Internet Thought

Vanished from the index of Google, Bing and Linkedin a rebuttal of an anti-net neutrality paper

noreply@blogger.com (Rudolf van der Berg) — Sat, 11 Jan 2025 20:44:00 +0000

What if something can't be found through Google Search? Does it still exist? That question I have to ask myself when I found that a blogpost I wrote two years ago can't be found anymore through either Google or Bing, though it is still listed on Brave. A LinkedIn post where I refer to the blogpost was also made inaccessible, even though Bing can still find that it exists and it can be partially read when browsing anonymously.

Given that it was delisted from 2 search engines and removed from LinkedIn, I suspect it's not just "death by algorithm". It appears someone asked for its removal from the search index and from LinkedIn. Who it was; staunch anti-net neutrality advocates, the publisher, authors or someone else, I don't know. What I do want to understand is:

How could someone file such a request?

Are there any records of such requests?

How could I as the author know of such requests? Should I have been informed?

How can I get my article added to the public record?

Debunking an anti-netneutrality study by four economists

My blogpost debunks an anti-net neutrality study by four academic economists. The anti-net neutrality study claims that net neutrality leads to 25% less investment in fiber broadband. It used to be that when someone searched for the academic paper (eg “Briglauer net neutrality OECD broadband”) that my post on Medium was the fifth or sixth result. The LinkedIn article would also pop up in the search results.

The academic article is cited by telecom lobbyists in the USA and UK as proof net neutrality is bad. However, the US FCC and Ofcom both dismissed the conclusions of the article on the effects. This shows I'm not alone in my criticism of the article. Anyone who tries to verify the claims used to be able to find my debunking of it too and make up their own mind. Now neither my Medium article nor my LinkedIn post can be found through Bing, Google or Linkedin. My Medium post still exists on the Internet, but it isn’t indexed and so it is unfindable. You would have to know that it exists and where I posted it, or otherwise you won't find it, same as the LinkedIn:

The original posted to Medium is here: Updated: Request for retraction of “Net neutrality and high-speed broadband networks: evidence from OECD countries”, by Wolfgang Briglauer, Carlo Cambini, Klaus Gugler & Volker Stocker, Rudolf van der Berg, Nov 17, 2022 https://medium.com/@rudolfvanderberg/request-for-retraction-of-netneutrality-and-high-speed-broadband-networks-evidence-from-oecd-d33ce2f8b749 a backup is now at: my old blog on blogspot

The LinkedIn article that I wrote about it, used to be at: https://www.linkedin.com/posts/rudolfvanderberg_request-for-retraction-of-netneutrality-activity-6998827449144778752-Y254/ This one is now inaccessible.

The academic paper is at: Net neutrality and high-speed broadband networks: evidence from OECD countries | European Journal of Law and Economics

Examples of how the article has become unfindable

The article on LinkedIn now only shows that the post can't be displayed. I received no notification, no request to change my tone or any other indication that the post isn't visible anymore.

When browsing anonymously/incognito, the Linkedin piece is still kind of visible:
Google will tell me to "Go Fish" if I search my blog for the author's names or unique words, such as the author's names, Ofcom or Australia https://www.google.com/search?q=Ofcom+site%3Arudolfvanderberg.medium.com

If I search for https://www.bing.com/search?q=net%20neutrality%20briglauer%20oecd%20retraction%20request I only get links to the LinkedIn article, but no reference to the post at rudolfvanderberg.medium.com. If I then click those links, the result is the same blocked page as above

https://www.google.com/search?q=net+neutrality+site%3Arudolfvanderberg.medium.com only returns the root, but not the specific article. If you search for Squid Game, it does return the specific article.
Brave does still show the article on the first page, which is how it used to appear in Google and Bing Search too! https://search.brave.com/search?q=net+neutrality+broadband+OECD+briglauer :

Conclusion

Being able to find things through search engines is essential for the public discourse on contentious subjects, such as net neutrality. If a rebuttal can be made to disappear from indices of search engines and removed from social media sites such as LinkedIn, then what does this do for our open society? I'm generally a techno-optimist, but this is a gloomy prospect.

Rebuttal of “Net neutrality and high-speed broadband networks: evidence from OECD countries”, by Briglauer, Cambini, Gugler & Stocker

noreply@blogger.com (Rudolf van der Berg) — Sat, 11 Jan 2025 19:30:00 +0000

Tuesday November 15th 2022 I sent an e-mail to the authors and editors of the European Journal of Law and Economics requesting the retraction of the article by Briglauer, W., Cambini, C., Gugler, K. et al. Net neutrality and high-speed broadband networks: evidence from OECD countries. Eur J Law Econ (2022). https://doi.org/10.1007/s10657-022-09754-5.

The New Year's surprise of 2025 was, that the article detailing my interactions with the authors on my blog hosted at Medium, had been removed from the search index of both Google and Bing and the LinkedIn post I had made about it were inaccessible. Read about that story here: https://internetthought.blogspot.com/2025/01/vanished-from-index-of-google-bing-and.html

To make sure that the article will be indexed again, I decided to post it here and see what will happen next.

Updated: Request for retraction of “Net neutrality and high-speed broadband networks: evidence from OECD countries”, by Wolfgang Briglauer, Carlo Cambini, Klaus Gugler & Volker Stocker

Update 09–01–2023: The editors have replied that they see no reason to retract the article. They have invited me to write my own article and submit it through the usual review process. The authors haven’t responded at all. Off the record I’ve had some positive and negative reactions from economics professors. Apparently it’s a biblical truth that IP-interconnection is a two-sided market, even though I explained why it isn’t. So basically nothing changes and this has been lots of effort with little effect. That said, there is a nice addendum on how interconnection pricing works at Deutsche Telekom.

Tuesday November 15th I sent an e-mail to the authors and editors requesting the retraction of the article by Briglauer, W., Cambini, C., Gugler, K. et al. Net neutrality and high-speed broadband networks: evidence from OECD countries. Eur J Law Econ (2022). https://doi.org/10.1007/s10657-022-09754-5. This was with a heavy heart, but the article contains too many methodological and factual issues and an even heavier heart that I publish the whole email and some background here on my blog. It is only a day later. I haven’t received a response to the email from the authors or editors of the European Journal of Law and Economics. I don’t expect a response anymore now <author X> blocked me on LinkedIn.

Background

For those who don’t know me. I’ve been working on telecom and internet policy for two decades now. I have a masters degree in Public Administration and Public Policy from Twente University in the Netherlands in 2001. It is in Twente that I started my carreer at the NDIX internet exchange, learning about interconnection. I worked as a policy advisor at the Dutch ministry of economic affairs. It is there where I wrote my first report on net neutrality in 2006, of which only a reference in Kocis and De Bijl 2007 remains. The section on peering and transit I later translated and was published on Ars Technica. I worked at the OECD from 2011 to 2015 doing analysis and writing reports for the Working Party on Communication Infrastructure and Services Policy. There I initiated and organised the two “BEREC expert workshop on IP-Interconnection in cooperation with OECD” in 2011 and 2012, because I realised that telecom regulators and IP-interconnection people didn’t know each other and this created misunderstandings. Those meetings had a positive effect on the position of BEREC and the EU towards the ITU WCIT conference in 2012, where ETNO (like before and still today) tried to impose the anti-competitive and destructive sending party network pays model.

Since then I’ve become a consultant and I consult among other topics on net neutrality to regulator and tech firms. I am and have been in favor of net neutrality and have always been clear of that. One of the roles I have and for which I’m sometimes thanked in academic papers is, that I can explain how the internet and interconnection works. For example when a telecom CEO says traffic growth is too much, I will show that average traffic on his network is 3Mbps and his firm has told investors that it can handle traffic growth with ease.

So whenever academic work is published on net neutrality I’m interested.

My interaction with economists on net neutrality through the years.

To be fair, my experiences with economists on the topic of telecommunications policy in general; broadband, investment, interconnection and net neutrality hasn’t always been positive. When working at the OECD it was normal to find that academics from the field of industrial economics would write articles slaughtering OECD reports and recommendations, based on their own datasets, research or a theoretical model that somehow questioned the empirical data and conclusions me and my colleagues had gathered. Surprisingly, not many who criticised would contact before publication or ask for clarification. Regularly a telecom firm was cited as having funded the research, though the statement generally also said this hadn’t affected the results. Surprisingly, the results were always more favorable to the telecom firm than the OECD’s results.

When contacting industrial economists through the years about what I perceived as mistakes, misunderstandings or possible options for improvement in their articles, I generally found them less then willing to make changes or address my concerns. Their articles almost always come with some comment on how policy makers should pay attention to their model and its results. At times there are even statements that they showed the government policy or the underlying research to be flawed. The reverse somehow doesn’t appear to be possible. Some responses I heard (if I got them):

Thank you very much for your comment, but I’ve moved on to other projects
Thank you, but the article we wrote is about a theoretical situation, so we understand that it doesn’t fit your empirical description of the telecom market. We may include that in a future article
Telecom firm X has told us how the telecom market functions and this is different from what you say is the case.
Our article reflects the state of academic discussion including those of nobel prize winners
The reviewers and my professor thought it was good.
Why don’t you write your own peer reviewed article on this.

Last year I received a similar response when I contacted author Briglauer on a paper he published then, that was sponsored by Deutsche Telekom. That article I had found through a blog post of Deutsche Telekom on repealing net neutrality, because it can’t handle 7MB of Youtube traffic per customer per day sustainably. (357Terabytes for 53M customers in Germany) The article argues against netneutrality and for sending party pays based on the imagined energy use of sending data in networks. It uses an out-dated and mistaken dataset on energy use of data networks based on the unscientific kwh/gb metric. The authors had missed that energy use has remained stable for a decade. Networks use energy primarily for achieving a certain distance with the signal, the modulation is irrelevant. I received a similar response to other academics. That the article was based on academic sources and had been peer reviewed. When I asked a Dutch professor on how I coud deal with this; the response was that sending a letter to the editor could be an option, but often editors choose not to publish the letter. I left it there.

Now I will admit that I may just be a fool and wasting the valuable time and brain cycles of industrial economists with my crackpot ideas. For some reason academics in other fields; law, sociology, technology etc have proven much more willing to interact and discuss.

My interaction with the authors

A week ago I learned of the article of the four authors through a comment of someone I’m linked to on the Linkedin profile of <author X> who promoted the publication of the article. I read the report and saw some elements that got me positively excited:

Not just a theoretic article, but empiric data
Data that had a novel way of looking at investment
Some of the authors I knew and had a positive image of.

When reading it, I saw some classic mistakes, such as a theoretic model that assumes a two-sided market and the assumption that net neutrality bans payments between content providers and ISPs for terminating interconnection, the so-called Zero Price Rule (ZPR). I asked some questions publicly on LinkedIn and then asked <author X> to connect on LinkedIn to allow a more direct chat.

I expressed my concerns and <author X> responded. I had misunderstood one element of how the econometric model was created, which he clarified for me. I expressed three concerns which he responded to:

On my comments that the ZPR wasn’t present in EU regulation and therefore the channel identified in the theoretic section of the paper that would affect investment, the response was: No, the ZPR is one — comparatively easy — way to illustrate formally the effect of imposing a constraint on profit maximizing ISPs, just as NDR would be (if binding, both reduce profitability, because if ISPs would otherwise engage in such practicies, it would increase their profits — which just follows from the profit maximizing assumption; in that sense the model is generic in simulating / describing the effect of some kind of binding NNR on ISPs profits and related investment decisons.) (NDR= Non-discrimination rule, NNR = Net Neutrality Rule) This shocked me, because apparently a theoretic model that doesn’t describe the empirical situation is not relevant for empirical conclusions.
On my comment that authors had found a correlation but failed to find a causation the response was: second we argue we do identify causal effects in the empirical analysis; the whole section “identification strategy” is about this, which has been accepted by anonymous referees in a rigourous review process. hence, if you argue that we still have identified correlations only you would have to argue this on solid econometric grounds. So because the reviewer didn’t catch that the theoretic model doesn’t apply to the empirical data, I have to show that the econometric model isn’t sound. So I did that in the e-mail below. Good thing I always loved broadband statistics.
To my argument that EU telcos, such as Deutsche Telekom and Orange were charging for interconnection and that almost all netneutrality cases in the EU had been about zero rating, which the theoretic model completely fails to address, the response was; third, yes, just because there are peering agreemes in the real world re IP transit etc does not imply that NNR (imposed on terminating part of networks) could have had no (partial) impact on investment. Same logic applies to mobile cases you refer to, in both cases you simply do not have the counterfactuals supporting your argument (e.g. NNR might have had an impact on wireline ISPs as they were obeying to the NNR … there might be dozens of reasons why we had mobile related cases only …. zero rating we dont have as such in fixed broadband, right?, etc …

I then started working on a detailed analysis of the article and also of the authors it cites as references and other work of the authors to see what the article is based on. I got increasingly worried the more I examined the article. I had first been primarily concerned about the theoretical model and that concern proved valid. As <author X> requested I then looked at the econometric work, because it had been in a rigorous review by an anonymous reviewer. What I found very disconcerting.

I did place a phone call to <author X>, who was surprised, but we spoke an hour on the phone. The call didn’t alleviate my concern, it increased them only more. The author was dismissive of my concerns and even after several requests and explanations refused to give insight into a sample of the datasets used for the article.

The problem with the result of the modelling is that the authors say it shows that countries with netneutrality (the EU) had 25% less investment in FTTH, relative to countries that don’t have netneutrality. That conclusion however hinges on Australia and New Zealand that don’t have netneutrality and didn’t in the period when the EU does from 2016 til today. Those just happen to be the years that Australia and New Zealand invested massive amounts of money in national broadband networks. So the relative difference is a lack of EU wide broadband plan, not the netneutrality regulation.

The econometric model is used in a number of articles by author Briglauer. I looked at these articles and suspect some of the criticism I give below also apply to those articles. Indeed I found it quite concerning that data on government investment in broadband, that was used for another article was not part of this article, despite it being of great relevance to the effect of net neutrality on the investment in high speed broadband networks.

What I hope to achieve

I hope to achieve better academic research. Particularly in the field of industrial economics. Not a week goes by without someone arguing based on economic research that the government should do XYZ in telecom/internet/tech policy. Every now and then I get asked to verify or give my opinion on the article. I really just want to be able to say: “oh by economist X … yeah, that’s probably rock solid!”. My current response is; did you check the assumptions? did you check who funded this article or his previous articles that say the same? do assumptions hold? any empirical data and if so what is the quality? It shouldn’t be like that.

I need to be modest too

Should anyone see mistakes by me, outright stupidities or smaller ones, please do point them out to me. I can generally handle it. I at least think I should! I’m critical of others, but generally to get science further and get better policies for better lives (the OECD-slogan is one of the best I think) I however do often make mistakes. I know my communication style is blunt and I tried to do better.

Addendum 1: An employee of a European content network send a message that the tariff charged by Deutsche Telekom for terminating traffic on its network is a blended price of transit and termination payment. So it may look 20cts per megabit per month capacity, but it really is a termination fee that is high and an additional lower tarriff for transit, which the other may not want or need.

