundation for Technologic Transfer. This resulted in the Adkintun platform, which allowed users to measure some of the quality indicators of their connections by themselves. (The Adkintun platform is described in more detail in Section 4.1, including its development into a mobile-only app.) •    “Technical Consultancy for the review and analysis of a measuring protocol for quality indicators on Internet access”, which was awarded to the University’s Faculty of Physical Sciences and Mathematics. The output of the work was the definition of measurement methodology using probes that simulate the behaviour of users. The methodology defined the temporal and geographical requirements needed to provide a statistically representative sample. The University of Chile’s initiatives were concluded in 2013. In addition to defining the scope and process for ISP reporting, the two initiatives provided knowledge transfer to SUBTEL, to assist in building its capacity for tackling net-neutrality issues. Another “tool” used by SUBTEL is a regular customer satisfaction survey. Every year SUBTEL carries out a survey called Encuesta de Acceso, Usos y Usuarios de Internet 60 (Survey of Internet access, uses and users). It takes also into account the evolution of customers’ expectations, as a measure of their satisfaction with network quality. Finally, SUBTEL is developing tools to analyse the data that comes in through the STI portal. In September 2017, SUBTEL is hoping to implement a tool that will help it to analyse the large amounts of statistical data it receives. In the first instance, the tool will be used for traffic flows, but in the longer term it could be developed to consider net-neutrality-specific issues. 3.4.2 India At present, TRAI does not undertake any net-neutrality-specific monitoring or supervision. ISPs do throttle traffic to manage QoS, but no tools are being used to measure this. 60     See http://www.subtel.gob.cl/wp-content/uploads/2015/04/Informe-VII-Encuesta-de-Acceso-Usos-y-Usuarios-de- Internet_VF.pdf Ref: 2009152-254                                                   .

Study on net-neutrality regulation | 23 TRAI’s Analytics Portal is focused on the general QoS of India’s mobile communications providers. 61 The portal allows users and service providers to explore and resolve various issues related to different telecoms services. It has three portals for different QoS parameters: •   TRAI MySpeed Portal: allows users to explore the mobile data experience of customers across India. Users can submit data by downloading the app and testing their data speeds •   TRAI Drive Test Portal: allows users to explore the results of drive tests carried out by TRAI to independently check coverage and QoS •   TRAI QoS Analysis Portal: allows users to explore the call drop rate of various telecoms service providers (TSPs). Through this portal the QoS performance of service providers can be identified for any specific location in India. Users can also navigate and view performance metrics from service area to district to city, and finally to the base-station (BTS) level. There are defined complaints procedures available to consumers in India: •   a portal on TRAI’s website (which receives a tracking number from the ISP involved), and •   the Consumer Redressal Forum, a quasi-judicial body which adjudicates to resolve issues. TRAI has also considered the issue of monitoring in its recent consultation. The options being considered for monitoring of non-net-neutral practices are: •   disclosures and information from TSPs •   collection of information from users (complaints, user-experience apps, surveys, questionnaires) •   collection of information from third parties and public domain (research studies, news articles, consumer advocacy reports). 3.4.3 USA Monitoring and supervision arrangements in the USA are primarily driven by complaints. Other than the Measuring Broadband America programme (see Section 4.4), there is no systematic or periodic monitoring. The Measuring Broadband America programme serves mainly to help consumers determine whether their ISP is honouring its QoS commitment. (In the USA, there is no specific legal or regulatory obligation to deliver the advertised speed; however, a failure to do so could possibly be viewed as a misleading practice under the Transparency Rule of the Open Internet Order of 2015, or might perhaps be viewed as a violation of Sections 201 and/or 202 of the Communications Act of 1934 as amended, the general law that governs telecoms in the USA.) At the same time, overall throttling of throughput would likely be visible in the aggregate statistics produced by the Measuring Broadband America programme, and so in that sense it can be viewed as a net-neutrality monitoring tool. 61    See http://www.analytics.trai.gov.in/ Ref: 2009152-254                                            .

