OECD on Global Opportunities for Internet Access Development

The OECD published a great interview with its senior economist Dr. Sam Paltridge on Global Opportunities for Internet Access Development. He also wrote a paper on the subject which is a great read. Some snippets that I liked alot

Q.What role do Internet Exchange Points (IXPs) play in the creation of Internet service?
Name withheld

A. IXPs are places where different Internet networks can physically interconnect to send and receive traffic between their networks. Following the commercialization of the Internet they rapidly spread around the world to enable service providers to economically and efficiently exchange traffic locally.

In the absence of an IXP, in any country, local traffic between two service providers will by and large be exchanged internationally. In these cases an email sent from one user to another, in the same country but using different service providers, may be routed via an IXP in New York or Paris rather exchanged domestically. By way of contrast, if that traffic is exchanged locally it can be far less expensive (i.e. avoiding expensive international circuits) and provide better performance for users (e.g. in some countries avoiding satellite circuits with their inherent delays).

Some 90 countries don’t have IXP’s today. There is a map created by the Packet Clearing House which keeps count of countries with and without an IXP: https://prefix.pch.net/applications/ixpdir/summary/
IXPs are relatively inexpensive (e.g. less than USD 40 000 per IXP) to establish and can benefit all stakeholders. For less than USD 4 million each country of the world without an IXP could have one.

Such expenditure, for example, in the form of development co-operation, however, would only make sense if the conditions are in place to enable the IXP to operate efficiently and become industry driven. The main challenge is often is creating awareness of the benefits of IXPs.

and from the report
The next several billion Internet users represent a commercial opportunity rather than a burden and this should be reflected in policy approaches.

Liberalisation of communication markets focuses competitive forces on the expansion of access and affordability for the poor as well as promoting innovation applicable to local circumstances as highlighted by the recent experience of some Asian and African developing countries. Improved access and lower communication costs generate general economic and social benefits. Historically, in markets typified by monopolies and little momentum for access growth, the cost of reaching and maintaining service to low income users was not sustainable.

And on Page 12 there is a explanation of Jipp's Law which used to mean that countries could over and underinvest in telecommunications and which was used to keep investment low and prices high in accordance with the development of the economy in that region.

and on interconnection it sais:

MPP and CPP markets have both exhibited strengths and weaknesses relative to each other in thedevelopment of wireless markets. Both have shown capabilities for innovation and growth. A number of open questions for the future relate to how will both models will deal with convergence with the Internet.Will one model, for example, prove more adept at fostering innovation at the interface with the Internet? How will financial flows between countries be affected by the different termination models? Will one system lead to greater integration in service offerings between countries or networks with the same termination model irrespective of geography? Finally, if termination rates are about compensating networks for costs what explains the vast differences between charges and how will this impact on the competitiveness of countries with high termination rates?

All in all it is a must read on development of networks as it covers many angles on the subject from submarine fibre, to IXP's, to interconnection

KPN, 3 and Play drop data roaming charges. It's cheaper to use VOIP

KPN, 3 and Play have dropped the data roaming charges between their networks to 25 cents/Megabyte. That's great for their customers roaming in the UK, Germany, Belgium, Poland etc. So how many seconds of VoIP does this buy you? A VoIP stream is between 10 kilobit per second and 90Kb/s. This translates to between 90 and 800 seconds. Given that the price of a roaming call on the mobile network has been set at 49 cents per 60 seconds and receiving at 24 cents/minute, it seems that it is now officially cheaper to use VoIP for mobile roaming!

Put differently: any (VoIP like) stream of 266 Kbit/s or 133 Kbit/s is now cheaper than a GSM call of 9600 bit/s.

There must be a catch here. Telco's are never this afraid of mrs. Reding and this is not a competitive market yet!

Reding: Blame it on the National Regulators

Ai Frau Reding has the audicity to blame the national regulators that they haven't brought mobile data roaming tariffs down!

Wasn't the problem that national regulators were incapable of bringing down these rates, because the rates were high at the wholesale level and therefore out of their jurisdiction (a foreign mobile operator was the problem). Wasn't it the case that our friends from the Commission therefore had to set the roaming rates? Wasn't it the same Commission that didn't want to regulate data roaming, but saved it for a later date?

Come on Frau Reding, you could have fixed this last year, don't blame it on the national regulators! Or is this all a scheme to get the Euro Uber Regulator (EECMA) in place?
Even better fix mobile roaming by introducing competition. Give all operators including MVNO's the option to offer access to anybody roaming in the country and require that the home operator is only allowed to charge a small percentage for the billing. For this see my previous entry.

A simple proposal for mobile roaming charges.