Any network that has significant traffic for Deutsche Telekom finds that commercial transit networks towards DT are saturated (an experience that Deutsche ForschungsNetz had too). The network than requests Deutsche Telekom for peering. Deutsche Telekom Wholesale will only offer transit. The price of transit is different for different AS-numbers, based on whether they are a telco like Deutsche Telekom, CDN or hosting provider. This is odd, because networking protocols can’t see a difference and there is no cost difference. Hosting apparently pays more than CDN and telco. The network than tells Deutsche Telekom it has enough transit capacity and doesn’t need more.

Deutsche Telekom however insists, it’s either transit or nothing. The transit price for 100Mbps may be 20 euro per month. The hosting network only has 25Mbps of traffic in either direction with DT. The price for terminating traffic translates to around €0.80/Mbps/per month. 4 times more than I wrote based on WIK. The content provider can then choose to use the transit or not. If the content provider wants only 50Mbps than DT will tell it the price is 40cts/mbps so €20/mbps/month too.

However requesting to only pay for the traffic that terminates on DT, paying only the blended price of €0.20/mbps/month is met with disbelief. It still costs €20/month for 25Mbps. If the content provider doesn’t agree, the only alternative is to route through networks DT has a friendly relationship with and that have a similar price point. This is similar to the Swisscom-Init 7 case referenced below. Mind you, Swisscom lost in court for exacting a termination monopoly with a revenue split with Deutsche Telekom for all traffic destined to Swisscom routed over Deutsche Telekom.

Addendum 2: The authors submitted a different version of the article to Ofcom as a response to the consultation on network neutrality held in 2021. This article contains a graph, that illustrates why I am concerned. The fiber investment growth rates that are graphed are for the whole EU, the whole US and Australia and New Zealand combined. This is exactly what I’m worried about, that by the different scale of these countries, combining 27 countries in the EU and by omitting that the national broadband plans of AU and NZ went full steam in those years, the authors found correlation but not causation. The graph was omitted in the final article, so maybe there is more wrong with it. I don’t understand what is on the Y-axis. Is it in percentages? Is it the growth of the growth? The USA goes negative at 2017, so is that less fiber, or just less growth than a previous year? To me it shows the authors use very few data points and a too short time scale for their anlysis.

In its 2022 net neutrality review Ofcom mentions the article in a footnote; 271 We are aware of some academic empirical evidence suggesting that net neutrality rules can exert a negative effect on network investments and other outcomes, although the scale of the impacts found and methodologies create some uncertainty about the probative value of these results (e.g. Briglauer W, Cambini C, Gugler K, & Stocker V, 2021, Net Neutrality and High-Speed Broadband Networks: Evidence from OECD Countries). This suggests that Ofcom isn’t convinced by the contents of the article either. Ofcom is currently doing a consultation on this review.

The e-mail with requests and criticism

Dear authors and editors,

I read with interest your article on “Net neutrality and high-speed broadband networks: evidence from OECD countries” in the European Journal of Law and Economics. https://link.springer.com/article/10.1007/s10657-022-09754-5 I hereby request a sample of your data to confirm my suspicion that the econometrical analysis mostly relies on the increase in fibre roll out due to the Australian NBN and the structural separation of Chorus in New Zealand and possibly some other countries. However, given the lack of theoretical basis it’s probably best to repeal this article straight away and fix the econometrical analysis in another future article.

I spoke to <author X> today to express my concerns. I understand modelling and data collection is hard. He has not just a unique dataset, he also created a hole new type of datapoint with regards to fibre based networks. By summing up the individual types of networks from VDSL2, Docsis to FTTH, I’m almost certain there will be some weird artifacts in the data that will make it harder to do proper analysis. I also realised that at least <author X> wasn’t aware of the legal context of netneutrality and didn’t know that termination payments are legal in the European Union. This means that the entire theoretical basis for the article falls away as the entire Appendix 1 describes a situation of interconnection. As BEREC states:

Para 50. As Article 3(3) concerns the equal treatment of all traffic “when providing internet access service”, the scope of this paragraph excludes IP interconnection practices.

https://www.berec.europa.eu/sites/default/files/files/document_register_store/2022/6/BoR_%2822%29_81_Update_to_the_BEREC_Guidelines_on_the_Implementation_of_the_Open_Internet_Regulation.pdf

It is very good that you have attempted to do empirical work on net neutrality, instead of the theoretical work that is so prevalent in industrial economics. It is unfortunate that I find so many methodological, data and factual issues with the article. These are in a large part caused by the theoretical errors in the work of your colleagues.

Request for sample of data

I would like to request access to a sample of the underlying data used to do the econometric analysis. In particular the proprietary data from the FTTH Council (I assume that is the iDate data?). The sample I would like to receive is:

the values used for fiber-to-the home (FTTH) and fiber-to-the building (FTTB), as well as the hybrid fiber technologies fiber-to-the cabinet (FTTC) and fiber-to-the last amplifier (FTTLA).
Values for basic broadband and basic cable used
Countries: Australia, Belgium, EU27 (totals), France, Japan, Korea, the Netherlands, New Zealand, United States
period 2009–2012 and for 2014–2019, or if that is no problem, the whole time period, however I can imagine you don’t want to hand-over proprietary data.

This sample would be enough to confirm or refute my suspicion of serious methodological issues with the compilation and analysis of the dataset.

Theoretical model doesn’t apply to the empirical situation.

Your theoretical model doesn’t describe the netneutrality regulation in the European Union and its effects. To make it worse, it also doesn’t describe how the telecommunications market functions for ISPs. I know the theoretical models you reference have been widely used and discussed in academic literature. I’ve contacted several of those authors through the years. The answer I generally get is that the model is theoretical and isn’t intended to be an accurate description of how interconnection between CPs and ISPs functions. So that puts the burden on you to verify whether your model actually does reflect the situation in the EU.

With regards to the theoretical underpinnings both in the literature review and Appendix 1 the major issues are:

The theoretical description of how net neutrality affects investment decisions lacks verification of whether the theoretical assumptions hold in the empirical situation analysed. In my opinion it should have been included.
The model in Appendix 1 assumes a two-sided model. The assumption that an ISP is a platform provider is very common in economic theory, but wrong. An ISP is not a platform provider in a two-sided market. It doesn’t meet the essential criteria (see https://pure.uvt.nl/ws/portalfiles/portal/1410181/2012-008.pdf for an overview of such criteria, for example positive network externalities). Indeed, as BEREC says in para 15. BEREC understands the term “connectivity to virtually all end-points” as a consequence of the fact that the internet is a distributed system where a single ISP controls a rather limited part. Due to reasons outside the control of an individual ISP (e.g. technical limitations, the policy of other ISPs or regulation in some countries), not all end points might be reachable all of the time. However, such a lack of reachability should not preclude that the service is defined as an IAS.: This can be easily demonstrated. The Internet is a network of interconnected networks. There are currently 75,219 of those interconnected networks with their own routing policy active on the Internet. https://bgp.he.net/report/netstats. 3000 of those in Germany, 700 in Austria, 1700 in the Netherlands, 1400 in Italy. Each of these networks provides access to all those other networks. The customer of the ISP doesn’t choose the ISP based on how many networks it connects to. It is assumed all can be reached. Indeed, if you read this at university, then your national research network ACOnet, DFN, GARR is how you access the Internet. Consumers choose ISPs based on who is available, services offered and price. Where for TV the number of channels might be relevant, it is definitely not relevant for the Internet. The theoretical model is not an adequate base to determine the channels through which regulation affects investment. Empirically the internet works more like 75 thousand individual networks who have a choice to buy transit, build out fibers and to interconnect directly or through others. To put It differently; let us say you choose to start a website about your cat and that website becomes really popular in Argentina. Neither your ISP here in Europe or the ISPs in Argentina have done anything in that market other than carry the bits. Just like the postal service carries the letters and packages. You as content provider didn’t choose to host the website with a European ISP to become popular in Argentina, but you nothing prevents it from becoming popular there.
The model assumes that net neutrality rules would lead to a Zero Price Rule, which would mean that ISPs can’t charge Content Providers for interconnection. There is no country that has net neutrality rules that contain a Zero Price Rule. The EU rules most certainly don’t have a ZPR. As BEREC says in para 50; IP interconnection is excluded. That networks can and do pay for terminating traffic can be empirically shown. It is why Deutsche Telekom could charge German universities money when students weren’t able to follow online classes during Covid lockdown. https://www.heise.de/newsticker/meldung/Deutsches-Forschungsnetz-und-Telekom-Peeren-in-Zeiten-von-Corona-4694172.html Arcep reports termination fees (paid peering) of up to several euro per mbps. https://en.arcep.fr/uploads/tx_gspublication/report-state-internet-2022-300622.pdf WIK reports that Deutsche Telekom charges around 20cts per mbps for terminating access. https://www.bundesnetzagentur.de/EN/Areas/Telecommunications/Companies/Digitisation/Peering/download.pdf?__blob=publicationFile&v=1 The Dutch ACM also explains this in significant clarity in chapter 4.3 of its market study on IP interconnection https://www.acm.nl/system/files/documents/study-into-the-market-for-ip-interconnections-2021_1.pdf
The theoretical model would be best served by looking at various types of interconnection models. In 2013 I explained the various models in this OECD paper https://www.oecd-ilibrary.org/science-and-technology/developments-in-mobile-termination_5k9f97dxnd9r-en The model to apply in this case is a combination of peering and transit, which I also explain here https://arstechnica.com/features/2008/09/peering-and-transit/ A termination monopoly best describes the case of telcos such as Orange, DT and Swisscom. For the Swisscom case see https://www.inside-it.ch/exklusiv-naechste-runde-im-peering-streit-zwischen-init7-und-swisscom-20220506
Your model can’t explain why many ISPs, including large ones such as BT choose to peer and don’t ask for a termination payments. The model of peering and transit does provide a basis for this. Furthermore your two sided market assumes that there is some benefit obtained in the quality of the network by paying the ISP. In practice we see that networks aren’t saturated and that payment doesn’t lead to better performance. Traffic growth is not an issue. See for example BT https://indico.uknof.org.uk/event/42/contributions/555/attachments/752/924/UKNOF40-MCRAE-WEBSITE.pdf
Your model can’t explain why Deutsche Telekom decided to route all of the traffic of its Dutch subsidiary through Germany in 2019, breaking the Internet for its Dutch customers. Nor why it reversed that decision within a few days. It’s the same platform, with the same customers, so why didn’t the other networks pay Deutsche Telekom to be connected to the customers they lost connection with? https://rudolfvanderberg.medium.com/t-mobile-nl-routed-all-internet-traffic-through-germany-and-broke-the-internet-for-small-firms-a176855d2b0
Net neutrality regulation in the EU and USA focus primarily on non-discrimination and anti-competitive bundling (eg zero-rating). How this can affect investment incentives, your theoretical analysis doesn’t address at all.
You have no estimate of the size of the theoretical effects based on empirical data. Even if ZPR would be a channel, we would be talking about 20ct/Mbps per month for Deutsche Telekom all across Europe, which given 80 terabit capacity would amount to at most €200M/year. Mind you, that is a very high estimate. Interconnection traffic in France is according to Arcep 43Tbps or 1.5Mbps/subscriber peak. So at 20cts/mbps we would speak of 30ct per subscriber/month and €9 million/month for all French ISPs. Given that average French broadband subscription prices are around 25–30/month ex VAT, we’re talking about 1% of revenue.
The quote of Akerman is altered and doesn’t represent the intent or the meaning of the original quote. You say: In a similar vein, Akerman et al. (2015), examining basic broadband investment, summarize as follows: “We find that 89% of the variation in broadband coverage can be attributed to time-invariant municipality and industry characteristics and common time effects, while less than 1% of the variation in broadband coverage can be attributed to a large set of time varying variables.”. There are 25 different versions of that paper available around the web, your quote appears to be from earlier versions. The quote should be about availability: We find that 88 % of the variation in broadband availability can be attributed to time-invariant municipality characteristics and common time effects, while less than 1 % of the variation in broadband availability can be attributed to a large set of time-varying variables. In addition the way you use the quote in the article is incorrect. The roll out of broadband in Norway varied greatly by municipality, but those differences were attributable to variables that are stable. In this article you appear to use it as a justification that the variation between countries is similarly attributable to time invariant characteristics, which of course can’t be done. Municipalities in Norway can’t just be scaled op to 34 OECD countries. If that was true, the North of Germany should have just as good broadband as Denmark or the Netherlands.
The conclusion states that: a regulatory preference for network investments over the use of network management to avoid long-lasting or recurrent states of congestion . That is a bold statement, particularly because you can’t show a recurrent state of congestion. I have explained this in some detail to NLNOG recently, but average peak traffic per household doesn’t exceed 5Mbps. Deutsche Telekom’s Terastream network is designed for 40Mbps per subscriber peak traffic. Other networks have more capacity. Total capacity per country for interconnection therefore is 5Tbps/million subs times 2 or 3 for redundancy and capacity growth. Given that Nokia sells 240Tbps routers, there is no telco in Europe with a capacity problem. https://nlnog.net/nlnog-day-2022-the-20th-anniversary-edition/

I could continue, however the model you develop in no way describes the market situation as it functions, nor the effects net neutrality would have on financial flows between ISPs and Content Providers, nor would the effects be very significant. I’m well aware that there are many other theoretical models like yours… However those academics said it was theoretical and didn’t try to apply to empirical situations. Or said differently, those models aren’t even wrong.

Of course, despite a lack of theoretical basis, there might still be a correlation between net neutrality regulation and investment levels! Theory is there to be improved by excellent empirical analysis. That would still mean a significant part of the article should be rewritten in chapters 2 and 3 and the Appendix 1 should still be deleted.

Econometrical analysis, first verify the assumptions for the model

As requested by <author X> I’ve done an assessment of the econometrical model. My issue isn’t with the arithmetic. The choice of data and the way those data are processed are already such a significant concern that the arithmetic performed with it is on quicksand. If you could give me access to the sample I can further confirm this suspicion.

You estimate empirical models of investment in new (fiber-based) broadband access capacities (fiber_inv). Considering the relationship between fiber investment and fiber subscription, the former is logically a pre-condition for the subscription decision of consumers. The way you do this, raises some significant questions. I would like to look at a sample of your data to see to what extent your dataset can support your analysis. Some points that show problems with the building of the model:

The way you calculate ln(fiber_inv) in a country is novel. Nobody has compiled this statistic ever before. “ln(fiber_inv). Fiber investment is measured in real terms as the absolute number of connections deployed, representing newly installed fiber-based broadband Internet access capacity in a given country”. As far as I can see from your paper and based on my own experience, I haven’t seen this approach anywhere else in literature. This puts a burden on you to describe the validity of the approach. Such a description is lacking. The other articles dr. Briglauer wrote based on this data have the same omission.
Ln(fibre_inv)=Ln(SUM:(FTTC, FTTN, FTTH, FTTB), which means that you see VDSL, Docsis and FTTH as equal technologies that are complete substitutes with no qualitative differences. This is technically not true, but also not true in the perception of customers. KPN Netherlands has 40% or less market share in areas where it has VDSL and competes with either Docsis or FTTH of another ISP. In areas where KPN has FTTH its market share against Cable networks is above 50%. So consumers in the Netherlands say FTTH is Better than cable, but cable is better than VDSL.
Ln(fibre_inv)=Ln(SUM:(FTTC, FTTN, FTTH, FTTB) also means you can’t see the difference between Ln(SUM:1, 2,3,4) and Ln(Sum:7,3,0,0).
Ln(fiber_inv) is the log of the sum of the total number of VDSL, Docsis and FTTH based homes passed in a country. Even before taking the log, there ae some elements to consider: This means that some countries have significantly more than 100% coverage. For example my own country, the Netherlands. However, what is unclear is how log of the sum of these numbers prevents you from missing replacement investment in fiber based networks.