Study on net-neutrality regulation | 24 Another tool that has been recognised by the US FCC is Netalyzr. Netalyzr was created by the International Computer Science Institute (ICSI), an affiliate of the University of California at Berkeley. It is a Java applet that runs on the end user’s device and communicates with ICSI servers in order to measure performance through the network, with a particular focus on practices of the ISP that directly serves the end user. We discuss Netalyzr in more detail in Section 4.3. 3.5 Legal mechanisms for enforcement of net neutrality by NRAs In this subsection we provide an overview of the legal mechanisms available to NRAs in benchmark countries to enforce national net-neutrality rules on ISPs. 3.5.1 Chile Enforcement of net neutrality in Chile is based on the 2010 law (see Section 3.1.1). In general, SUBTEL acts mainly in response to complaints from customers. One of the most common complaints relates to degradation of service. This does not make use of net-neutrality law (which specifies no minimum level of service), and can be addressed under general telecoms law (by refunding users). If a net-neutrality violation does occur, a fine cannot be imposed immediately. Once SUBTEL learns of the incident, it launches a “charge procedure”. It then visits the ISP and inspects the infrastructure. A technical report is sent to the legal team, which prepares a statement of fine. The operator can respond to this, but then the state responds in turn with a final assessment of the matter. This is standard procedure under telecoms law, and is not specific to net neutrality. 3.5.2 India TRAI cannot enforce net neutrality without a related regulation, and therefore it currently can only take action in situations relating to differential tariffs such as alleged zero rating (see Section 3.1.2). The current consultation on wider net-neutrality issues may result in a wider-ranging regulation under which TRAI can act on a broader range of issues. In the event of a violation of any of its regulations (once they are defined), TRAI is able to impose monetary penalties. In particularly serious cases, the ISP’s licence could be removed by the Department of Telecommunications. 3.5.3 USA The FCC Enforcement Bureau takes responsibility for investigating complaints that are serious enough to potentially warrant fines or penalties. These investigations remain confidential until and unless a determination of apparent liability is made, at which point they become public. Ref: 2009152-254                                              .

Study on net-neutrality regulation | 25 The procedures followed in the case of apparent net-neutrality infractions are the same as those for any other potential violation of FCC rules. A number of specific enforcement actions have been attempted or taken, most notably the USD100 million fine imposed on AT&T Mobility for throttling the traffic of customers who had subscribed to “unlimited” plans (see Section 3.2.3). 3.6 Reporting by NRAs In this final subsection we explain the regime under which NRAs in benchmark countries report/publish on net-neutrality monitoring, including the items that are reported and the frequency of reporting. 3.6.1 Chile There are no specific reports on net neutrality. ISPs provide QoS and traffic-management-related information on a quarterly basis. Every six months, SUBTEL prepares a report on the state of the telecoms market. This is a general state-of- the-market report which includes some QoS data, but no specific reporting on net neutrality. 3.6.2 India There is currently no net-neutral-specific reporting from TRAI. QoS information is available via the Analytics Portal, but no specific report is issued. 3.6.3 USA The only regularly published report that relates to network neutrality is the annual report from the (fixed network) Fixed Measuring Broadband America programme. The report helps consumers to determine whether they are receiving the QoS they were promised. 13 ISPs participate in the voluntary data collection; our interviewees estimated that this covers roughly 90% of all fixed broadband subscribers in the USA. The Measuring Broadband America programme report is discussed in more detail in Section 4.4. There are annual reports that summarise competitive conditions for the fixed, wired network and for mobile networks (produced by WCB and the WTB, respectively). These reports contain some information that might have possible relevance to net neutrality, but they are not specifically designed to address net-neutrality issues. Ref: 2009152-254                                           .

