IGF 2018 - Day 1 - Salle IV - DC Net Neutrality: Measuring Discriminatory Practices (DCNN)

The following are the outputs of the real-time captioning taken during the Thirteenth Annual Meeting of the Internet Governance Forum (IGF) in Paris, France, from 12 to 14 November 2018. Although it is largely accurate, in some cases it may be incomplete or inaccurate due to inaudible passages or transcription errors. It is posted as an aid to understanding the proceedings at the event, but should not be treated as an authoritative record. 



>> MODERATOR: This works, this it really works!  So we have some good news, which is that no one knows when we're going to finish.  Did you see the program?  So if we get really animated, we can carry on beyond.  If we get not animated at all, we can be done by 1:00.  Let's introduce the panel, my name is Chris Marsden.  I'm co‑chair, from University of Sussex. Oh, I have a prop!  Luca, you launched a book, right? 

     >> LUCA BELLI: Yes.

     >> CHRIS MARSDEN: I have a prop.  For those who don't know about it, I will hold this up since we're on webcast.  This is me.  A book called "Net Neutrality."  You are applauding, but it came out last January, 2017.  Thank you.  I assume everyone has read it already.  It is available online.  You can buy it if you want a physical copy.  If you think the web will crash as it has today at the IGF.  Online is gratis.  For paper, we have to pay for the frees.  Manchester University Press.  My co‑Chair, Luca Belli the dynamic behind the Dynamic Coalition on Net Neutrality.

Luca, should we get everyone to say their name down the table, and then we'll begin. 

     >> LUCA BELLI: Good morning, everyone.  Sorry if I was running a little late.  If we can have a quick round, first, panelists introduce themselves quickly.  I apologize for the reduced time slot due to the fact that the MAG decided that the Dynamic Coalition only deserve 60 Minutes.  I encourage you to complain, write in to [email protected] to ask to have back our 90 minutes.  Tomorrow, there is the Dynamic Coalition main session, you can also go there and complain physically.  That is better.  Without further ado, please, let's introduce ourself quickly so we can begin the debate.  Please, Mark. 

     >> MARC LEBOURGES: I'm Marc Lebourges.  I'm with Orange.  I am in charge with European and economical issues and I represent the African part of Orange activities.

     >> JORDANA VIOTTO: Good afternoon, I'm Jordana Viotto from Universite Paris‑Dauphine, I'm a researcher there with my coauthor who is here, Manuel Lorenzo.  We're developing a study on net neutrality and departures from that.  This is what I will talk a little about today.

     >> SERGE ABITEBOUL: I am Serge Abiteboul.  I am an computer science researcher on the board of ARCEP, the French regulation authority for telecommunication.  And I will speak about measuring net neutrality.

     >> MODERATOR: As this is the main topic of this session, Serge will introduce  with keynote remarks what we will speak about.

     >> FRODE SORENSEN: Hello, everybody, my name is Frode Sorensen.  I'm representing the Norwegian Communications Authority, and I'm also one of the cochairs of the net neutrality working group.  Thanks. 

     >> LUCA BELLI: I'm Luca Belli, I'm a professor of (?) on regulation with Fundacao Getulio Vargas,  which is both a think tank and major academic association in Brazil.

     >> I'm Chris Marsden.  (Audio skipping) As Facebook tells, you can never have too much advertising.  I'm a professor of law at University of Sussex and the founder director of the Sussex for information governance research, which is close but no cigar.

     >> AMBA KAK: Hi.  My name is Aba Kak.  I'm a public policy advisor with Mozilla Corporation, most famously known for our browser but I actually work on developing positions on laws and regulations globally.  When it comes to net neutrality and zero rating in particular, I have been introduced and worked on this issue as mostly a researcher, first with the Oxford Internet Institute and later as the Mozilla technology policy fellowship.  And I will talk about my research findings in the session.

     >> SIMONE BASSO: Hello.  My name is Simone Basso, I'm a researcher and developer with Tor Project.  I will talk about the research we have done in the past to measure performance and how this is related to network neutrality, with the UNI project which is the project that relates and studies Internet interference on the flow of communication.

     >> LUA FERGUS:  Hello. My name is Lua Fergus.  I am a research assistant at ‑‑ actually ‑‑

     >> The microphone is really strong.

     >> LUA FERGUS: ‑‑ law school.  I am the moderator online for this session. 

     >> LUCA BELLI: Ok.  Without further ado, I will ask Serge to introduce our session. 

     While the PowerPoint is prepared, I just wanted to reiterate how interesting and how productive it has been to work with this coalition over the past six years, not only producing a model framework that has been directly used by the Council of Europe to collaborate the recommendations on net neutrality and utilized by more than 50 organizations around the world for net neutrality.  The beta version was presented last year and I will present a refined version, it is included in the State of Internet report, which is the first example of actually an IGF project being included in a national regulator's report.  Which is very, very good.  And then without stealing any further time from Serge, please, go ahead.  Thank you very much.