This can also be shown by data from Australia: The Australian NBN has replaced the Telstra network. That looks like this snapshot below: https://www.accc.gov.au/regulated-infrastructure/telecommunications-and-internet/national-broadband-network-nbn-access-regulation/nbn-wholesale-market-indicators-report/june-quarter-2022-report The data from NBN in the excel says the total of FTTC, FTTN, FTTB, FTTC and HFC is 8.2 million, out of a total of 8.7 million connections. When looking at the speed tiers provided, a different image emerges. 2.1M connections are 25Mbps maximum. That’s basic DSL and wouldn’t qualify as fiber based in most countries.

To further complicate the analysis.. In the Netherlands some HFC networks have shut down their cable network and completely moved to FTTH. It would appear that the iDate data would show a move from one category to another for this shift, but it wouldn’t show a difference in ln(fiber_inv). Similarly KPN has VDSL2 active in much of the country, but actively moves customers to FTTH. It intends to shutdown its copper network. Such a shutdown would show as a decrease in fiber investment.

Ln(fiber_inv) would need a thorough analysis for the various categories it comprises of and how the changes in the categories affect the dependent variables.
Activation of technology can mean that a connection moves from one category to another without significant investment. VDSL which your data probably qualifies as Fibre based compared to DSL was switched on in France on October 24th 2014. It was a big bang. Many French ISPs had been ready for a few years already, but for technical reasons had to wait. In my apartment bandwidth increased six fold on that day.
<author X> appeared to think that all fiber based networks require significant build out to street cabinets etc. This is not the case for many VDSL based networks and the fiber build out for most HFC based networks was finished at the start of the millennium. At the same time loop lengths in France were very long because of France’s late deployment of telephony. It had to deploy fiber much deeper to villages at the start of the millennium to be able to do DSL at all, because DSL is limited to 5KM.
There appears to be a mistake with the definition of EU+. For some reason table 6 omits: Bulgaria, Croatia, Finland (OECD), Iceland (EEA, OECD), Latvia (OECD), Lithuania (EU, OECD), Luxembourg (EUOECD),
With regards to the NNR dates the date that the regulation went into force is more correct.
Table 6 has bad and out of date references. In general it would be appreciated if you at least cite a government or other document with full name and title, not just some link.
NNR dates of Table 6 have several errors:

The OECD Communications Outlook 2013 table 2.9 isn’t referenced properly in the references, despite being mentioned in the table. I was responsible for that table when I worked at the OECD. Which also means I should have paid more attention to the responses of the countries. https://www.oecd.org/sti/broadband/2-9.pdf My apologies for not catching Finland’s error.
The NNR date for the EU is set at the time of the EU Directive adoption, however it went into force on April 30th 2016.
Finland didn’t have net neutrality regulation. The reference is probably from the OECD table, but article 68 which is mentioned in the OECD table and which you appear to refer to was already in the telecom law of 2003 and is the same non-discrimination rule present in other EU telecom laws. https://www.finlex.fi/en/laki/kaannokset/2003/en20030393.pdf
Israel had a netneutrality law for mobile in 2011 extended to fixed in 2014.
The Netherlands; the date the regulation went into force is 1–1–2013.
For some reason Norway is omitted in your list. It had NNR in 2007

The description of L.NNR also shows a misunderstanding of how public policy is formulated. It generally takes significant time for a country to formulate policy of any kind. The formal introduction in parliament is the result of a process of several years. What you currently estimate is how controversial the rule was and how efficient the country’s parliament, not necessarily whether the industry could be prepared. In the Netherlands our debates on NN started in 2006, when I wrote a discussion paper on the topic for the Ministry of Economic Affairs.

Canada started in 2008 already https://crtc.gc.ca/eng/archive/2008/pt2008-19.htm
The EU had controversial debates ever since 2011 https://www.manchesteropenhive.com/view/9781526105479/9781526105479.00024.xml?chapterBody=pdf
Luxembourg had a parliamentary decision to call for net neutrality in 2011, though it pointed to the EU to realise it, it was clear Luxembourg wanted it.

The political variable “left wing” and its description shows a bias by the author towards the policy of net neutrality. One could just as well say that right wing parties favour collusion, terminating monopolies and collusion, whereas left wing parties favor level playing fields and competition.
The values chosen for left or right wing should be available for verifications. The way it is obtained I was first critical of, but on second thought looks quite clever. The EU ellections are at least less often than election in some EU countries. That said, it would need references to political science studies to show the values fit with the results of other academics.
Population density: common mistake in broadband studies is to take the total surface of the country, better is to look at the landmass where 80–90% of the population lives.
Wireless isn’t generally part of the competition of fibre networks in a country. If anything the bundling of mobile subscriptions with fixed line broadband hinders some competition in fibre roll out, as it raises the switching barriers for consumers
You say: Gov_spend measures the overall degree of governmental spending in the economy. We expect that more left-leaning governments, as well as governments showing higher levels of public spending, will favour regulatory measures such as interventionist net neutrality regulations.

This shows a bias by the authors that would need a reference. Take for example the Bush administration and the way it ballooned the US Debt.
Gov_spend is not related to broadband spending and the authors have provided no support for it. Worse,
A better variable is broadband spending by the government. Which you could use Briglauer and Grajek for. Particularly because “The most interventionist form of public engagement is in terms of direct infrastructural investment where the government owns parts of or the entire network infrastructure. Australia and New Zealand favored such interventionist approaches and so far have provided the highest per capita state aid funds.6” and The most striking difference is between Australia and the Netherlands, which spent in total $1,721 and $1.5 per capita, respectively, during the period 2003–2018. If we then realise that most of that money was spent after 2015 we can see why Australia saw growth in fiber

The Hausmann style panel comparisons shouldn’t have just lumped the EU together in a big batch. It basically changes the comparison from 34 to 8 countries. This has greatly limited the validity of your results. 2 outliers (Australia and New Zealand) can mess up your results dramatically, because it is 25% of your sample size. Some ways you could’ve improved it is:

Distribute the EU in 5 random groups of 5 countries and see how it holds
Distribute the EU in 5 groups with high FTTH compared to countries with low FTTH.
Distribute the countries based on how much cable competition they have
Distribute the countries based on the amount of government investment in broadband they have.
Etc.

You could have a look here for additional panel data https://stats.oecd.org/Index.aspx?DataSetCode=BROADBAND_DB#
The number of countries in the table, though technically correct appears to overemphasize how many you actually use for comparison. You often lump the EU 27 together
The timeline you do the research over looks long, but since much of the action happens after 2010 and particularly after 2016, the relevant number of observations is much smaller and therefore more susceptible to data problems

This also goes for government investment, the numbers for Australia and New Zealand should be corrected by year and not spread out over the analysis period. It is quite a difference; 20K per household in 5 years or in 20 years.

I am sorry it has become such a long mail… particularly because it shows how much is wrong with your work… I will just assume these are honest mistakes and stem from an academic disconnect with the empirical situation outside your institutions. If you have any additional questions feel free to call or mail. A more complex question you should ask is why your field is so focussed on making ISPs two-sided platforms. Do economists get bonus points for working on two-sided markets?

I hope you retract this paper and explain others why science benefits from your honesty.

Sincerely,

Rudolf van der Berg

Your academic research doesn’t support mobile mergers? Just say it does!

noreply@blogger.com (Rudolf van der Berg) — Wed, 19 Sep 2018 20:23:00 +0000

On September 7, Prof. Frank Verboven of KU Leuven wrote an article on Politico arguing in favor of the proposed merger between T-Mobile and Tele2 in the Netherlands. Verboven claims in the article that research he did proves investment rises significantly, while prices will rise modestly. Interestingly, the Politico article declared it was sponsored, but not by whom. After I called this out to Ryan Heath of Politico a notice was added that the sponsor was Deutsche Telekom, since then advertisments by Politico on Twitter have promoted the article.

I was alarmed that the article was commissioned by Deutsche Telekom and based itself on research by Christos Genakos (Cambridge) as lead author, with professor Valletti (Imperial College and European Commission) and Prof. Verboven as second and third author for CERRE, a think tank with many telcos, energy companies and some regulators as backers. Valletti is the head economic advisor to Commissioner Vestager. She has to approve or block this merger. That made me suspicious; is the merger a done deal already? However, Valletti commented and denied the article said anything in favor of the merger. In a subsequent tweet he clarified that mergers don’t improve investment but do increase costs for consumers.

So how can it be that Prof. Verboven claims that the same research said mergers benefit investment, when another author says they don’t? Is it just a mistake? Is there a misunderstanding? Is it just hyperbole? Or is there dishonesty at play? I looked around and I saw Verboven gave the same comments to the Belgian newspaper De Tijd on the possible entry of a new operator in Belgium. Here he also cites research he did with Marc Bourreau and Yutec Sun of Telecom ParisTech on the French market. Note: At the OECD I wrote a report on mobile markets with four and three mobile operators and the effects of increasing or decreasing their number.

Verboven uses three arguments:

Prices for consumers will increase modestly. (citing Genakos)
Mergers lead to considerable increases in investment, more than the increase in prices and don’t decrease total industry investment (citing Genakos)
New operators lower the quality of service(s). (citing Bourreau)

I will show that the research does not say this. It says the exact opposite. Prof. Verboven misrepresents the conclusions of the research. Not only do prices increase, revenues increase by hundreds of millions, whereas investment rises by only a hundred million. New networks however increase quality and service offerings. It is impossible that he wasn’t aware that the research doesn’t support his conclusions. It is also not hyperbole or a consultant answering his master’s voice. This is an academic incorrectly citing academic research he (co-)wrote.

Prices for consumers will increase.

Verboven states in the article that: “On the one hand, standard economic analysis suggests that mobile mergers tend to result in some price increases, but the magnitude of these are, on average, modest in light of the general sharply decreasing price trend observed over the past two decades. He then argues that the introduction of Tele2 in the market didn’t lead to lower prices, other than what could’ve already been expected. In addition that Fixed Mobile Convergence puts Tele2 and T-Mobile at a disadvantage and there are the many MVNOs in the Dutch market, that will constrain price increases.

These statements are difficult to square with the observed effects and contradict the report he himself was a co-author of. The introduction of Tele2 in the Dutch market did lead to the introduction of “unlimited” offers, first by T-Mobile for €35 and later by Tele2 for €25. Such offers are not common in three MNO markets. According to Rewheel, the price per GB is significantly lower in four MNO markets and the gap between three and four MNO markets is growing. Interesting are comparisons between Austria, Belgium and Germany on the one side and the Netherlands on the other side over the same period. The moment the Netherlands went from a three to four market, the prices per GB dropped and usage went up. This didn’t happen in those other countries, who have three mobile operators, to the same extent, so saying it is part of a trend and implying it is not because of competition is very odd. The effect in France in 2012 was similar with the fourth operator adding subscriptions for as low as €2 (100MB+100Min+100SMS) and €20 (for 25GB and unlimited calls). We can see similar trends in other markets with new entrants and the reverse in markets with mergers (See OECD-report)

Investment will not increase more than prices.

With regards to investment Prof. Verboven says: On the other hand, such mergers also lead to a considerable increase in the investment per operator. As a result, our findings indicate that mobile mergers do not reduce total industry investment, contrary to theoretical concerns raised by competition authorities.” This is odd because Prof. Valletti said the research doesn’t actually make those claims. The Politico article suggests that they show a trade-off between quality and price or more precise between network investment (which he uses as a proxy for quality) and price. He states that the Cerre report shows that investment grew more than the prices raised. It is however the opposite by a five to ten times.

The report states: “A hypothetical average four-to-three symmetric merger in our data would have increased the bill of end users by 16.3%, while at the same time capital expenditure would have gone up by 19.3% at the operator level, always in comparison with what would happen in the case of no merger. More realistic asymmetric four-to-three mergers (between smaller firms in European countries) are predicted to have increased the bill by about 4–7%, while increasing capital expenditure per operator by between 7.5% and 14%.” These numbers are however relative. But they are relative to different base numbers:

Dutch mobile operators had € 4.749M in retail revenue in 2017 and €167M in investment per operator (670M total). A 16.3% increase would lead to €774M more revenue. A 19.3% increase in investment per operator would lead to (670/4)x19.3%= €32M more investment per operator (€97M total). The smaller numbers are even less favorable (€356M increase in prices and €70 increase in investment). That’s nowhere near a significant rise in investment. It’s like me saying my niece got 700% older in the last week, while I aged only 0.05%. It also explains why the author’s couldn’t find an overall increase in investment in the market. Using their numbers it’s clear that in the Dutch market the fourth operator’s investment wouldn’t be fully replaced. At the same time consumers would pay over half a billion more per year… for what exactly?

There are some arguments that might explain why investment is roughly the same in a three and four operator market. One example we found in the OECD-report is that in a number of countries with four operators, there were initiatives to share some of the fixed infrastructure, such as sites, poles and sometimes even active equipment. This was less prevalent in markets with three operators. One could also imagine that third parties, such as tower companies are more active in four operator markets, leasing out locations to multiple operators. This allows costs to be shared over multiple operators and also saves CAPEX.

New operators increase quality

With regards to quality he states that “The merger will create a mobile telecom operator with a sufficient customer base to invest in a nationwide 4G and 5G network and compete effectively with the two larger fixed-mobile operators.” The arguments he gives are that it will allow the two to spread investment, particularly for 5G, over both their customers and that locations to place antenna are scarce and less competition would allow better coverage. On the Belgian market he even states that an extra operator would lead to decreased coverage of rural areas. As stated in the previous paragraph in four operator market there is often more sharing of infrastructure. Which is for example the case in the Netherlands between Tele2 and T-Mobile. This doesn’t mean that the coverage is significantly less. Indeed, coverage in the Netherlands is for all operators in the top of the world according to various metrics.

According to professor Verboven his research showed that the introduction of a fourth operator in France led to lower priced, but also lower quality offers. With regards to the situation in France, the article by Marc Bourreau prof. Verboven references states that: “We show that the incumbents’ launch of the fighting brands can be rationalized only as a breakdown of tacit collusion in product lines: before entry the incumbents can collude on suppressing low-cost brands to avoid cannibalization, and after entry this semi-collusion becomes more difficult to sustain because of increased business stealing incentives.” Though in the article low quality is mentioned a few times, it is unclear what the author’s with low quality as the incumbents use their challenger brand on their own network. It appears its more perceived quality and maybe the internet-only nature, than the actual offer.