Update: this would work great if combined with an iPhone/Android App, that would do the selection of networks for you. Either the app maker or the European commission could have a matrix of all costs charged by host mobile networks for mobiles roaming on their network


The EU has finally succeeded in capping the roaming charges that customers have to pay when they use their mobile in another country. Ms. Reding’s proposal was quite simple: a cap on wholesale per minute charges and a cap on retail per minute charges. The caps for the coming years are:

Eurotariff maximum	Summer 2007	Summer 2008	Summer 2009
Mobile calls made abroad	49 cents	46 cents	43 cents
Mobile calls received abroad	24 cents	22 cents	19 cents

This is nice and an improvement to the current situation. However it will not lead to a competitive European market for roaming. For the market to become competitive it would be necessary:

1. The customers choice of roaming network will have a direct impact on the price paid by the customer.
2. When a destination network lowers its roaming charges this is immediately felt by the end-customer

In reality neither is the case:

Ad 1. When a customer uses a network abroad, it is often charged two tariffs; one low tariff for the preferential network and one high tariff for the 2-4 other mobile networks in the country. The preferential network often is a sister company of the users home mobile phone company. If a customer decides to use a different network than the preferential network, this will only result in higher charges, since the retail price has been fixed by the home network. The customers choice is limited by its home operator.

Ad 2. When a destination network decides to lower its roaming charges, out of the kindness of its own heart, this doesn’t oblige the home network to charge the customer a different retail price. As shown in the table, there is nothing wrong with charging 46 cents for making a call even if the destination network only charges 10 cents. Not only does the destination network receive less, it also hasn’t become more attractive for the consumer, so there is no way it can make up for its losses in margin by increasing volume.

A simple proposal to end this mess (so it will probably never happen)

Roaming in another country is technically and economically quite simple. There is this person who is not a customer of the destination network, but is a customer of another network, that wants to use the destination network to place a call. For the destination network setting up a call is the same as it does for its own customers, except for one thing: Authentication. It will have to authenticate the roaming caller on its network and see if it has a contractual relationship with the home network of the roaming caller. If it has such a contract, the roaming caller’s network can be billed and therefore the caller can place (or receive a call).

The ONLY thing a home network does for a destination network when one of the home network’s customers roams is to do the authentication and the billing. It doesn’t carry traffic. It is not involved in the routing of the traffic. It acts just like a credit card company. It authenticates that the customer can make the purchase (place or receive a call) and it bills the customer the price (in the process charging the shop a modest fee of a few percent). This is beneficial for the customer, it can buy (place a call anywhere) and it is beneficial for the shop owner (destination network) who knows it will get paid and doesn’t have any risk of the customer not paying.

When applied to roaming a better model would be to allow the home network to only charge a small fee for authentication and billing (2-5%) to the destination network. This way it’s billing and authentication costs are paid for and it’s adequately compensated for its troubles. (Visa has made quite a living on 2%-5% margins) The destination network would then be free to charge the consumer a charge it deems fit, whether it’s a receiving call, outgoing local call, international call etc. When it wants to lower its prices to attract more customers it can do so. The price conscious consumer will be able to switch.

Type of call	Network A	Network B
Local call	25 cents	15 cents
International call	40 cents	25 cents
Receiving call	20 cents	22 cents
SMS	10 cents	17 cents
Internet data	1 euro/Mbyte	20 cents/Mbyte

The effect for the consumer would be, that when arriving abroad he/she will be greeted by billboards and SMS-messages of the large networks explaining their tariffs for placing and receiving calls, sending and receiving SMS and even using mobile internet. There will be competition because the network that charges least is most likely to get the most customers.

A smart consumer will even change networks to get the lowest price for each action be it an outgoing local call, incoming call, international call, a call home etc. A smart customer will use Network B for local and international calls, but use A for SMS and receiving a call. (A very smart consumer would even compare the prices abroad to the prices of calling nationally. Think of it , if in France a network would only charge 3 cents per minute to call to a fixed line in Germany, a German might be cheaper off using a French network to call a German fixed line when being in Germany if in Germany this costs 8 cents a minute)

This will promote genuine competition between networks. All of them will want to get some of the revenue of the visiting customers. If there is a network C, it could decide to offer the lowest price for all types of calls and data. When MVNO’s would be allowed to offer roaming to visiting tourists and business people as well, prices certainly would drop. For MVNO’s it’s the simplest way of running a network. No contracts with individual end-users, no need for hand set subsidies, no risk of non-payment, contract disputes etc, just a contract with the home network is necessary.

From a regulatory point of view this idea would be great. It wouldn’t require any kind of regulation in a competitive market. The consumer is free to choose from multiple providers that can all independently set their prices. There is no need for a cap. It’s only necessary to be vigilant for market power and cartels, which is all in a days work. And they should watch out for home networks charging higher percentages, or destination networks posting wrong prices and/or weird schemes. It get’s better. By requiring the home network to only charge what the destination network is charging them, it becomes possible for citizens of the EU to enjoy low tariffs globally. When a Turkish operator is charging $1 per minute, that plus 5% will be the retail price for the end-user. If they have the option to choose multiple operators in Turkey, they can choose the operator that charges the least.

Would this scheme be implemented anytime soon? No of course not! It would promote competition and declining revenues and profit margins. One could expect the whole industry to be heavily against the idea. But it’s great thinking up such an idea.

Update October 29, 2007: Documents obtained by The Times show that the UK government is heavily opposed to getting the consumer a good deal in mobile roaming. The UK governement was giving the operators minute by minute updates on the state of the negotiations in the EU Council, including such remarks as "UK not happy bunnies" when the proposal was geared too much towards the consumers. All in all this shows that an idea like the one above will never see the light in real life.