     >> SERGE ABITEBOUL: Thank you. 

So I'm going to talk about measurements.  The first difficulty we meet when talking about measurement, it is difficult to measure the performance of Internet.  For a lot of reasons, including topology, we don't know very well.  It is depending on how you access the network and a lot of other reasons.  This is the first thing that makes it difficult to actually measure net neutrality.  Because the underlying network is already hard to measure. 

Actually, I can give you an example for that.  In France, a while ago, there was this idea that Netflix was struggling.  It was throttled on free.  Then, you know, people looked.  The reason was not an actual throttling, it was a question of interconnection.  And also all the conflicting demands for actually doing ‑‑ violating authority for sometimes good reasons.

I assume that we all want to enforce net neutrality, and the question is how do you verify it.  The first good way to approach is using crowdsourcing.  Crowdsourcing allows good input from users.  Users are brought to this discrimination and can feel when they're discriminated.  They can provide input.  That can be used (?) are for instance.

We can get valuable input from third parties, stakeholders, academics.  This is a complex problem where everybody has to contribute.  ARCEP the regulation authority for communication in France has developed and worked on this area for a while. 

I will mention briefly some of the work.  First, signalling platform to monitor relevant practices and assess the impact.  Q is the measurement tool and (?) coconstruction approach to develop best practice of measurement.  One of the issues with measurement is depending on the institution or depending on who you are, you measure different things.  When the aggregate the measures, it is difficult.

People measure different things.  Sometimes the measures themselves are biases.  The idea here is to put together all the people involved in the activity and try together to come up with best practices and really understand what it means to measure the network. 

Last an IPI to put in the ISP box to characterize the end user environment, because the environment from which you take the measure is an important stronger impact on what you measure. 

I want to spend time on an app that is demonstrated here at the ARCEP booth at 2pm in the entrance hall.  It is called Wehe.  It is an app to download and perform measurements on the telephone. 

Essentially, it is detecting how whether a particular ISP is throttling or not a particular application that you like.

The idea is simple, though the technical details are complex, the idea is to compare with an instance of a video and when you ask it in an encrypted manner.  When you ask in the encrypted manner, the ISP cannot know that it has to slow it down.  So if you see differences, you can see it was throttling.  This was used by Wehe to determine that Sprint is throttling Skype. 

This is a partnership with Northeastern University.  I encourage you to look at the demo in the booth.

I get to a conclusion.  And when you speak of net neutrality, there are difficulties in measuring it and in defining the net neutrality that you want.  It is not easy, but regularly there are new changes, new practices like zero rating or new rating,  like 5G, that basically reask the question or if you want is the occasion for some to question the net neutrality.  If you believe net neutrality is important for society, there is no choice.  So you have to understand more precisely what net neutrality is so you can articulation regulation. 

Crowdsourcing is a good way of getting information, understanding what is going on, and also research development and measurement is important because of technical issues to solve. 

I would like to conclude with a sentence that actually is enlarging, a bit, the problem.  If the network is neutral, that is very good.  That is a good starting point for better freedom of choice for users and for more fair business competition.  That is not the end of the story.  You have also to make sure the entire ecosystem is neutral.  And in particular, ARCEP has put some work ‑‑ has done enough work on the platform ‑‑ not platform, sorry ‑‑ on terminal devices, neutrality.  That is important.  Again, the network can be very neutral if the server that you help you get to the Internet is hiding it from you, neutrality is not complete.  Thank you.

     >> MODERATOR: Thank you very much, Serge and also explaining how important it is to crowdsource, the information and also the application of the regulation.  I think ARCEP is leading in this together also with other regulators in Europe.  We will see that further.  I would like to commend the work led by Sebastian, the president of ARCEP, sitting here.  And keeping with the regulatory debate, I would like to ask Frode Sorensen, also a net neutrality hero for the past years in Europe to explain what is the work in this moment that has been undertaken by Nkom. 

     >> FRODE SORENSEN: Thank you very much for this introduction.  I'm really looking forward to discuss with you today I will first of all give a brief intervention.  It is divided into three parts.  Title is net neutrality measurement in the European context.  Not because European context is the best, but because that is the context I know.

My presentation will be divided into three parts, first, I will paint the background of the discussion.  Secondly, I will present a European measurement tool that is being developed.  Finally, we'll discuss some of the potentials we have with such a tool and measurement platform. 

The background of my discussion is that as many of you know already, European net neutrality regulation was adopted in 2015.  This regulation gave direct the mandate to develop guidelines for the implementation of the regulation.  These were published in 2016 and set out how regulators should apply the regulation.  Article 5 obliges regulators to closely monitor and ensure compliance with the regulation. 

As a regulatory control to accomplish this, the net neutrality working group, which I Chair, developed a measurement methodology on one hand and on the other hand, we specified a measurement tool.  Both of the documents were published in 2017.