Quality is harder to measure, as it can mean many things to many people. The experience in France is that the new operator Free increased quality in service offers by adding contracts that could be cancelled any time, free international calls (now more than 100 nations), free roaming (also in a number of nations outside the EU) and innovations in purchasing and leasing handsets, all in a fixed price offer of €20 (or €16 for fixed line customers). The competitors followed in their offers. There were complaints about coverage, however it met its coverage obligations. It also leased access on Orange’s network for increased coverage, so it’s hard to say this is a lower quality offer than the one of its competitors. For France the OECD report found also that operators pulled their investment in 4G forward by 2 years to better differentiate from Free. So for the French quality increased on many different scales.

Indeed the article states that low-cost brands aren’t necessarily lower quality as they often introduced new features that appealed to customers in various ways. The article ends with the statement that: “A general conclusion from our analysis is that concentrated market structures may facilitate tacit collusion on restricting product variety. This adds to traditional concerns that increased concentration may facilitate collusion in prices or quantities, with a potential of generating considerably larger losses to consumers and welfare.” So it is difficult to see how Prof. Verboven reaches the conclusion that in France the introduction of Free led to lower quality for consumers, or that a fourth operator in Belgium would lead to a significant impact on quality and lower investments.

Conclusion

All in all, I find it very strange that a professor of a renowned university makes claims in sponsored articles, aimed at influencing decisions by the competition authorities of the European Commission that aren’t supported by the studies referenced. In addition that the claims he makes obfuscate the relative size of effects. Yes, investment may grow a bit more in percentages, than the costs to consumers, but in Euro the last number is an order larger. His call on the Commission to apply cutting edge economic analysis, begs the question whose analysis he’s referring to as it appears his own research doesn’t support the statements he makes in this article.

Is the EU going to make it illegal to put church services on the Internet?

noreply@blogger.com (Rudolf van der Berg) — Fri, 29 Jun 2018 22:17:00 +0000

This is a call to action to my fellow Christians (Jews, Muslims, Buddhists, Hindus and others join too). Share this widely. Please contact your party’s representation in the European Parliament as they are about to hurt churches through copyright law and they appear oblivious to it. The new EU copyright law proposal https://ec.europa.eu/digital-single-market/en/modernisation-eu-copyright-rules that went through JURI Committee last week and is up for EU Parliament vote next week July 5 is the cause. There are two articles in it that churches should oppose, Article 13 (filtering) and Article 11 (Link-tax). The effects will be devastating, both for copyright law and for churches. Fortunately there are fixes possible.

It struck me that when I researched it, none of the representatives of Christian parties (ECR, Conservatives and Christenunie and EPP, CDA, CDU etc.) in the EU Parliament are opposing it. Worse, it’s actively proposed by the EPP-faction’s mr. Voss. They are heavily in bed with Big Media, forgetting most churches rely on small media and social media. The law is opposed by some liberal and left-wing parties amongst which the copyright oriented Pirate Party. So we must act and tell the Christian parties they are hurting the very people that send them to Brussels in the first place.

Now, I’m not the greatest copyright buff, I dabble. I’m also not the best Christian. I go to an non-denominational evangelical church occasionally (monthly??). I do care about both of them quite a bit as both have brought me great joy (and bitterness on occasion).

I base myself here on the works of EDRI.org https://edri.org/files/copyright/20180626-ReDeconstructing_Article13.pdf and of Julia Reda of the Pirate Party https://juliareda.eu/2018/06/the-internet-after-axel-voss/ who has some solutions. Some readers may grumble that those left-wingers and churches don’t always see eye to eye. However, those people take their copyright law as serious as many churchgoers take their Scripture, so let us work together and stop this bad proposal. If enough parliamentarians vote against this on Wednesday, it can still be blocked.

Article 13: the upload filter brings bad luck.

Anyway, why should churches (and other religious communities) be worried by the proposed changes to EU copyright law? Well, it’s particularly the proposed article 13 that would require intermediaries to filter any type of copyrighted content, up front, automatically, before it’s being published. This article makes the intermediary, Youtube, but also, your national Anglican-Lutheran-Reformed-Catholic-Orthodox-Evangelical church organisation is liable for all damages through copyright infringement. Let’s say you want to have a nice site with all the services of all the congregations you represent. Now, you get treated like the party responsible and liable.

That means that if you want to put your church service online, the intermediary has to check if anyone owns the copyright to the content of the service. Of course they do! A church service is 6 copyrighted songs, with a preacher citing from one or more copyrighted books (Yes, the Bible is copyrighted (except the KJV or Statenbijbel) and for good reasons, but still). And don’t tell me your church only sings century old hymns that are out of copyright, because some choir, singer or organist will have recorded these songs and their record label may have registered the copyrights to that performance. Youtube has a filter to check whether you violate any copyrights, but it’s not yet set to the toughest setting. If your church has a nice site with Wordpress or a local hosting firm and puts its services there, that site and hosting company is liable too and so they have to filter too. Unfortunately computers aren’t smart enough to know all intricacies of copyright. It doesn’t know if your church paid its copyright dues. or that the one who registered the content is using the same centuries old song, but played on an organ in your church by a famous organ player and then put it on a CD. So it all gets filtered and blocked.

A redress mechanism has been proposed that if you complain hard enough the content should get reinstated. Well, we all know how that works, it will be slow difficult and cumbersome and most won’t bother.

Article 11, just to make life of religious publishers harder

Article 11 is known as the link tax. It’s intention is that companies like Google should pay for using small snippets of an article on Google news. The way these link taxes have worked out in Germany and Spain is that anyone who cites news has to pay for it. That means any church or religious organisation’s newsletter has to pay. You can’t run a blog, a Facebook page, or anything where you point people to news that’s relevant to them, without paying the publisher. In Spain the effect was that these small aggregators ceased business.

In addition the effect in Spain and Germany was that Google News stopped indexing news sites. Traffic to these sites dropped. However it hurt smaller publishers way more than large ones. A Christian oriented newspaper, like Trouw, Reformatorisch Dagblad or Nederlands Dagblad in The Netherlands is by definition focused on a small group of people. Google News at least shows those that search that these papers exists and their opinion can carry weight. And when it’s visible, people click on the link and read. Without Google News people go to the Daily Mail, Bild and de Telegraaf.
Just think of how often religious organisations share snippets of news in their newsletters. What if you had to clear the rigths for that?

So why didn’t I write this earlier?

Well, as I said, I’m not a copyright buff. I thought that things would solve themselves. If the inventors of the Internet, Vint Cerf and Tim Berners-Lee oppose it, If Wikipedia opposes it, if a whole bunch of scholars oppose it, we should be fine, but we aren’t. Big media and national broadcasters hate Google, Facebook, Netflix and others. This is their payback for 20 years of them having to change their business models. The sad thing is, the effect won’t be that they will do better, just that everyone is worse off, but at least it’s impossible to compete with them. With regard to churches, this means that some big churches and churches abroad (in the USA) can afford to put content online. For smaller ones it will become too hard. We all like an hour of Songs of Praise on the BBC, but that won’t get people into a local church. Protecting the national broadcaster so that his content doesn’t get uploaded is nice, but it shouldn’t prohibit a church or religious organisation from uploading theirs.

Oh and if you think I might be wrong, have a look at these tweets and the reactions to them. The EU said funny pictures (memes) weren't affected. A quick reply from me caused a furor. (I never had this many reactions to a tweet before!

pic.twitter.com/4oXlg0uIzj
— European Commission 🇪🇺 (@EU_Commission) June 22, 2018

Because you (parliament) put it in a new directive #article13. Which requires people to filter for copyrighted content. And that picture of Jean Luc Picard is copyrighted, though its use might be fair use. Except nobody knows until after the lawsuit, so platforms will block it!
— Rudolf van der Berg (@internetthought) June 22, 2018

I am at the IGF 2014 in Istanbul

noreply@blogger.com (Rudolf van der Berg) — Mon, 01 Sep 2014 07:18:00 +0000

The Internet Governance Forum is held in Istanbul this year. I am fortunate enough to go to the meeting as part of the OECD team, we're here with a team of six. It will be my first IGF.

I organise one meeting on "Policies and practices to enable the Internet of Things". The proposal was ranked third! That's great and scary!

I am invited as a speaker in two others:

As always, please feel free to contact me. Twitter works, but I also have a local Turkish number +90 553 5471473. (Not a bad deal, 1GB of data, 100 minutes to Europe and 250 minutes and 250SMS local for €31 euro with Avea)

Speaking on Connected Television at IIC/Ofcom on May 2nd

noreply@blogger.com (Rudolf van der Berg) — Wed, 01 May 2013 09:23:00 +0000

It has been very quiet here for a while. I am of course active on Twitter, but I heard some of my readers don't use Twitter (you should, it is practical, don't tweet, just follow)

May 2nd I will be in London speaking on Connected Television. If you are around, leave a message. Maybe we can do coffee on May 3rd (or a beer late in the evening)

Connected TV and Platforms: evolution or revolution?

International Institute of Communications
UK Chapter Event

Connected TV and Platforms: evolution or revolution?
Market trends, dynamics, and policy implications

Thursday 2 May 2013, 17:00 for 17:30 Start

Kindly hosted by Ofcom
Riverside House, 2a Southwark Bridge Road, London SE1 9HA

In yet another sign that convergence is finally upon us, TV sets are increasingly connected to the Internet, and advertising for television sets seems increasingly to focus on which ‘cool apps’ they feature, rather than their size or picture quality. And it’s not just TVs: DVD players or PVRs now come connected to the Internet. The possibilities are vast, yet this is only a nascent market – and one which already raises many questions.

How prevalent will Connected TVs become in coming years? What impact will they have on the delivery of online and broadcasting content? How will broadband television change the TV market? What changes can be expect from Satellite TV? How will demand and usage patterns for connected TV evolve? Are there risks of vertical integration and generally for competition? Will the second screen be the interactive part of television? And for the legal wonks: what are the implications for content regulation and audience protection, and should the Audiovisual Media Services Directive (AVMS) be revised or renewed to accommodate the development of a connected device market?

To discuss this exciting and rapidly evolving phenomenon we will hear from a range of speakers, from different industries and from the public and private sectors:

- David Mahoney, Director of Policy Development, Ofcom, will host and moderate the discussion

- Marcel Boulogne, European Commission

- Rudolf van der Berg, OECD

- Chris Hutchins, Vice-President, EU Affairs, Liberty Global

Numbers are limited and you are advised to register as soon as possible. Attendance is free to IIC Members and £30 for non-Members.

Please note that registration for this event is now closed. Please contact j.grimshaw@iicom.org with CH1-13 in the subject heading for further information.

For IIC membership details please contact the IIC at enquiries@iicom.org.

DATE FOR DIARY: 21 MAY: 4G – Prospects, concerns and policy implications.

For further information or to register your interest contact j.grimshaw@iicom.org.

For enquiries please contact the IIC secretariat on +44 (0)20 8417 0600.

Interview and Conference on The Internet of Things

noreply@blogger.com (Rudolf van der Berg) — Tue, 25 Sep 2012 09:31:00 +0000

In COMMUNICATIONS & STRATEGIES No. 87, 3rd quarter 2012 Internet of Things: A new avenue of research there is an interview with me conducted by Martin Cave. I go into the many aspects of IoT and discuss the policy implications. The interview has been published online and you can see it here

There is also a conference at Telecom ParisTech where those who have written articles and those that were interviewed will present. http://innovation-regulation2.telecom-paristech.fr/wp-content/uploads/2012/09/Programme.pdf (the conference is tomorrow, so you need to be quick.

Presentation on M2M and market failure at Internet of Things-week

noreply@blogger.com (Rudolf van der Berg) — Tue, 26 Jun 2012 07:59:00 +0000

Last week I was invited to the European Commission's expert group on the Internet of Things to speak on OECD work. It was in Venice during Internet of Things-week, which sounds great until you realize it was 35 degrees Celsisus outside and 40 degrees inside. The European Commission currently is doing a consultation on the Internet of Things, so if you have opinions, voice them! Unsurprisingly maybe, I put the focus on market failure that currently impedes the growth of the Internet of Things.

Update: Here are some links to where I got the data:
- http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/not-spots/PA_Consulting_main_report.pdf on coverage
- https://ripe64.ripe.net/presentations/172-Mobile_Broadband_Measurements.pdf that 20% of locations are unavailable more than 10 mins/day

Policy Brief on Internet of Things and Transport

noreply@blogger.com (Rudolf van der Berg) — Wed, 18 Apr 2012 08:23:00 +0000

The International Transport Forum at the OECD is an intergovernmental organisation with 53 member countries. It acts as a strategic think tank for transport policy and organises an annual summit of ministers. For this years summit on Seamless Transport I wrote a policy brief on the Internet of Things and Transport.

Building the "Internet of Things"
In the coming years the Internet will move from connecting people to connecting things. In a new report entitled “Machine to Machine (M2M) communication: Connecting billions of devices”, the OECD analyses the
impact of this phenomenon. The report

finds new sources of growth
identifies significant barriers to the functioning of the market
proposes liberalisation to further open the mobile telecom market, to enable new entrants that may be transport, energy and healthcare companies, not telecom providers
argues to support trade and travel for manufacturers and service providers in providing these services across borders,
argues that removing barriers will result in billions in direct and indirect savings on mobile connectivity, and additional billions in new revenue from new services.

In 2017, in OECD-countries an average family with two teenagers could have 25 things that are connected to the Internet: telephones, TV, tablets, printers, sports gear and health devices. But that is not all. Companies will dramatically change the way they design machines and devices, starting from the type of data they need to operate efficiently and effectively and then building the machine or device around it. Tens of billions of connected devices by 2025 is not farfetched. The combination of the data will allow smart transport, smart
cities, smart energy and smart health. for more see http://www.internationaltransportforum.org/jtrc/PolicyBriefs/PDFs/2012-04-04.pdf

Presentation at Telecom Reform Conference 2012

noreply@blogger.com (Rudolf van der Berg) — Thu, 22 Mar 2012 16:12:00 +0000

I presented at the Telecom Reform Conference in Copenhagen. Good conference. The 1969 reference in the disclaimer is to Bill Melody's presentation at the OECD on the Carterfone decision. Bill Melody was also one of the organizers of this conference. BTW 1969 is way before I was born. Nice little anecdote, in 1969 liberalisation wasn't even an acceptable topic to the US delegation to the OECD. We've come a long way since then.

(Just noticed that the nice graphs I had on Roaming and MTR's didn't get included in the Google Docs version. If you have a solution, please mail me. If you want to see them, look here (roaming) and here (MTR)

OECD report on "Developments in Mobile Termination"

noreply@blogger.com (Rudolf van der Berg) — Thu, 01 Mar 2012 12:27:00 +0000

As part of the OECD’s ongoing work to stimulate competition and innovation in mobile phone markets (see here for most recent news release on mobile roaming), this new working paper aims to contribute to the current debate among regulators in OECD countries who are reducing or considering the phasing out of the fees telecommunication network operators pay for delivering telephone calls to mobile wireless providers, known as mobile termination rates (MTRs). This is because mobile operators have a monopoly over the termination of calls on their networks.