Study on net-neutrality regulation | 26 4 Case studies This section contains a set of case studies providing more detail on processes or tools for monitoring and enforcing net neutrality in the benchmark countries. The case studies are: •    Chile: Adkintun (including Adkintun Mobile) •    Chile: Sistema de Transferencia de Información (STI) •    USA: Netalyzr •    USA: Measuring Broadband America programme. 4.1 Chile: Adkintun SUBTEL commissioned the University of Chile to develop a QoS monitoring platform for fixed and mobile Internet connections in 2011. The programme was designed to automatically and passively measure traffic passing between a user device and the network, to create a detailed picture of network characteristics.62 A pilot programme monitoring fixed Internet connections was launched in Santiago to evaluate the effectiveness of the system, with plans to place around 1000 devices around the country following the pilot. 63 A mobile app was also developed, to monitor and evaluate the QoS delivered by mobile ISPs. Both the fixed and mobile monitoring platforms passively monitored traffic passing between the Internet and user devices, without generating any new traffic. The resulting data sets were then sent to the university for analysis. Participation was voluntary for both the fixed and mobile monitoring platforms. The fixed platform monitored traffic using either a modified router placed on the customer premises to check connection speeds,63 or an application installed directly on a computer. 64 The fixed monitoring project was closed in 2013, once the objectives of the project had been achieved (see below). The mobile platform monitors traffic via an app installed on a user’s phone. The user receives daily reports on traffic consumption from a dashboard in the app, and the data is transmitted to the NIC Chile Research Labs when the phone is connected to a Wi-Fi network. The app is designed for phones running the Android operating system, and as of May 2017 it had been downloaded between 1000 and 5000 times. 65 62    See https://issuu.com/niclabs17/docs/memoria_niclabs_2016 63    See https://www.fayerwayer.com/2012/04/chile-adkintun-el-proyecto-para-medir-la-calidad-de-la-banda-ancha-en- el-pais/ 64    See http://download.cnet.com/Adkintun-Client/3000-2085_4-75765749.html 65    See https://play.google.com/store/apps/details?id=cl.niclabs.adkintunmobile Ref: 2009152-254                                                    .

Study on net-neutrality regulation | 27 The Adkintun project was developed to help define the scope and process for ISPs to report the information required by the net-neutrality law. In addition to defining the scope and process for ISP reporting, the two initiatives provided knowledge transfer to SUBTEL, to help it build capacity for tackling net-neutrality issues. 4.2 Chile: Sistema de Transferencia de Información (STI) SUBTEL provides the STI, a secure online portal for ISPs to submit QoS information to the 66 regulator. ISPs are required to provide information on commercial and technical aspects of their services, along with QoS data (see Figure 3 below). Data must be provided either monthly or quarterly, in plain text format. 67 SUBTEL requires data submitted via the STI to be collected according to a protocol established as part of the net-neutrality regulation. We assume that these requirements follow the ITU’s recommendations regarding the collection of quarterly QoS information for reporting. Reported statistics are split by technology (and by segment, for fixed connections). Local traffic is defined as traffic within the ISP’s network, national traffic remains within Chile, while international traffic is transmitted via submarine cables or other international transit agreements. Local and national traffic is expected to be faster than international traffic.              68,69 Figure 3: Summary of QoS statistics reported by operators [Source: Operator websites, 2017] Fixed                                 Mobile Technology split                                     Copper / fibre-optic                       3G / 4G Segment split                                 Local / national / international                     n/a Statistics reported                                               • Upload/download speed • Average speed (and deviation) • Speed in 5th and 95th percentiles • Rate of failed transmission • Average delay (and deviation) The data provided to SUBTEL via the STI portal contributes to SUBTEL’s twice-yearly reporting on the state of Chile’s telecoms market. SUBTEL uses a tool called QlikView for data visualisation. It is developing further tools to analyse the large volumes of data, and hopes to launch these in September.67 Although the portal is currently only used for QoS-related data, SUBTEL believes it could be developed to consider net-neutrality-specific data. 66     See http://sti.subtel.cl:8080/sti/jsp/login.jsp 67     Source: interview with SUBTEL. 68     See http://www.movistar.cl/web/movistar/neutralidad-en-la-red 69     See http://personas.entel.cl/PortalPersonas/appmanager/entelpcs/personas?_nfpb=true&_pageLabel=P2760099104132 4329037346 Ref: 2009152-254                                                         .