Furthermore, during the spring this year, there was the opening of the net neutrality development tool.  In August a contract was awarded and in September the development of the tool started.  The measurement tool is planned to be available after a year's development whereby regulators could start to deploy September of 2019.

Secondly, turning to the characteristics of the European measurement tool, the goal of the tool is to foster consistent approach for measurement of Internet services.  This will cover two different aspects.  First of all the traditional speed measurement and cultural parameters.  And it will cover, to some extent, traffic management practices that can be assessed against net neutrality rules.

The main characteristics of this tool are the following.  First of all, it is based on open source software.  Whereby the software of this tool can be reused by the different regulators but also by other parties without any additional cost.  Secondly, this tool is based on standardized architecture, it is developed by ITF, the organization developing and specifying the IP technology. 

The tool is based on crowdsourced measurements, whereby users across Europe contribute to measurement results.  We had the possibility to achieve measurements from a huge amount of different users of the tool.  Tool will be deployed nationally and autonomous and regulators maintain full control of the tool and the data collected by the tool.  This is particularly interesting and important regarding the privacy and integrity of the tool, which is highlighted in the European regulation and some national regulations.  That is why it is important that each instance of the tool has control and remains full control of the data collected by the tool.  The tool will cover metrics like speed, latency, and packet loss.

The achievement of this tool is that data is available across different regulations and nations.  So this provides a database of old measurement results throughout Europe which can be used to produce statistics.  If the tool is used by others, the same would apply to that. 

Finally, the tool will cover traffic measurement metrics.   Examples that were implemented in the first version were core blocking, DNS manipulation, some application specific measurements like web browsing and video streaming. 

Finally a brief discussion of what are the potential of the European tools?  National rollout of the tool is foreseen.  The regulators will deploy the tool nationally.  The tool could contribute to a federated system.  It will be possible to measure from one country to another one across borders.  Measurement results would reflect the end user's experience.  And in addition, it would also be possible to cover measurements across Europe whereby you can have statistical data base for items of traffic over the continent.

It is also foreseen the measurement tool could contribute to work within the quality of service measurement.  And this measurement platform since used by many different regulators would provide a platform for further development of methodology or researching of net neutrality supervision of Europe, but to the extent that others are contributing, that could be sourced into this tool. 

So possibilities are relatively great for this tool.  Of course, it depends on the acceptance of the tool in the market, but the support of this organization like (?) behind it, we believe it could be broad, at least.

     >> MODERATOR: Thank you, Frode, and than you for highlighting the tool you are developing, which is open source, which could be useful as an example for other regulators around the world.

Before opening for a brief break for a couple of questions, I would ask to Simone Basso to complete the first measurement sections with his speaking about his work at Tor Project.  After Simone's presentation we will take only two, maximum three ‑‑ we'll see three quick questions.  Then we will get into the second segment of the session with Z rating, policy, and research on different issues, and then have an open debate at the end.  Simone, when you want?

     >> SIMONE BASSO: We're trying to figure out how to go into presentation mode.  I have never used Adobe.  I have not also ‑‑ I do not use Windows.  That is it, full‑screen mode.

     >> MODERATOR: I think it is working now.

     >> SIMONE BASSO: I apologize for that, I should practice at home with Adobe to be more ‑‑ okay. 

I'm Simone from the Tor Project.  The main reason I was invited here is because we did this campaign with Fight For the Future one year ago in which we promoted our OONI app, a measurement application to install in ‑‑ from mobile stores to try to measure say the parameters of video live streaming.

I'm here, more than talking about these in specific, I would like to share with you my view of how we were doing measurements in the past, what we're doing now, and what are we doing more broadly and also discussing stuff mentioned before and giving the perspective of frustration that might be had.

2008 ‑‑ this is not the situation as the Internet was working like this.  But many researchers were using this mental model and using this kind of measurement.  What do we have here, briefly?  Couple of users, couple of laptops.  They're connected to the Wi‑Fi, and it has to be taken into account because there are issues when you are measuring parameters.

This situation was basically the Internet was for sure much more tiered.  There was the transit tier, and there were these tiers.  So come a little here on the flows we can see.  In 2008, many persons were doing peer‑to‑peer.  That was the main net neutrality issue.  You may recall Comcast started throttling Torrent at that time.

That is the red flow.  The reason it is bad to ISP because it was using the transit. It cost them.  In fact the green flow is what was ideal for many ISP ‑‑ this is a bit simplifies, I apologize.  But they said okay, I will traffic like user wants to caches that are inside me, and people will access content from there.  The reality is many times the caches were working and other cases they were accessing content that was not that far from them.  That was a thing at the time.  This will be the yellow flow.  That was not very far, but still, that was the place from which people was fetching content.  When Facebook, in 2008, Facebook was using (?) and fetching content like that.