While it finds that rates have decreased across the OECD by 53% between 2006 and 2011, from USD 0.1406 in 2006 to USD 0.0650 per minute, there is still much divergence between countries. Rates are at zero in Canada and vary from the lowest (the United States (USD 0.0007/min), Israel and Turkey (USD 0.0203/min) to the highest, including Estonia (USD 0.142/min) and Chile (USD 0.165/min).

The complexity and difference in the way that operators charge fees makes it difficult to draw a link between rates charged and prices paid by users for voice calls in different countries. But cutting rates to zero would strengthen competition in voice and other services, says the report. It could also speed up the introduction of innovative new VoIP services and encourage providers to offer a range of tariff models to meet the needs of their users, free from prices reflecting monopoly power on the networks of others.

Developments in Mobile Termination is available at http://dx.doi.org/10.1787/5k9f97dxnd9r-en.
All the data in an excel file is available here: http://www.oecd.org/dataoecd/25/44/49805921.xlsx

The author Rudolf van der Berg is also available to answer any questions at +(33) 1 45 24 93 67, (+33) 6 58 15 85 08, or rudolf.vanderberg@oecd.org

OECD publishes report on Internet of Things and M2M

noreply@blogger.com (Rudolf van der Berg) — Wed, 01 Feb 2012 10:56:00 +0000

Look around you for a second and count the number of electronic devices, machines and gadgets. All of them -- light bulbs, cars, TVs, digital cameras, refrigerators, stereos, cranes, beds – will be connected to the Internet over the next 15 years, if they aren’t already.

Farmers will connect their livestock, machinery and fields through sensors. Medical practitioners and patients, fitness fanatics and those of us that require greater encouragement, will be monitored for everything from our heart rates to glucose levels to the special needs of pregnancy.

This is the potential of the “Internet of Things”: billions and billions of devices and their components connected to one another via the Internet. 50 billion devices by 2020, according to companies like Ericsson. The “Internet of Things” will radically alter our world through “smart” connectivity, save time and resources, and provide opportunities for innovation and economic growth.

The basic building block of the “Internet of Things” is machine-to-machine communication (M2M), devices equipped to communicate without the intervention of humans. The trends are already visible: Internet-connected TVs are now widespread; eBook readers must have a Wi-Fi or 3G connection; smart electricity meters have already become standard in many countries. More often than not, however, M2M is hiding in plain sight: information terminals in trains and buses, traffic lights and bicycle sharing systems, like the ones in Paris and London.

This new “Internet of Things” is the subject of a new OECD-report, “Machine-to-Machine Communication: Connecting Billionsof Devices”, and examines:

1. New Technology: the drivers behind connecting devices to the Internet

2. New Markets: user and business demands, and whether they are being effectively met

3. New Policies: what governments can do to promote this new source of growth

New Technology

- The Internet of Things is enabled by ever cheaper communication modules, new innovations and applications of existing technology and the fact that the Internet has become available almost anywhere that people are.

- M2M is very different from traditional telecommunication. It does not need a human to start or stop a communication. As a result it has the same economical life span as the product in which it is embedded (10 years for consumer electronics, 15 years for cars, 30 years for smart meters). It can process extremely low data volume or very high amounts.

- Different networking technologies can be used to connect M2M devices, depending on the amount of mobility needed and dispersion over an area. 2G/3G/4G mobile wireless is, however, often an ideal technology for most applications.

- SIM-cards could become the standard for authentication to use networks, as this removes or reduces the need for human interaction. Devices can be connected straight from the factory.

Table 1. Machine-to-machine applications and technologies, by dispersion and mobility

Geographically dispersed	Application: smart grid, meter, city remote monitoring Technology Required: PSTN, broadband, 2G/3G/4G, power line communication	Application: car automation, eHealth, logistics, portable consumer electronics Technology Required: 2G/3G/4G, satellite
Geographically concentrated	Application: smart home, factory automation, eHealth Technology Required: wireless personal area (WPA), networks, wired networks, indoor electrical wiring, Wi-Fi	Application: on-site logistics Technology Required: Wi-Fi, WPAN
	Geographically fixed	Geographically mobile

New Markets

- M2M creates a new player in the mobile market: the “million device” user. These new large scale M2M users will potentially manage hundreds of thousands of smart meters, cars, and consumer electronics, possibly in higher numbers than some countries have citizens.

- Large scale M2M users may offer their services in 10s to 100s of countries, selling the same devices globally. Their customers may buy the devices abroad and travel with them. As a result, manufacturers need to offer international connectivity solutions. The telecommunication industry, however, is still largely organised and regulated on a per country basis. Large M2M users will thus place new demands on telecom companies, and regulation and business models will have to adapt.

- Companies creating innovative M2M-based services are currently locked into 10-30 year mobile data contracts and high roaming fees; this dependency hinders the role out of new services and innovation.

New Policies

- Governments hold the key to set large scale M2M users free, by giving them access to wholesale markets. They will need to change the rules so that large M2M users can have access to numbers and SIM-cards, just like telecom companies have now. This will open up the market, break lock-ins, make large M2M users responsible for their own innovation and create a competitive market for roaming for M2M services. This liberalisation of the market will be a major paradigm shift, but might by some estimates lead to billions in savings and new services.

- Privacy and security need to be designed into products from the start. M2M could allow a detailed view of people’s lives, and parliaments have already curbed or changed some projects as a consequence. For example, cars are increasingly using onboard M2M services (GM Onstar, Ford Sync, BMW ConnectedDrive) and the European Union is now mandating their own service (eCall) to be built into every car from 2014. Since EU legislation requires telephone companies to record a person’s location at the start of each mobile communication, and since turning a M2M car on will itself start a communication, these companies will be inadvertently tracking the start and end of any trip! The current challenge requires privacy by design, because as this example shows, even if the automobile company does not register the location, the telecommunication company by law will have to.

- Governments have tried to make spectrum policy more flexible in recent years, allowing companies to change networking technologies when new technology becomes available. M2M may rigidify spectrum policy, however, because any time M2M uses a particular networking technology, it expects the spectrum to be there for the lifetime of the device, which is 10 to 30 years. The consumer-oriented wireless technology works on a timescale of a maximum 10 years.

- Countries may run out of phone numbers in their current numbering plans as a result of M2M, because 2G and 3G equipped M2M devices require an E.164 telephone number to work. Only when 4G is used can M2M work with just an IP-address.

- It is unclear which country should issue telephone numbers for any given device. Is it the country that device is used in, the country of the mobile operator, or the country of the large scale M2M user?

- Combining data generated by M2M devices may offer great insights to improve society. Cars could notify local governments of icy roads or bottlenecks in infrastructures. This may not always be seen as positive, however, as shown by a case in The Netherlands where anonymous and aggregated data from GPS-systems was used by the police to identify prime locations for speed cameras, which led to a public outcry.

What is certain from the report is that governments will have to change regulations in the telecommunications market, will have to be vigilant to apply privacy and security regulation and stay innovative to make use of the many possibilities it offers. Doing so promises to transform the economy, promote growth in the telecommunications sector, and produce growth and efficiency savings in government and society.

Slides available: BEREC expert workshop on IP-Interconnection (Peering and Transit) in cooperation with OECD

noreply@blogger.com (Rudolf van der Berg) — Tue, 20 Dec 2011 10:18:00 +0000

Update: BEREC website was redesigned and the slides fell off. They can still be found here in this open directory http://berec.mp.bi.lv/files/doc/berec/oecd/
On November 2nd, BEREC organized a workshop on IP-interconnection together with the OECD. All the slides are now available online. And to save you from reading the whole post, here they are:

Dennis Weller, Senior Advisor, Navigant Economics "IP Traffic Exchange; Market Developments and Policy Challenges"
Martin Levy, Director, Hurricane Electric "The Background of Internet Interconnection"
Scott Marcus, Director, WIK-Consult, "IP interconnection, traffic trends, and wholesale and retail prices"
Bill Krogfoss, Strategy Director, Chief Technology Office Alcatel-Lucent "Internet Economy and Content Peering"
Falk von Bornstaedt, Head of Peering, Deutsche Telekom "IP Interconnection and differentiated QoS."
Patrick Gilmore, Chief Architect, Akamai
Eric Ralph, Chief Economist of the Wireline Competition Bureau, FCC (Video message
and teleconference); (no guarantee that the video works).

This was a direct result of the OECD High Level Meeting on the Internet Economy in June, where through conversations in the halls we noticed that there was a difference in the way regulators and Internet peering coordinators discussed interconnection. The workshop was a huge success, several people from the internet peering community flew in from Vienna for one day even though there was a RIPE-meeting going on too.

A couple of points to take away from the meeting are:

Peering agreements are for 95%+ handshake agreements, without any involvement of lawyers. They take 3 minutes to set up technically.
Both content providers and eyeball networks are working on getting content closer, faster and cheaper to the consumer. This benefits both parties. An interesting example were Google Global Caches which are now placed in networks around the world, but whose effect can most profoundly be seen in Africa where for instance the traffic over the Kenya Internet Exchange Point increased by several hundreds megabit/s peak after a cache was installed, saving the local internet community hundred thousands of dollars every month
The market for peering and transit is highly competitive and highly dynamic, with switching barriers being extremely low. A change in routing from one transit provider to another can literally be done in minutes. The effect of a peering agreement is almost immediate.
The people whose job it is to interconnect IP-networks speak of themselves as a community, even though some of them disagree quite considerably on how it should be done and who plays what role.
Peering and transit is done between all kinds of networks, even the European Commission has an AS-number and could set up it's own peerings.

I think everyone looks back at a very successful event, that was seated to capacity.

BEREC expert workshop on IP-Interconnection in cooperation with OECD

November 2nd, Bloom Hotel Brussels, 9:00-17:30

The goal of the workshop is to bring experts from the IP- interconnection community in contact with experts on interconnection from national regulatory authorities and to discuss future interconnection in an all-IP world.

The Internet’s way of using peering and transit as the basis for commercial negotiations differs considerably from the telephony’s world of Calling Party’s Network Pays. As a result even when talking the two worlds seem to be speaking about different things even when using the same words.

BEREC has been looking into these different approaches to interconnection in a series of papers since 2007 (ERG Report on IP-Interconnection 2007, ERG Common Statement on Regulatory Principles of IP Interconnection 2008, BEREC Common Statement on NGN future charging mechanisms, 2010). The OECD has studied Internet traffic exchange in a series of reports in 1998, 2002, 2005 and in a forthcoming paper in 2011. Furthermore, it has studied Internet traffic exchange in relation to the development of local content in cooperation with UNESCO and it has also organized a workshop on the topic in 2001 in cooperation with the German government. IP-interconnection markets are global markets crossing national borders and even continents. Therefore the OECD is singular in its analysis of trends in Internet interconnection taking a global perspective. Following the same objective of safeguarding competition, BEREC and the OECD take this workshop as a starting point hopefully to be continued in the future.

The format of the 4 session is intended to allow for an extensive discussion with the audience.

Programme
9:00-9:30 Registration and Coffee

9:30 – 10:00 Opening words by Monica Ariño, BEREC and Sam Paltridge, OECD

10:00-11:30 Session 1: The background of Internet interconnection
The goal of this session is to outline the basics of Internet interconnection.
Technical background: Peering (paid or free), transit, partial transit, variants (reciprocal transit
etc.), “Public” versus private, application needs for QoS.

Dennis Weller, Senior Advisor, Navigant Economics "IP Traffic Exchange; Market Developments and Policy Challenges"
Martin Levy, Director, Hurricaine Electric "The Background of Internet Interconnection"

11:30-12:00 Coffee break

12:00-13:00 Session 2: IP interconnection, traffic trends, and implications for
wholesale and retail prices
The interconnection of internet networks has been described in terms of two-sided networks.
The network provider stands in the middle and can receive money from either the content
provider or the consumer. Is this description of the market accurate? What can the theory of
two-sided markets teach us? What contribution can content providers give to the deployment
of networks or alternatively what is the role of network providers in content?

Scott Marcus, Director, WIK-Consult, "IP interconnection, traffic trends, and wholesale and retail prices"
Bill Krogfoss, Strategy Director, Chief Technology Office Alcatel-Lucent "Internet Economy and Content Peering"

13:00-14:00 Lunch

14:00-15:30 Session 3: IP Interconnection and differentiated QoS

Mike Blanche, Program Manager - Network, Peering and Content Distribution, Google,
Falk von Bornstaedt, Head of Peering, Deutsche Telekom "IP Interconnection and differentiated QoS."
Patrick Gilmore, Chief Architect, Akamai

15:30-16:00 Coffee Break

16:00-17:10 Session 4: The Future of Interconnection
The Internet hasn’t done away with telephony as a very important means of communication. The growth of mobile telephony has even been more dramatic than the growth of Internet. There have been calls in academic journals and on regulators directly to both impose the Internet’s way of interconnection on telephony and vice versa the telephony’s way interconnection on the Internet.

Can we expect the market to evolve into one model or the other for all traffic?
Could a hybrid model evolve?
What role will new services on networks play? Will new demands be placed on
interconnection?
What characteristics of both models should survive?

Panel:

Eric Ralph, Chief Economist of the Wireline Competition Bureau, FCC (Video message
and telco);
Andreas Sturm, De-Cix;
Mike Blanche, Falk v. Bornstaedt, Patrick Gilmore; Martin Levy

17:10-17:30 Wrap-up

Cara Schwarz-Schilling, BEREC and Rudolf van der Berg, OECD

Will speak at FTTH Forum Budapest on Nov 9

noreply@blogger.com (Rudolf van der Berg) — Tue, 08 Nov 2011 11:46:00 +0000

I will speak at the FTTH Forum in Budapest tomorrow. Always interested in meeting up with people. Many people of the so called Fiber Ring will be there ie Benoit Felten, Costas Troulos and James Enck

New presentation on Machine to Machine communication

noreply@blogger.com (Rudolf van der Berg) — Thu, 30 Jun 2011 16:47:00 +0000

I recently gave a presentation on Machine to Machine Communication to the Dutch Telecommunication User Group, BTG. This was based on work I have done at Logica. The presentation represent my own thoughts and not those of my employer. I am still working on the topic and am interested in hearing anyones views on the topic.

M2 m implications for btg

View more presentations from Raindeer

Born: Tycho Rudolf Willem

noreply@blogger.com (Rudolf van der Berg) — Sun, 12 Jun 2011 12:01:00 +0000

The blog is on semi-hiatus, but I thought some of you would appreciate this news. On Wednesday June 8th our son Tycho Rudolf Willem was born in Almere, The Netherlands. He is the little brother of Grace :-) Mother and son are doing great. Pictures available on request :-) And yes, I was present, the two days before were spent in OECD working groups in Paris, discussing two papers I wrote with member countries.

I'll be speaking on WIK's conference on Bundling and Multi-Play

noreply@blogger.com (Rudolf van der Berg) — Sat, 28 May 2011 22:38:00 +0000

I'll be presenting at "Bundling and multi-play: Does it require a new regulatory paradigm?". This conference organized by the German WIK-Consult is held Monday and Tuesday in Brussels. I'll present OECD-work on Bundling, which was published last year. It's not my work, but the responsible colleagues were otherwise disposed or had changed to other jobs. The line up of speakers is quite interesting.

First blogpost for the OECD... on International Women's Day (of course)

noreply@blogger.com (Rudolf van der Berg) — Thu, 10 Mar 2011 09:39:00 +0000

Working for the OECD may mean this blog is on hiatus, but that doesn't mean I stop blogging completely. Fortunately the OECD has some blogs and I've been allowed to write on International Women's Day. It's not a topic I'm an expert in, but as always I can have an opinion and I've got some skills in writing it down, so for your pleasure:

May the best (wo)man win

MARCH 10, 2011

by Guest author

8 March is the centenary of International Women’s Day. This year, we mark the occasion with a series of blog posts about initiatives to strengthen gender equality worldwide. In this post, Rudolf van der Berg of the OECD’sScience, Technology and Industry Directoratediscusses sex discrimination in management.

As we celebrate the centenary of International Women’s Day, in OECD countries still only 5%-30% of senior management is female. This is often discussed in the context of right and wrong. But let’s look at it from the perspective of competences and economic performance.

The argument for there being so few women in management positions is that they just aren’t available and that the best person for the job needs to be promoted. However, with educational equality being the norm and girls and women actually performing better than boys and men in education, this argument seems to hold less than it did a hundred years ago on the first International Women’s Day.

The simple conclusion we have to draw from the numbers is that if competences and chances were distributed equally, we would see roughly 50% women on boards. Unfortunately we aren’t seeing this. We still see far less.

The sobering conclusion is that if our societies were more equal, 40% of the men in some form of management today wouldn’t be in that position. They are not competent enough, given the pool of talented women. However, since managers are continuing the practice of selecting men over women for any management position in the organization, this means they are dipping deeper into the pool of available men and pushing lesser qualified men up to positions they shouldn’t have been in.

The consequence of this is that the position of men in organizations is supported by a whole column of lesser qualified men, crowding out women and in their own self-interest keeping those women out.

That this has negative effects on the performance of companies is clear. In a series of studies byCatalyst, Fortune 500 companies with more women board directors are shown to have significantly better financial performance, including 53% higher returns on equity, 42% higher returns to sales, and 66% higher returns on invested capital.

If shareholder value or government effectiveness and efficiency is dependent upon having the right people in the right place, it is time CEOs and Ministers start looking around and down and wondering which of the men they see in management is the weakest link, and where in the company the more qualified woman has ended up.

Useful links

OECD work on corporate affairs

OECD work on gender

WikiGender

Wikiprogress

This blog is going on (semi-)hiatus

noreply@blogger.com (Rudolf van der Berg) — Wed, 05 Jan 2011 14:44:00 +0000

It has been a blast to be writing blogposts here. It was conceived as a way for me to vent of the ideas in my head and to be in contact with others who are interested in telecoms. Over 350 people now subscribe to my feed, articles get read and linked to. I've actually become a bit known in the industry because of you my readers, reacting and forwarding the articles.

However, as you know, I've started a new job at the OECD. It is a fantastic workplace with brilliant people and high level discussions of telecommunications policy. It also means that no matter what I blog about in the telecom industry, it is directly related to my work. The consequence of that is, that it will be very hard to separate the two and so I've decided to put the blog on semi-hiatus for the time that I work in Paris. I may write and Twitter occasionally or link to something I've published elsewhere. So don't delete me from your feed readers and Twitter just yet.

If you are in Paris anytime, send me note. I love to meet people and actually will go and have dinner tonight with someone I met through the blog!

Celebrating Bill St. Arnaud and his wireless vision

noreply@blogger.com (Rudolf van der Berg) — Mon, 03 Jan 2011 22:14:00 +0000

James Enck used to call the readers of his blog "Eurotelcoblog", super mega über valued readers or something of the like, when they provided him with interesting tidbits and insight. In Bill St. Arnaud's case the list of superlatives that I would need to give him in thanks to his efforts would be a bit too long and not do him justice. Ever since I published about how opening up the market for MVNO's to end-users, Bill has been a major supporter of the idea, often blogging about it. Today he wrote such a great post on what universities and research networks could do with the idea, that I have to give him his own article, anything else would just be too little.

If you want to know where the future of innovation in networks is, then read Bill's blog of today. It's very enlightening, about how networks could be build that seamlessly integrate traditional mobile and wifi networks can be integrated into one large network, that could support many innovations.

Without readers and supporters, this blog wouldn't be anywhere so that's why I'm celebrating Bill today!

The truth about 3 digit Mobile Network Codes (MNCs) E.212

noreply@blogger.com (Rudolf van der Berg) — Wed, 29 Dec 2010 22:51:00 +0000

Despite the title, I don't know the full truth about Mobile Network Codes and whether they can be 2 digit or 3 digits long. However, I'm getting an ever bigger suspicion no one really does. So after talking to experts from various telco's and standardisation organisations for more than a year and just before I head into a new world, I thought I write down everything I've learned about Mobile Network Codes as defined in ITU recommendation E.212. This is not all there is to know and I hope others will chime in. There is so much misinformation and Fear, Uncertainty and Doubt, that it is hurting the development of the industry as a whole. I do have the sincere feeling, I'm the first person in a decade to fundamentally look at these numbers and the impact they have on the telecommunications market. I haven't seen any studies by any other researcher looking at these numbers at all. (Please prove me wrong, I would so love to see some other nutcase who looks at this stuff in-depth, preferably one who is way better than me and teaches me something new).

You're a policy maker and you need to know what to do

Before dismissing the potentially billions of euros of unrealized potential in the market and not changing a damn thing to the way MNC's are assigned, please ask a very very good and honest researcher to look into this. As far as I know there is no reason to be against this plan. But I am one person. I'm often right, but I may not be. After a year of searching though nobody I met who has serious mobile network knowledge has said I'm seriously bonkers and it cannot be done. Here is what I know. Don't be afraid to be critical. Do know, no one is paying me for this position.

A Mobile Network Code is digits 4 to 6 of a 15 digit IMSI.
As you may know, your SIM-card carries a unique number, the 15 digit IMSI number. (there are more unique numbers for an IMSI, but this is the important one). This number is used in the network to identify the mobile subscriber. Based on the IMSI the network knows whether the SIM should be allowed access, the type of access it should be granted and where to route the bits too.

The pre-paid SIM-card from the Dutch supermarket MVNO Albert Heijn I bought yesterday as a temporary replacement has IMSI: 204122030140166. Decoding it leads to the following knowledge:

204 is the Mobile Country Code of the Netherlands. Why it is necessary for mobile networks to know countries, I don't know. But it is in the spec. There are some operators like Digicel, who use one MCC 338 for their whole network in many different countries, deeming it too hard to work with an MCC for every Caribean Island and technically speaking they are right. Vodafone could easily use on MCC for all nations it operates in. Nothing breaks if you use an Italian MCC in Vatican city. Though it somehow is against best practices and the ITU doesn't like you doing it and wants operators and regulators to report extra territorial use of MCC's and MNC's
12 is the Mobile Network Code of Telfort. This used to be an indepent MNO with its own mobile network. It was bought by KPN, who switched of Telforts network and now Telfort is an MVNO on KPN's network. MNC's are issued by national regulators. In Vatican City this may well be Radio Vatican or the pope himself. The combination of MCC and MNC uniquely identifies the network. The E.212 spec says that the number can be 2 or 3 digits long. In Europe it is two digits long. In the rest of the world Wikipedia seems to show 3 digits are used as well. For instance Honduras, Columbia and the United States.
2030140166 is the Mobile Subscriber Identitification Number: In this case it is 10 digits long, per the standard it could also be 9 digits long. That gives every mobile operator at least a billion and up to 10 billion numbers. Though it could be issued completely at random, it does seem there is further logic in the number. For instance I heard that the first x digits are sometimes used to point at a specific Home Location Register, which keeps track of where SIM-cards are in the network. It seems HLR's can only accomodate x million devices per HLR. If you expect more than x million devices, using the first digits allows for easy internal routing of traffic. Furthermore specific MVNO's, like my supermarket MVNO may get their own range, so maybe 2030 is unique to Albert Heijn. Again, if they grow, it becomes easier to shift them to different or dedicated HLR

MCC + MNC are also broadcasted by the network base stations.

Apart from being used on the SIM-card, the MCC+MNC is also broadcast by the networks base stations. Your mobile phone receives these announcements and will select these networks either based on a list of preferred networks that it has listed on the SIM or by just trying all these networks. The number broadcasted should always be 5 digits I heard. This has something to do with base stations, for instance those used for CDMA. It is hard to get good information on this, but it seems to be a hard restriction for some networks. It may be more than just legacy in the core network, it may be that this is hardcoded into end-user devices too. So it may be that your phone only expects to see a 5 digit number broadcast from a base station and will not respond to a 6 digit one. What this doesn't mean is that your device only logs on to a network that has the same MCC+MNC combination as the one broadcasted. Again, take my phone. The SIM uses 20412 as MCC+MNC, but it doesn't latch on to Telfort's network anymore. That network doesn't exist anymore instead it chooses to work on 20408 or 20410, which are KPN's. It shows this in the home screen. In a foreign country, a phone is very promiscuous and latches on to whatever network has the strongest signal, that is willing to give it access.

Proposition to let large scale end-users manage own Mobile Network Codes.

As said in previous posts, I propose to open up access to MNC's to anyone wanting to become an MVNO for instance for their smart metering project. The benefits are enormous. It would allow:

- easier switching of mobile operators, without having to physically switch the SIM-cards of one operator for another

- better coverage and less black spots through national roaming

- cheaper international roaming through more competition

- more innovation, because it is the end-user who is in charge.

This is scary stuff.. or to quote a telco guy it is quite extreme, controversial and maybe can not or should not be implemented. The main reason why I think it could work is that no argument is ever made against the 4 points as being something a customer needs. Even better, the 3GPP is researching solutions for these problems, except it only looks at technical solutions and these technical solutions fail because of the administrative overhead they require and the fundamental breach of GSM security models. The proposition looks at changing the rules under which someone can get a number, instead of trying to change numbers post-hoc. Every argument against it seems to focus on how hard it could be and how this is untested.

The proposition isn't untested. All I'm proposing is roaming

What everyone seems to forget is that mobile roaming is doing exactly what I'm proposing. Every telco has hundreds of roaming agreements. Every telco allows hundred of thousands if not millions of devices on its network every day, with whom it has absolutely no relationship what so ever. The only thing the network is interested in is; what network does this SIM belong to. Do we have a roaming agreement with it. If yes, grant access and start billing. If no, deny it access.

The arguments agains doing this are mostly based on FUD.

A well known tactic in the IT industry to discredit any product or idea is to use Fear, Uncertainty and Doubt. The proposition discussed here has seen its fair share. The work I did on it was mostly aimed at finding out if there was any basis for this FUD, so that it was valid criticism. The easiest to dispel are arguments about:

Demand: Critics say customers don't want this, we never heard them ask for it. First of all, for a regulator this should be irrelevant. Limiting supplies to a resource shouldn't be about whether or not someone wants it or not, but about whether they could get it if they wanted it and whether or not that would be a problem. And yes there is demand in the market, quite a lot actually. The Dutch Defense department already has its own MNC and is very happy with it. (but they have guns, so its a different request)
Operational difficulty: Critics say that running a telco is hard. Customers can't be trusted to run HLRs etc. This is true, operations are hard. That is why many telcos outsource their operations to companies like Ericsson. Yes you heard it right. Ericsson runs many networks totalling well over 750 million active customers. So why can't an end-user hire Ericsson, Huawei or a smaller company to do the same for it.
It is too expensive: Not what I heard. Seems you can be done with a figure in the low hundred thousands euro for a full set up. Some speculate it could become less if the market becomes bigger.
It will become too big and need too many numbers. I tried to find out the numbers and it seems most people I speak to intuitively think the amount of interested companies is too big for this to work. However, just looking at the statistics for the Netherlands, out of 844,000 companies in The Netherlands only 15,000 have over 50 employees, and only 7850 of those over a 100 employees. Most of these companies aren't interested in these new ideas. They are accountants, schools, NGO's etc. I guessed that for the Netherlands the upper limit would be 1500, but a realistic number for the next 15 years is closer to 200 or even lower. The Netherlands might run out of MNC's, but the whole of Europe, with dinky toy countries like Luxembourg, Andorra, Liechtensteyn, Jersey, Guernsey etc. could get very far. And really, if I could get 50,000 end-user MVNO's online in the next three decades, with this idea, I would be ecstatic. Just imagine the competition and the benefits. But to counter this and to not get into prolonged negotiations in the ITU for nations to reform the way MCC's are used, I propose to move to 3 digit MNC's and according to some, this will unleash network Armageddon, shatter the space-time fabric and as a result destroy the entire functioning of the mobile network. This last one is rather nasty and based on a GSM Europe letter from the year 2000, that I will delve a bit deeper into.

Moving to three digits will not lead to mobile network Armageddon

In the past Europe and the USA have discussed moving to three digit mobile network codes. This was always faced with heavy opposition from the GSM Association and its predecessor. Though the people at the GSMAs central office are very nice, I do have some misgivings about its perspectives on public policy and the functioning of the mobile market. Mobile roaming or Mobile termination access anyone? ;-)

The argument is outlined in a letter from 2004 which is available on the ATIS-website. The nice thing is that it is a letter which has enabled track changes. Running it through Google docs and looking at the HTML, the entire text is below the fold. It turns out the original letter is one sent to CEPT/ECTRA in 2000. And yes the letter is still seen by some as valid as I've seen these points pop up in discussions and even saw some points quoted verbatim in mails. There is also a message from the 3GPP on the topic, which provides a bit of background. However it was written after the letter to CEPT/ECTRA and it quotes those points directly, so I'm not thinking it was written indepently.

The argument is:

Call back of the released 2-digit SIM card – this could affect all current GSM users (around 500 million in Europe).
The handset display may not be ready for such a change – this could cause severe confusion with all existing GSM handsets.
Change of all existing roaming agreements with the need to re-run all the performed tests.
Enormous efforts in terms of investment and manpower in order to change network elements and billing systems, thereby reducing the innovation power of mobile operators.
Inability for many operators to roam with the second operator where 3-digit MNCs have been issued in a country with no ability to identify an operator by the first 2 digits.