What did people running measurements do?  That is the blue flow.  We tried to put servers in places as near as possible to content.  This is exactly the way Measurement Lab works.  This is a platform for running experiments of this time.  What people tried to do, in a crowdsourced way, access content.  See what happened.  That could be anything.  Attached to a web page, big resource to be downloaded from the measurement server or could be something that  looks like Torrent.  Then the idea was to see if there are rules inside ISP number two, like in this case, that might be throttling Torrent.

Fast forward 10 years.  What changed?  There is of course people that run peer‑to‑peer applications.  The relevance of videos, CDNs, increased dramatically.  Also the architecture, topology of Internet changed.  Now the shift towards peering and private peering is huge. 

Callers are good.  Great.  What we see now is the following situation.  The violet flow is you streaming from NetFlix.

You don't go from the transit ISP, you go straight into Netflix.  This is how it works in many cases.  Do you remember when the ARCEP colleague talked about interconnection issues? 

That is problems in the router over there.  Problems because there is nothing of capacity.  There that is where you are slowed down.  This is the thing that happens.  This not only happens because you know we need to upgrade the port, create more capacity.  Sometimes there are also business reason because you have the ISP number or not to want to increase the capacity.  Rather than taking the violet route, you take the orange route.  Orange route in many cases is worse because it is far away, you have other traffic and the quality is lower. 

This violet Orange split is called in the past, the tussles or issues between content providers and ISPs.  This is something we'd like to measure, however it is complex.  The infrastructure we're using for any measurements is not run in the Cloud.  We don't have a server sitting side‑by‑side with NetFlix or YouTube.  But what we measure at the moment is the blue thing. 

There are cases with there are no performance issues in the ISP, but you can see issues in the interconnection.  But we don't see them because we are taking another route.  There are other approaches like the Wehe one that was mentioned before in which they replay traffic and the replay is not going towards the violet line.  It is going to the measurement server to the blue part.  There are rules in the ISP that see for example, the YouTube naming the HTPS exchange and throttle.  Say this is the current situation, some approaches work like the Wehe one is good, but still there is a big trunk that doesn't measure well.  We can't measure what we want to.  Measure real content providers.  And the main issue at the moment is this one.  This is the context of the project they work on.  I think I am done.

     >> MODERATOR: Thank you very much, Simone.  Measurements it has a lot of benefits.  It is promising, but it has challenges to overcome.  Before I pass you to the second section, I would like to open the floor for three questions.  Try to make brief questions, not long comments and as precise as possible.

     >> CHRIS MARSDEN: While we wait for the first question, we will do the audience participation moment.  We have online 100 countries, we have in the room 50 countries.  We would like people to tweet the measurement tool they use themselves in their country.  We have heard of Wehe and OONI.  And heard mention of a previous measurement tool like Sound nodes before.  So please use the hashtag DCNN, as in Dynamic Coalition on Network Neutrality.  Tell us what measurement tool you're using.  Hopefully in half an hour we'll gather a decent collection of  we're crowdsourcing the actual use of the tools and the tools themselves.

     >> MODERATOR: Go ahead.

     >> AUDIENCE: My name is (?) from University Japan.  I have been working for the net neutrality for the last couple years and currently the part of the community (?) right now.  I have to say the I have the phrase, that the entire ecosystem should be neutral.  I have the phrase.  And also net neutrality is important for our society.  It is not for operators or content creators, but for the people.  I feel that QS is not an appropriate measurement for net neutrality. 

I think the experience, the QE, is more important.  So because not guaranteeing experience, It is sometimes beneficial for the end user to experience. For example, during the condition period, prioritize (?) videos on YouTube may be beneficial for end user to experience because the video needs more real‑time transmission of the packets.  Compared to the SNS on Facebook.  So my question is:  How do you incorporate this possible contradiction between the Qs and QAs into your measurement system or do you think it is necessary?  Thank you very much.

     >> MODERATOR: As it is a quite long question, I will take one more question, and let the panelist reply.  Is there anyone that wants to add something or clarify the question? Okay.  So we can leave ‑‑

     >> DCNN and the measurement tool you're installing right now on your mobile device. 

     >> MODERATOR: Please you are head panelist in the next session.  Be my guest if you want to reply to the question.

     >> Brief comment from me.  I understand as it being requirements related to the application and not regarding measurement.  Under the European context, there are different answers on how could service be provided for an application? 

First of all, I emphasize Internet technology is dynamic and developing all the time.  Currently video streaming is provided with quality over the Internet with CDN service.  It seems like sufficient service is provided for the video on the Internet service even though it is the best service ‑‑ secondly under the European regulation, specific streaming services like IPTV could also be provided as a service that is not going over the Internet.  That is something we refer to often as specialized services in Europe.  So that could be provided for IPTV or for example, video conferences where it is not possible to use a CDN.  There are two answers.  The technology is developing on the Internet and provided streaming with good quality.  Second, provide communications separate from the Internet and thereby secure the quality without compromising quality on the internet.

     >> CHRIS MARSDEN: Frode, do you want to say something ‑‑ I believe that Barack has recently stated 5G does not have an automatic exemption from net neutrality.  Maybe we should direct people to the fact there has been a statement by the Chair of Barack, right?

     >> FRODE SORENSEN: Yes, there has been a statement from the Chair of Barack in some conferences explaining we don't see a problem to implement the 5G technologies, but at the same time ensuring net neutrality on the Internet.  Barack will publish an opinion to the European Commission December this year.  And we'll have a more substantial explanation to that in that paper.  Generally speaking, the way we see it in the Norwegian regulator but also operator, in Europe, 5G could be implementing with the current net neutrality and it should be compatible with each one another.  That is the view of the European regulators.

     >> MODERATOR: Thank you.  That is helpful.  I'm sure everybody in many of the countries we're covering, I'm sure many are thinking of 5G and net neutrality and specialized services.

Let's move on to our further speaker (?).  Amba, you will have the floor.

     >> AMBA KAK: I don't know if I can have the title of being part of the open source community, I'm more a lawyer than part of that.  I work for Mozilla, and that is the value we stand for.

I want to talk, we will shift gears.  We heard why technical discrimination is a hard problem.  We will shift gears a little during this session. We talked little about how zero rating and forced discrimination is a hard problem.  To step back, whenever you make a regulatory choice to restrict or ban a particular technology or business practice, there is always what is counterfactual.  What is it that is not going to happen because you decided to ban something.

In the case of zero rating, the pro consumer zero rating what is the control that you will not see because of the fact you banned it rather than evaluating it later?  You will see on this map, in blue, the regulator did decided to ban any phone that is zero rating. 

In the memorandum, they explained (audio skipping) one of the many reasons it was decided the way they did, it would probably be hard or difficult to identify in an exposed analysis.  It would be difficult to show the impact on innovation exposed, for that and other reasons, the Indian regulators decided the way they did.

What I was thinking about last year, in the case of zero rating, we can actually ask that question of what about counterfactuals.  What I did, I studied regulatory decisions in Europe, in Canada and in the US before the repeal of net neutrality to discuss when regulators evaluated the zero rating cases how did they pick good from bad zero rating?  What were the factors that were important, what factors were missing in the assessment?  And what did it tell us about regulatory strategies going forward? 

For those interested, I will be happy to show the results later.  I will start with the broad trends I saw across regulators. I give the caveat, in the U.S., they want regulatory decisions as much as review of the Wireless Telecommunication Bureau.  They did put forth conditions they thought were concerning and those that were not concerning in zero rating. 

First, whenever it was zero rating content that was through the ISP or content provider that was a red flag across jurisdictions. 

Secondly, in Europe, when a particular service discriminated within a class, that was frowned upon.  For example, in Canada they found that picking a class of service for example, online streaming to be zero rated and not online radio was viewed as discriminatory. 

The broader question or perhaps the prior question, is it zero rating available for open content providers to apply, too?  Can I,  as a content provider that is not zero rated, apply to the ISP to be part of a content provider that is not zero rated, what is the terms of that decision?  What is the fee to be zero rated?  And finally, this cropped up a lot, particularly in Europe, Is the throttling of the nonzero rated content, particularly after data caps have been reached? 

These are factors that intuitively make sense, but broadly, my final conclusion related to the regulator decisions was there seemed to be factors that were missing, completely missing or were not being analyzed with any degree of rigor.  That meant the true impact of allowing zero rating or plans on innovation, the net neutrality was really not have been evaluated the way it should have been.  And the four factors, I will end with that.

The first is data caps.  It is easy to understand if your data caps are low in general, there will be an incentive to use zero rated content more than others.  In Portugal, for a bundle of applications which were chosen by the ISP, the data caps were five higher than they were for normal content.  If there is a distortion, we can argue it is amplified when data caps are low.

There is another question, this is what folks at (?) are putting forth evidence for.  How is zero rating in general affecting the data caps in general on the Internet?  Are they falling, rising?  From there we found it seems the practices are making it more expensive.  General data caps are not expending.  It is the opposite trend. That is concerning.

The second, we heard a lot that ok the zero rating plan is open for you and you to apply.  But where is the information?  How easy is it for me as a content service supplier to apply to be zero rated?  Going to do a plug, since that seems to be the flavor of the panel, I will give a plug.  I finished a report with Daniel O'Maley,  where we looked at how news organizations applied to be part ‑‑ did they know there was Facebook in their country?  Had they applied?  We heard two interesting things.

First, we heard many of them had no clue they could apply.  It was buried in some part of  many facts in the web and didn't know it.  Which is an important factor.

The second one, many of them  felt the technical specifications were too complex and those not adapting to the digital age felt they wouldn't be in a position to apply. 