None of these arguments is true. They can be debunked I think. (If you feel otherwise, please tell me so)

Callback of all SIM-cards isn't necessary. The proposal isn't to split all existing two digit MNC's in 10 3 digit MNC's and hand these out to different organisations. So the plan isn't to split 20412 into 20412X and to let Telfort only keep 204120 and give 204121 to a smart metering project. Nope, what has been issued has been issued. So none of the SIM-cards have to be revoked.
The handset display has developed massively in the last 10 years. But even then. We aren't talking about changing existing networks. It might even be necessary for backwards compatibility to let those that run networks still use 2 digit MNC's by effectively giving them a block of 10 consecutive 3 digit MNC's. Nothing says this can't be done.
Roaming agreements only need to be tested for new networks using the three digit MNC, as all existing networks will remain doing the same thing they did before. So no changes there. Furthermore roaming agreements are tested regularly. Things always change in networks. Testing is essential to this.
An enormous effort in network elements and billing systems. This may be true, but only for those networks that didn't comply with the standard. According to my sources, some networks assumed standards compliance in their networks and needed to identify more digits then 6 to find MVNO's and HLR's anyways, so they are ready for any length of an MNC. Furthermore, that is business. Things change. Windows versions get upgraded and loose backwards compatibility. If this was a real argument, cars still wouldn't have safety belts.
Inability to roam seems rather unlikely. The 15 digit IMSI is what is used. That doesn't change by looking at either the 5th or the 6th digit. And as said, it isn't necessary to change the base stations. If they need 5 digits, that can still be accomodated.

Loose end: Global Title Translation

Realizing I don't know everything. Do know I see Global Title Translation as a loose end. I don't know precisely how it works, but it feels from the descriptions that there shouldn't be a fundamental difficulty there either. It may however require some work.

Update Loose end: How it is processed "on the air"
Good point by my friend TT. A serious question is how an MNC is sent over the air. Is it that it is sent as part of an IMSI string and indistinguishable from the rest of the IMSI, if so the network wouldn't care about the length of the MCC+MNC bit. If not, the length may be of interest. He found in some ETSI documentation that the IMSI is broken down into separate parts of MCC+MNC+MSIN. The MNC is sent separately over the air as 3 digits. That is nice, because this means that three digits should be understood by the network as it gets a 3 digit field that it needs to deal with. Will investigate further.

So what to do

GSM Association

Mr. Alain Doisneau

Chairman Project Team Numbering

Subject: Mobile Network Codes (MNCs) – Change from 2digits to 3 digits

Dear Mr. Doisneau,

GSM Europe has recently been following the developments concerning CEPT/ECTRA’s intention to formally adopt a Decision on the expansion of Mobile Network Codes from 2 to 3 digits (draft Decision attached).

GSM Europe is the European regional interest group of the worldwide GSM Association, the premier global body behind the world’s leading wireless communications standard. GSM Europe represents around 159 mobileoperators in 50 regions in Europe with around 200 million subscribers.

I understand that the adoption of ECTRA’s Decision, which was originally scheduled for the CEPT/ECTRA meeting on 21st June 2000, was postponed further to concerns raised by industry. Further to its preliminary analysis, GSM Europe shares such concerns and believes a change in MNC from 2 to 3 digits is most likely to have severe implications for almost the entire GSM system (See Annex). The main consequences from our point of view would be:

¨ Call back of the released 2-digit SIM card – this could affect all current GSM users (around 200 million in Europe)

¨ The handset display may not be ready for such a change – this could cause severe confusion with all existing GSM handset

¨ Change of all existing roaming agreements with the need to re-run all the performed tests

¨ Enormous efforts in terms of investment and manpower for the changes in the network elements and the billing system, thereby reducing the innovation power of mobile operators

GSM is not convinced of the urgend need to initiate those changes. It does not feel that the costs and risks to be incorporated by the operators only are in any way justified by potential advantages. Especially new operators, operators which a dense infrastructure to provide special home services and operators building up a UMTS network could face more problems for migration than others.

In light of the abovementioned issues, GSM Europe strongly supports the idea of having a workshop open to all parties concerned in order to identify the technical difficulties and assess if and when these problems can be overcome. Such discussions, it is felt, would bring some clarifications to CEPT/ECTRA and other parties concerned on the real implications of an alteration in the MNC digits number. With regards to the timeframe, it would seem appropriate to hold a workshop organized by ETO in September 2000, so that adequate measures can be taken in a timely manner.

You will find attached in annex an analysis paper drafted by our experts expressing the views of GSM operators. The paper seeks to outline the technical issues relating to the introduction of 3-digit Mobile NetworkCodes and the possible consequences associated to this change.

I hope the attached document will be of interest to you.

Yours sincerely,

……..

Copie: Marco Bernardi

Alisdair McLead

Leo Koolen

Annex

Technical Issues Relating to the introduction of Three Digit Mobile Network Codes (MNCs).

Introduction / Background

The MNC is a component of the IMSI (International Mobile Subscriber Identity), the other components been the Mobile Country Code (MCC) and the Mobile Station Identification Number (MSIN).

Figure 1 IMSI Structure

The MNC is an inner element in the IMSI with a defined length of 15 digits. The whole IMSI is marked on the SIM-card. To this extend, the IMSI must not be mismatched with the subscriber's number which is e.g. in GSM an E.164 number.

Various GSM entities will interpret the extra MNC digit as part of another field (MSIN), which will consequently lead to inconsistencies within the interaction between various parts of the whole GSM System.

The IMSI is used to:

· Identify a (roaming) customer, for

· Network internal purposes used in all signalling in the PLMN, in

· Interaction with the Mobile Terminal and for

· BackOffice applications like charging, billing and accounting

As the National Regulatory Authorities (NRAs) are responsible for the establishment and publication of conventions for IMSIs (IMSI = MCC + MNC + MSIN), the change of the current MNC allocations (e.g. 03 may become 030 or 003) may lead to severe problems as the length of the IMSI is fixed and the structure / usage of the MSIN is subject to each individual operator.

It should be noted that for the time being, out of a total of 300 MNCs actually in use, 14 are 3 digit MNCs. Furthermore, there are no assignments for 3 digit MNCs outside the US, where 5 US Operators where assigned such MNCs. In terms of operators this corresponds to 1.7 % of the GSM Operators.

It should be noted that this paper does not claim to be exhaustive and further in depth analysis is required.

Technical implications for the introduction of 3-digit MCNs:

SIM Cards:

The IMSI is an inherent part of all SIM cards,. This would hence require a replacement of all existing SIM cards with new ones. Taking into account the number of mobile users and prepaid cards, such exercise is practically not feasible (e.g. in Germany about 35 Millions of SIM Cards are produced, or are submitted to the customers, with a rapidly increasing proportion of prepaid cards). The costs of this procedure are expected to be quite high due to production of new SIM cards, SIM card distribution and logistics, customer service, activation of the new SIM cards, and customer handling. The activation/provisioning part raises questions about the capacity of the IT and networks systems. It would be quite likely that most operators would be forced to build up additional capacity to be able to manage the migration. Some customers would experience problems and would temporarily unable to use their mobile, which means loss of revenue from the operator's point of view. In addition to this it is doubtful, that SIM card chips can be produced to serve all carriers in due time.

As a result the co-existence of 2 and 3 digit MNCs can only be the subject of further analysis.

Handsets:

The introduction of a 3 digit MNC will certainly result in a number of compatibility problems between existing mobile equipment and the SIM, and also very likely between MS and the network. With respect to compatibility problems in the interface between SIM and the software of ME, two cases can easily be identified:

1 SIM with a new IMSI structure built up by three digit MNC combined with an ME running old software that isn't prepared for the new IMSI structure. The ME will in this case attempt to read the extra MNC digit as part of another field. This case will be difficult to solve as there are about 250 Million legacy mobiles in operation throughout Europe.

2 SIM with an old IMSI structure build up by two digits MNC combined with a ME running new software, expecting the new IMSI structure. Here it is much easier to find a solution, but the problem has to be addressed in order to find some kind of IMSI type identification on the SIM.

In other words, either a large number of terminals would have to be replaced, a very long lead time (longer than the typical life span of a mobile terminal) would have to be allowed, or the problems in the transition phase would have to be accepted by users and operators.(Further Input is welcome by other GSMA groups !!!!)

Network:

The IMSI analysis would need to be redefined .

All subscriber records would need to be modified to change the MNC from 0X to 0X0 or 00X. In the case of multiple HLRs, as is the case with nearly all operators, the level of planning and coordination would be substantial.

IN the case of redundant HLRs redundancy the data would need to be changed on both the live and backup locations.

In addition to this, many of the operators would have to reconfigure their Base Station Subsystem so that it transmits a LAI which contains the new 3 digit MNC. This causes operational costs. In addition it would demand additional operational capacities, which especially operators, who are still in the phase of 2G network rollout or who have (some additionally) plans to built up a 3G network may not be manage in a sufficient way, so that customers will not be frustrated.

If FNR was in service it would also need updating.

Billing & Customer Care Systems:

The billing systems of most networks use the IMSI for the generation and gathering of billing information. A change in the MNC would require severe modification to billing systems and protocols. All effected subscriber records would need to be modified TAP-Files (TAP-Incollect/Outcollect) TAP 2 / TAP 2+.

Roaming:

Regarding Roaming, the IMSI analysis for all European operators (e.g. approximately 100 in Eircell’s case) will have to be changed in every MSC.

All roaming testing would have to be redone and would be more complex.

Finally, operators would have to change their roaming contracts with each of their roaming partners. This will have an effect on almost countries using GSM.

Hence, if the roamed-to network cannot scrutinise 3-digit MNCs , roaming would stop overnight

Fraud Prevention / Management:

The impact from a Fraud & Security point of view would be the cost of modifying Fraud Management Systems and the length of time it would take to run the additional reports required.

(Further Input is welcome by other GSMA groups on specific issues relevant to their groups!!!!)

The GSM Association (GSMA) has recently been following the developments on the expansion of Mobile Network Codes (MNCs) from 2 to 3.

The GSMA, as the premier global body behind the world’s leading wireless communications standard, protects and enhances the interests of 678 GSM mobile operators from 209 countries and territories with more than 1 billion customers throughout the world today.

The GSMA has concerns related to the allocation of MNCs and its change in some cases from 2 digits to 3 digits and would like governments and regulatory authorities to take these concerns into account when allocating numbers to operators.

European operators have been issued 2-digit MNCs by their regulators for many years, therefore they have built their systems/infrastructure to support 2-digit MNCs.

In the past only North America and a few other countries have used 3-digit MNCs. These 3-digit MNCs always ended in '0' (zero) and therefore could be treated as 2-digit MNCs. The '0' (zero) is dropped/ignored and operators are able to uniquely identify their roaming partner based on the first two digits of the MNC.

The GSMA is now aware of countries where multiple 3-digit MNCs have been issued by regulators where the first two digits are exactly the same. Although according to the specifications this is acceptable it is a problem for many operators whose systems are only set up to accept 2-digit MNCs.

Enlarging the length of the MNC to accommodate 3-digit MNCs (filename length, field and record length) is a major change to the entire systems such as the rating engine, the TAP engine, the billing system and the data warehouse.

Further to its preliminary analysis, the GSMA shares such concerns and believes a change in MNC from 2 to 3 digits is most likely to have severe implications for almost the entire GSM system (please see Annex). The main consequences from our point of view would be:

Ø Call back of the released 2-digit SIM card – this could affect all current GSM users (around 500 million in Europe).

Ø The handset display may not be ready for such a change – this could cause severe confusion with all existing GSM handsets.

Ø Change of all existing roaming agreements with the need to re-run all the performed tests.

Ø Enormous efforts in terms of investment and manpower in order to change network elements and billing systems, thereby reducing the innovation power of mobile operators.

Ø Inability for many operators to roam with the second operator where 3-digit MNCs have been issued in a country with no ability to identify an operator by the first 2 digits.

Therefore we strongly encourage governments and regulators to undertake a cost/benefit analysis when allocating MNCs. We would suggest allocating either 2-digit MNCs or alternatively 3-digit MNCs where the operator can be uniquely identified by the first 2 digits.

You will find attached an analysis paper expressing the views of the GSMA operators. The paper seeks to outline the technical issues relating to the introduction of 3-digit MNCs and the possible consequences associated to this change.

I hope the attached document will be of interest to you.

Yours sincerely,

Tom Phillips

Government & Regulatory Affairs Officer

GSM Association

For any questions, please feel free to contact Isabelle Mauro, Programme Director, GSM Association: Tel: +44 207 518 0548; Fax: +44 207 518 0531; E-mail: imauro@gsm.org

Annex

Technical Issues Relating to the Introduction of

Three Digit Mobile Network Codes (MNCs).

1 Introduction / Background

According to IMSI structure defined in ITU E.212 an MNC can have two or three digits.

The IMSI structure and format are as shown in Figure 1.

(Figure 1 IMSI Structure)

The IMSI is used to identify a (roaming) customer, for network internal purposes used in all signalling in the PLMN, in interaction with the Mobile Terminal and for BackOffice applications like charging, billing and accounting.

As the National Regulatory Authorities (NRAs) are responsible for the establishment and publication of conventions for IMSIs (IMSI = MCC + MNC + MSIN), the change of the current MNC allocations (e.g. 03 may become 030 or 003) may lead to severe problems given that the length of the IMSI is fixed and the structure / usage of the MSIN is subject to each individual operator.

It should be noted that for the time being, out of a total of 300 MNCs actually in use, 14 are 3-digit MNCs. Furthermore, there are no assignments for 3-digit MNCs outside the US, where 5 US Operators where assigned such MNCs. In terms of operators this corresponds to 1.7 % of the GSM Operators.

2 Technical implications derived from the introduction of 3-digit MCNs

Below we find some of the technical applications derived from the introduction of 3-digit MCNs. However this paper does not in any case try to be exhaustive and further in depth analysis is required.

2.1 SIM Cards

The IMSI is an inherent part of all SIM cards. This would hence require a replacement of all existing SIM cards with new ones. Taking into account the number of mobile users and prepaid cards, such exercise is practically not feasible (e.g. in Germany about 35 million of SIM Cards are produced, or submitted to the customers, with a rapidly increasing proportion of prepaid cards). The costs of this procedure are expected to be quite high due to necessary production of new SIM cards, SIM card distribution and logistics, customer service, activation of the new SIM cards, and customer handling. The activation/provisioning part raises questions about the capacity of the IT and networks systems. It would be quite likely that most operators would be forced to build up additional capacity to be able to manage the migration. Some customers would experience problems and would temporarily be unable to use their mobile, which means loss of revenue from the operator's point of view. In addition to this, it is doubtful that SIM card chips, which are already due to the strong demand in short supply, can be produced to serve all carriers in due time.

As a result of this, the co-existence of 2 and 3-digit MNCs can only be the subject of further analysis.

2.2 Handsets

The introduction of a 3-digit MNC will certainly result in a number of compatibility problems between existing mobile equipment and the SIM, and also very likely between the mobile station (MS) and the network. With respect to compatibility problems in the interface between SIM and the software of mobile equipment (ME), two cases can easily be identified:

1) SIM with a new IMSI structure built up by three digits MNC combined with an ME running old software that isn't prepared for the new IMSI structure. The ME will in this case attempt to read the extra MNC digit as part of another field. This case will be difficult to solve, as there are about 250 Million legacy mobiles in operation throughout Europe.

2) SIM with an old IMSI structure build up by two digits MNC combined with a ME running new software, expecting the new IMSI structure. Here it is much easier to find a solution, but the problem has to be addressed in order to find some kind of IMSI type identification on the SIM.

In any case a large number of terminals would have to be replaced and very long lead time (longer than the typical life span of a mobile terminal) would have to be allowed.

2.3 Network

The IMSI analysis would need to be redefined. All subscriber records would need to be modified to change the MNC from 0X to 0X0 or 00X. In the case of multiple Home Location Registers (HLRs), as is the case with nearly all operators, the level of planning and coordination would be substantial.

In the case of redundant HLRs, the data would need to be changed on both the live and backup locations.

In addition to this, many of the operators would have to reconfigure their base station subsystem so that it transmits a Location Area Identity (LAI) which contains the new 3-digit MNC.

2.4 Billing & Customer Care Systems

The billing systems of most networks use the IMSI for the generation and gathering of billing information. A change in the MNC would require severe modifications to billing systems and protocols. All effected subscriber records would need to be modified TAP-Files (TAP-Incollect/Outcollect) TAP 2 / TAP 2+/TAP3.

2.5 Roaming

Regarding roaming, the IMSI analysis for all European operators will have to be changed in every MSC. All roaming testing would have to be redone and would be more complex.

Finally, operators would have to change their roaming contracts with each of their roaming partners. This will have an effect on nearly every country using GSM. Hence, if the roamed-to network cannot scrutinise 3-digit MNCs, roaming would stop overnight.

2.6 Fraud Prevention / Management

The IMSI is a data element commonly used by the operator community to detect fraud. To change the format of the IMSI would render most detection and prevention useless thereby exposing networks to increased financial loss.

Historical searches for IMSIs spanning the transition will have to be split into two reports, one before and one after search. Usage variation alarms for all the numbers that change will effectively start from scratch. Similarly, any IMSIs that had previously been hot-listed for the purposes of fraud monitoring would need to be replaced.

One impact, from a fraud and security point of view, would be the cost of modifying fraud management systems that are currently used by home networks to monitor for fraudulent calling patterns based on individual IMSIs. It also likely that the length of time it would take to run the additional reports required would increase.

GSM operators currently have mechanisms in place that allow for the exchange of IMSIs, and other data, which may identify fraud. Such exchanges can be transacted electronically through the use of near real time data exchange systems across the SS7 signalling network. Such systems would need to be upgraded to accommodate changed MNCs.

Leaving Logica to go and work for the OECD in Paris

noreply@blogger.com (Rudolf van der Berg) — Fri, 24 Dec 2010 05:52:00 +0000

It is with mixed emotions that I’m telling you that I’m leaving Logica to go to work for the Organisation for Economic Co-operation and development in Paris as an economist/policy analyst in the information, communications and consumer policy division. Logica has given me great opportunities to work in the fields of telecommunications and utilities. I had the pleasure to work on a variety of projects ranging from shaping the Belgian energy sector, to smart metering and Machine to Machine communications and the future of ENUM in the Netherlands.

I’ll start at the OECD on January 3^rd, initially for a period of 13 months. I will work on the high level meeting on the future of the internet economy, which will be held in June, and on other subjects regarding telecommunications policy. It’s an exciting step. I don't know yet how this will affect this blog as I don't know the OECD's approach to new media yet. It might be that the blog and twitter will go on hiatus.

I still hope to write two blogposts:

1. everything I know about 3 digit MNC's. There is hardly any information available about it and what is there is mostly misinformation

2. How we should get rid of the difference between public and private networks in most of our telecommunications law.

And I'm tempted to write a third as a reaction to the Berec mobile roaming consultation

So many ideas, so little time.

Wrote an article for Gigaom on Apple and SIMs

noreply@blogger.com (Rudolf van der Berg) — Sun, 21 Nov 2010 12:07:00 +0000

Gigaom just published an article by me called: "How to bypass carriers Apple style", on what Apple could do with the SIM-idea I developed for M2M. Apple is only in there because of the marketing value, Samsung, Sony, Philips, LG, Asus, all of them could be doing this. It works best for laptops and other data devices. These devices often don't have data roaming enabled, even if you use a dongle.

Have fun reading it!

GIGAOM Readers: please have a look at some of the other articles I wrote on the topic:

A very simple presentation to introduce the idea given at the Telco2 conference (click or scroll down)
A more in depth management summary of the idea, focussed on embedded M2M
There is also a 55 page document for the government of the Netherlands (in Dutch)

Other things I wrote:

The common errors of Telecom CEOs

a primer on peering and transit

Telecom Cool Wall

Why all telecom marketing and product management is wrong

The GSMA changes opinion on programmable SIM-cards

noreply@blogger.com (Rudolf van der Berg) — Thu, 18 Nov 2010 14:49:00 +0000

In a surprise move, the GSMA today announced that it will standardize a new form of embedded SIM. This will allow remote activation and updates. The spec will be ready in January 2011. No word yet on whether or not the IMSI will be changeable after initial activation. This is however a complete reversal from the GSMA's position from the past:

Appendix A of 3GPP TR 33.812 V9.2.0 lists the opinion of the GSMA as
[…]Furthermore, one of the major concerns of MNOs is the potential weakening of the well-established and trusted SIM-based GSM/3G security architecture. Extended OTA (any kind and via any bearer of over the air data download to the USIM) capability to facilitate download of new subscriber keys and possibly authentication algorithms represents such a potential weakening of security.[...] not allowing MNOs to fulfil their obligations towards regulatory and other governmental authorities to guarantee secure authentication and billing.

[…]The only proposal in the TR that was acceptable to GSMA SG representatives was the Alternative 2 where operator change was performed by physical replacement of the UICC in the device.[…]

This would allow the likes of Apple to release mobile phones without removable SIM-cards. The biggest question is of course, why the GSMA changed its position and how much control they are willing to relinquish to the end-user, be they Apple or a smart metering company.

Though this will alleviate the problems I identified in my work on Machine to Machine communication, it's only 1 out of the 4 problems that I identified that now could be fixed (seeing is believing). Customers may be able to switch operators in the future, but roaming, national roaming and innovation, will probably still be limited.

Your customer wants to be mvno v2

View more presentations from Raindeer.

My presentation on M2M at Telco2.0 and my impression of the conference

noreply@blogger.com (Rudolf van der Berg) — Thu, 11 Nov 2010 22:48:00 +0000

Tuesday and Wednesday I was at the Telco 2.0 conference in London organised by STL. I was an invited speaker on the M2M track on Wednesday morning. Like many of you, I know the Telco 2.0 blog and knew of the conferences. Good friends in this community, like James Enck were on stage there before me. Other friends in this community stood at the basis of the Telco 2.0 ideas. So I was quite happy to be invited to the event.

Telco 2.0 on M2M
My presentation was part of the dedicated M2M track. Not all the attendants were there, but only those interested in M2M. The crowd was treated with a series of presentation from carriers and various forms of system integrators and suppliers.

What I found interesting in the telco presentations was, what they were and weren't about. What they weren't about is what their customers were doing and the problems that these customers were facing. What they were about is, potential use cases, the organisation of the telco's M2M unit and how it was embedded in the organisation and the design of the M2M platform. The whole reasoning was supply driven. Build it and they will have to come. As someone had warned a while ago, most telco diagrams of a problem start with their network in the upper left part of the page and the customer in the lower right part of the page. This is true. It also shows where their focus is; themselves, not their customer.

The difference between the telco's wasn't too great. Orange seems to go for a more fully integrated solution, where they will deliver the network, the platform and all sorts of other devices. Telenor has a dedicated group for M2M networking solutions, called Connexion and another group Objects that focus on service enablement, regardless of the network. So Connexion can sell to the whole market without involving Objects and vice versa.

Most of the presentations were at least of a decent quality and sometimes quite better. It deserves metntioning, because often this is different. However, the feeling these presentations give me in retrospect, is that of enormous budgets and very little idea of what the customer is looking for. This idea came from a remark by a quite brilliant and quiet Eastern European financial controller of a mobile telco, who remarked privately that his company just didn't have the size and the budgets necessary for such a platform, certainly given the 5 euro max revenue/month per SIM. This remark is so incredibly true. The large behemoths of European Telecoms are still swimming in the free cash flow. Free cash flow that is supposed to do something and might as well be spend on M2M. Running after every hype in the business, including M2M is not so much a necessity as it is a luxury.

The interactive format of Telco2.0 also means there was quite some feedback from the room, through instant polls. What it showed was that the audience was effectively split over the idea whether mobile telcos have a larger role to play in the M2M ecosystem.

My presentation on how M2M customers becoming MVNO's
I gave a presentation called "Your M2M customer wants to be MVNO". In generally the audience quite liked it. They thought it was thought provoking. I do think it was the most referenced presentation during the day. Both the facilitator and the speakers quite often referred to Rudolf or Logica. This was of course nice for my ego. I really focussed on the 4 central business problems and why becoming an MVNO was a solution.

What was also quite clear is that the industry really doesn't want to hear what I had to tell. The idea that a large scale M2M user is going to become a wholesale customer of the telco is just not done. There were some interesting comments why not:

Your customer wants to be mvno v2

View more presentations from Raindeer.

It is all about money and prices are coming down. If the price is low enough, the customer doesn't need freedom, flexibility and innovation.
It is too complex to run an MVNO (so let an MNE or Ericsson run it. Ericsson already runs most telcos)
Lock-in is everywhere, so why not with a telco? (Because people only accept a lock-in if it is because of network effects, not because of market power)
This cannot happen, because we need to make a return on our platform
This cannot happen, because competition would kill margins and then the industry would be very badly off.

The best argument against my presentation was given to me in private by a large telco. The reason we can't have private M2M MVNO's now, is because otherwise all the budgets of telco M2M plans would be killed by the executive board. Private M2M MVNO's could happen in three years, when the initial investments in M2M platforms have been done.

There were also some quite positive reactions to my presentation. Certainly those that represented large scale users (they were scarce but present) were very positive about the content of the presentation.

My impression of the Telco 2.0 conference
Last year november I spoke at Ecomm on how all telecom marketing and product management is wrong. Now I spoke about a different but also controversial subject. What was interesting was the difference in audience and their reactions. The Ecomm audience is you, hip and happening. The cluelevel of the conference is high. Unfortunately most of those attending are not of such a level yet, that they can actually implement their vision. What a difference with the Telco2.0 conference. The power level of the attendants is high. Proper CTO's, Vice Presidents, Heads. Those that can say: "Make it so".

The difference shows up in the way the problems are discussed. For the Ecomm crowd a problem is a problem. It is something that needs to be solved and removed. For the Telco2 crowd a problem is either a business opportunity or Somebody Elses Problem. And with the amount of money sloshing around in the industry, almost every problem is attacked to see if it can generate more money. If it cannot generate money, it's somebody elses problem.

The harsh reality is that those high level execs attending are grasping at every straw to see if it will generate some yield. This was clear from the discussion of clouds. Every major telco thought they were going to be big and uniquely positioned for clouds. After a while I asked out of desperation, how the assembled telcos thought to solve a very basic problem, namely getting the right people to build their cloud. You see, Amazon and Google are now sponsoring university courses in cloud computing, because finding the right people is hard. I didn't see telcos finding a person like Vijay Gill or Werner Vogels, let alone a team of them. This was well answered by Oracle, who came with a promise, that they would provide the products that could be a building block for a generic telco cloud. I personally would more expect Oracle to build it's own cloud, which it would then resell through telcos, But who am I ;-)c

And that brings me to my last bit. Who was I at the conference? I probably was a bit the court jester. The outspoken guy who challenges firm beliefs of the industry. It wasn't just clouds or M2M that I challenged the current opinions. I also got in a small verbal fight with a network vendor trying to push QoS as a solution. I've never been able to understand why anyone would use the argument: You're only right on the past 15 years, the next 3 years will prove that QoS is necessary for wireless. I was the weird guy who promoted happy pipes and mentioned Pareto distributions as the reason why we have bandwidth hogs. I don't know whether that makes me a good consultant, or just a good analyst. A good consultant seems to profit from his customer's mistake. An analyst just gives the problems and solutions as is. Keeping that in mind, I've agreed to become an analyst at the OECD in Paris in the coming year, working on M2M, peering and transit and some other stuff.

All in all great conference. Thanks to STL for having me and I hope to be back in the future.

Internet Thought

Vanished from the index of Google, Bing and Linkedin a rebuttal of an anti-net neutrality paper

Debunking an anti-netneutrality study by four economists

Examples of how the article has become unfindable

Conclusion

Rebuttal of “Net neutrality and high-speed broadband networks: evidence from OECD countries”, by Briglauer, Cambini, Gugler & Stocker

Updated: Request for retraction of “Net neutrality and high-speed broadband networks: evidence from OECD countries”, by Wolfgang Briglauer, Carlo Cambini, Klaus Gugler & Volker Stocker

The e-mail with requests and criticism

Theoretical model doesn’t apply to the empirical situation.

Your academic research doesn’t support mobile mergers? Just say it does!

Prices for consumers will increase.

Investment will not increase more than prices.

New operators increase quality

Conclusion

Is the EU going to make it illegal to put church services on the Internet?

Article 13: the upload filter brings bad luck.

Article 11, just to make life of religious publishers harder

So why didn’t I write this earlier?

I am at the IGF 2014 in Istanbul

Speaking on Connected Television at IIC/Ofcom on May 2nd

Connected TV and Platforms: evolution or revolution?

International Institute of CommunicationsUK Chapter Event

Interview and Conference on The Internet of Things

Presentation on M2M and market failure at Internet of Things-week

Policy Brief on Internet of Things and Transport

Presentation at Telecom Reform Conference 2012

OECD report on "Developments in Mobile Termination"

OECD publishes report on Internet of Things and M2M

Slides available: BEREC expert workshop on IP-Interconnection (Peering and Transit) in cooperation with OECD

Will speak at FTTH Forum Budapest on Nov 9

New presentation on Machine to Machine communication

Born: Tycho Rudolf Willem

I'll be speaking on WIK's conference on Bundling and Multi-Play

First blogpost for the OECD... on International Women's Day (of course)

May the best (wo)man win

This blog is going on (semi-)hiatus

Celebrating Bill St. Arnaud and his wireless vision

The truth about 3 digit Mobile Network Codes (MNCs) E.212

GSM Association

Annex

Technical Issues Relating to the introduction of Three Digit Mobile Network Codes (MNCs).

Introduction / Background

SIM Cards:

Handsets:

Network:

Billing & Customer Care Systems:

Roaming:

Fraud Prevention / Management:

Annex

Technical Issues Relating to the Introduction of

Three Digit Mobile Network Codes (MNCs).

1 Introduction / Background

2.1 SIM Cards

2.2 Handsets

2.3 Network

2.4 Billing & Customer Care Systems

2.5 Roaming

2.6 Fraud Prevention / Management

Leaving Logica to go and work for the OECD in Paris

Wrote an article for Gigaom on Apple and SIMs

The common errors of Telecom CEOs a primer on peering and transit Telecom Cool Wall Why all telecom marketing and product management is wrong

The GSMA changes opinion on programmable SIM-cards

My presentation on M2M at Telco2.0 and my impression of the conference

International Institute of Communications
UK Chapter Event

The common errors of Telecom CEOs

a primer on peering and transit

Telecom Cool Wall

Why all telecom marketing and product management is wrong