     I would ‑‑ actually Nkom is a regulator,  and in of one of his decisions noted that the process to be zero rated should be simple, should be open, transparent and clear.  That is very important. There is one conclusion drawn is there isn't enough actual delegation of how easy is it for applications to apply.

That is the second order, third problem, can you negotiate the terms of zero rating.  For example, if the terms include deep packet inspection, or implications with privacy harms, I'm as content provider in the position to resist that?  We don't have enough information. 

And that relates to another point that Epicenter and Agee pointed out in their report, what about new technologies like blockchain or others that might be left out?  That is a broader question of is it open, but who is it open to and how easy is it to be zero rated? 

I will end with, the point of this research is not how to make the process of zero rating more fair or equal, but to ask the broader question which is the rules we choose today will shape the direction of innovation tomorrow.  What are the data points, the evidence, the principles we should think about when we decide regulatory strategies?       I think in Europe and elsewhere, that is a moment to think the question afresh.  I hope we do that over the coming year.

         >> CHRIS MARSDEN: Is ‑‑

>> LUCA BELLI:  I just wanted to specify, for those that don't know it, maybe they've not been present to in the past or this session last year or maybe the mailing list.  This is the zero rating map we are developing.  I will get into the details of the map in a couple of minutes.

     >> CHRIS MARSDEN:  I'm conscious, I'm staring at the U.K. not entered on the map.  I haven't turned that into input.  Luca, my apologies.  I can see Australia there as well.  Where is Angela Daley?  She can be filling in that section for you.  This is something we can complete very soon.  We should pass to Jordana Viotto, you have the floor now.  Thank you for joining.

     >> JORDANA VIOTTO: Thank you.  Good afternoon. Thank you, Luca, for inviting Manuel Lorenzo and I for participating on the debate.  We have been developing a research project on that and I will talk about why and how it connects to the debate today.  We live in a world that is more and more digital.  Traditional businesses are using online tools to connect to customers.  Who benefit from a reduction in transaction costs, verification costs, search costs in order to find and confine and prove products and services that better match their preference and needs.

For the more interesting increasing number of startups are digital natives.  When we talk digital natives, we think about uber, lyft, Facebook and Twitter, and there is a myriad of new digital native firms emerging every day.

As a result, the Internet service providers are growing in their position of the economy as they intermediate transactions between businesses and consumers.  It is important to understand the existing data regimes ones adopted. Net neutrality, rate prioritization, so on, so forth. 

One of the main concerns regarding the potential departure from net neutrality which was very highlighted today, I guess.  And what moved us to investigate this matter, is that might steer the consumer choice of the online content to the content provider, that is included in the zero rate plan or paid prioritization plan. 

Content providers rely on traffic to generate revenues, on advertisement, from subscription, the zero rate and other nonneutral practices could be given barriers to providers that are excluded from the plan.

And might reduce a variety of content providers available to users.  And interfere with incentives for content providers to invest in quality.

These are the concerns we started from.  In the paper, we aim at understanding whether the departure from net neutrality could affect the incentives of content providers to invest in quality.

The results are preliminary and we are focusing on the net neutrality and not prioritization.  But they show, to our surprise, the incentive varied by demand for the zero rated content provider.  In the preliminary result when demand for the zero rated content provider is inelastic, the nonzero rated providers would degrade the quality in order to minimize cost and in order to battle the zero rated content.  However when the demand for zero rated is elastic,  the other content providers would have incentives to invest in quality due to the attractiveness of the zero rated content. 

Of course, these are still preliminary results, and there are many questions not treated in the question.  For example, the levels of data caps. 

Results support the heterogeneity of opinions in the debate right now.  We also realize the defects will be different for distinct types of content, and the results might be taken into account by policymakers. 

And we saw today one of the things we wanted to do with our paper is investigate empirically how it happens.

We have difficulty in finding data, but today, we saw some tools that might help us shed a light and get more insight from data from the real world to see how results resonate in this context.  Hopefully we can offer policymakers and content providers with more insights in the near future.  Thank you very much.

     >> MODERATOR: Thank you very much. Yes, we want more evidence.  We as reseachers (?) I would love to see in the black box, AICAMI, to see what they're streaming.  That would be fascinating to see specialized services for the future.  We have heard from academics, regulators, lawyers, Civil Society, and the free software community.  We haven't heard from someone that makes it work in terms of pipes.  We're delighted it is from Marc Lebourges.  The floor is yours.

     >> MARC LEBOURGES: Thank you very much.  Thank you for inviting our engine.  This is one of the most important principles of Internet governance. It has been translated into several legislations.

Telecos have strong incentive to provide quality Internet access.  The value of most content available unintended Internet, and the value of the access itself depends on the availability of those contents available together online.  Think of search engines.  Only neutral networks allow such change.  In comparison, the specific interfaces provided on parallels of cannot exchange online from other applications.  It is depends on proprietary content and cannot be substitute to most content on that available on the Internet.  Thus for end users, specialized services are complement and not substitute for Internet access, which means their value increases when the Internet value increases, too.

European regulation requires the check of if it is degrading the general Internet quality.  The appropriate test in this respect should be whether, for given set of  usages, if the quality was degraded or improved if usage served by a specialized service was instead provided on Internet.  That specialize service used optimum technical solution to minimize research for given outcome.  Not using such an optimization can be expected to reduce the efficiency, at the expense of the general open Internet quality.

For instance, if instead of providing TV through specialized service the operator provides TV using the (?) solutions.  The operator  service would occupy much more capacity at the expense or quality of other services available on the Internet.  So providing specialized services when appropriate improves the general efficiency of the network and the guarantee quality of Internet service.  Zero rate is an important theme of this session.  Operator's motivation to provide zero rates are basic.  Proposing goodies for consumers for commercial purposes. 

The recent Hecht report on net neutrality in Europe mentions no prohibition of that right, as such.  This appears as right interpretation of European regulation.  In‑depth analysis proves that zero rate entails no material restriction of the end user choice, except in very, very specific circumstances.

On the contrary, zero rate generally increases the usage boost of zero rated and nonzero rated services because the treatises they're reporting for end users.

By increasing usage of all services, it increases the provision of all services in the long run.  Operators my propose to content providers to sponsor data, data traffic.  This is modern version of free phone numbers 800, phone service for which providers pay for traffic which is free for end users .  This is efficient when providers have reasons they want to promote usage of the online services.  Perhaps they are part of their customer care package.  Free phone numbers are not in contradiction with nondiscriminatory ideation for classical telephony.  The same should apply for sponsored in the Internet world. 

There are around 4 billion Internet access in the world.  All of them are the result of the technical, financial, and commercial initiatives of telecom operators.  There are more than 2 hundred million customers and more than half of them in Africa.  African operators start switching from voice base to database business model as European operators did.

This requires a dialogue between authorities and operators to maintain the system ability and achieve high penetration of (?) Internet in Africa.  Thank you for your attention.

     >> MODERATOR: Thank you very much.  Okay.  We have time ‑‑ I say we have time for questions ‑‑

     >> (?) (Off mic)

     >> MODERATOR: I'm reminded that we have one more speaker.  I turn the floor to Luca.

     >> LUCA BELLI: It is just as I promised, a very slight explanation of the zero rating map.  This is the main outcome for this year.  We started it last year with the beta version at the IGF 2017.  At the time, only featured 22 countries.  Now 100.  And added further criteria and differentiated also the colors to identify in which countries ‑‑ the main goal is to identify where zero rating and net neutrality are available.  You see blue the countries where net neutrality regulation is existing, zero rating is not available.  And there is a harder criteria that is emerged as key in the debate that was previously completely not considered.  Data protection.

Why is that key?  A merger from the first version, for example, in Brazil, the case, then emerged as tendency in the developing country, operators or those who sponsor this zero rated application may do this in order to collect data.

I just tweeted an article I penned last year called "Scramble for Data."  The first time I didn't see this emerge as very clear from this project.  It was the players that we're sponsoring this application maybe were doing it because it was becoming convenient to sponsor the data that were consumed by users to get access in order to get their personal data. 

That first completion was confirmed by the fact that in virtually all countries where zero rating is possible, the Facebook group applications are available.

That (?) became a little bit more scary when we started to apply the data protection criteria.  All of the orange or red countries, no data protection at all.  And no net neutrality rating.  And zero rating  is available.  It means that people that utilize zero rating may be a people of very low income.  Usually with low education and don't have enough literacy understanding to know what is the risk of not being mindful of one's own personal data.

We have the southern part of the world, with exception of India, in which zero rating available.  Only in the yellow country, there is net neutrality and zero rating and data protection.

In the orange country you have, there is no zero rating or net neutrality present but you have, at least, data is present.  No zero rating is present, you have the safeguard of at least a basic protection framework, that is due because the data protection debate emerged much before the net neutrality debate. I would be very scared if I were living in one of those red countries and that was also democracy.

This was formally democracy given the recent distortion that we are now used to hear about how personal data from people are harvested, used to psychological profile people and then divert democratic processes.  What is also emerging now it is not any more a question of freedom of speech innovation, but a question of democracy.

Brazil, where I live has formally net neutrality and zero rate regulation and suggested we adopt a law on data protection  if we force in two years, formerly the yellow country, really recently what is the kind of distortions?  I don't want to transform into political debate.  Think about this.  If very few players in the world can collect data and very low income with no digital literacy of all people are connected to those few applications and those few applications  they regularly leak their data or share data with people not in good faith.  That is an area of real problem fork policymakers.

That is the last thing with which I wanted to close my speech.  I will let Chris moderate this last ‑‑ this 10 minutes of open debate.

     >> MODERATOR: So have a look at the blanks, think about your personal responsibilities to help produce this official outcome of The Dynamic Coalition.  Russia.  China.  Iceland, I know passed data protection and net neutrality laws in June.   Anyone from Iceland in the room?  Sweden, Greenland, the world's most largest producer of original music, per inhabitant.  Drago knows that about Greenland. 

So think about no the Dynamic Coalition can do for you but what you can do for the Dynamic Coalition when you are filling in the map.

Let's take the question here. Anyone else have a question?  As we know, we are about to go into the very VIP, very official, very formal session of the IGF this afternoon.  So we have plenty of time to have an interesting discussion in here.  Question here and then there.

     >> AUDIENCE: Question and comment, basically, from (?) we have basically the role as everybody in Europe.  We don't have zero rating.  I don't see how to map that to the information to your gathering exercise.  Not because of the net neutrality privacy regulation, it is due to the commercial reasons why the operators hasn't implemented any zero rating plans in Finland.  That would be interesting information.  It is not only the regulation, but the market impact and how the market has decided.

     >> MODERATOR: For those that don't know, I will mention why we need to be careful about zero rating.  In Finland, everybody has unlimited data on mobile.    You mention to a Finland consumer zero rating, they say what are you talking about. We have unlimited data.  We don't need... 

(Background noise)

Really good point.  We are together tomorrow.  Can I say, Luca, tomorrow 4:40, another session on measure of net neutrality.  This is the official Dynamic Coalition session, and a further session tomorrow.  Talking about Finland and also Alyssa Cooper from IGF is speaking as well.

     >> To encourage you to add this information on the zero rating map.  Go to zerorating.info, it is also on the right bottom corner, the link so than  that everyone can feed the map.  A crowdsource effort.  This kind of information ‑‑ it is a country where there is regulation, there is data protection that would be a blue country like Canada or India.  Thomas?  New haircut. 

     >> AUDIENCE: Yeah.  Hello. (?) I have two questions for our colleague from Orange.  We are entering a reformed debate in European Union about our Internet net neutrality framework.  It is vital to have a factually based debate. 

     My two questions would be: You mentioned multicast versus unicast as one distinction criteria for the specialized service.  I think this is a well known example that we heard about a lot.

What about 5G? Can you mention any concrete specialized services that you do know the industry is thinking about?  Because I think everybody in the room is wanting to hear about the concrete business models that we want the regulation to allow or prohibit.  Having those examples would be very helpful. 

Secondly, you mentioned sponsored data as one potential zero rate business model.  To come back to a factual level, do you know if Orange is implementing or going to implement sponsored data‑like schemes anywhere?  We only of AT&T in the U.S. applying such a model yet, but maybe you can enlighten me. Thank you. 

(Background audio distortion)

     >> MARC LEBOURGES: On 5G, the position of Orange, on the net neutrality regulation 2015 regulation.  It is compatible 5G is compatible with the current regulation.  They come into different detail on the ‑‑ what type of ‑‑ what type of modeling of the category regulation it can be.  It can be long segmentation, specialized service, reasonable traffic management.  All of the categories are indicated in the regulation.

     Appropriate ways to address or consider 5G within the scope of net neutrality.  So Orange is not requiring or asking for any review of the Internet regulation of 5G.  It is the question of interpretation and regulation which is in the end of the NRA.  And our view of the NRA record the actual record making of most NRA in Europe.  He would say it is a reasonable record of implementation of the net neutrality regulation.  It is a question of philosophy. 

(Background discussions distorting audio)

you start by considering it is a breach of each and you make up circle and limits.  Then, I think if the state of mind of regulation will weigh‑in, it would prevent reasonable development of 5G.  If there is a reason to think the industry initiatives in the 5G are reasonable and we have strong incentive to be fulfilling quality neutral Internet access and we are not going to have any initiative to prevent quality access.  And if this is the state of regulation, I think things will go nicely. 

For now, there are 4 million relatively neutral Internet access in the world, all of them are the initiative of Telecom operators.  So it is strange to say telecom operators are against the net neutrality market and the Internet access market that they provide themselves.  It is surprising to be considered as not supporting the Internet that we ourselves have provided.

     >> MODERATOR: Sorry to interrupt you, Marc, but I thought ‑‑ we thought we could take our time, but there is the Secretary‑General of the U.N. waiting outside for this room.  We cannot.

     >> Can we take the last question as a tweet or e‑mail?

     >> LUCA BELLI:  Yes.  There is a session on 5G net neutrality tomorrow in Room 2 at 11:15 where I think these kind of questions will be thoroughly debated.  Thank you for your comments.

     >> Thank you, Luca and Panel


     >> MODERATOR: And if you need the tweet ‑‑ if you think we need more than 60 Minutes or 90 for the discussion of the Dynamic Coalition, you can reply directly Chengetai directly on this.  Thank you, Chengetai.

     >> CHRIS MARSDEN:  Hashtag, occupy IGF.

     >> LUCA BELLI:  Exactly.  Hashtag, give our 30 minutes back.