The African Internet and The Third World Dog

Johann Botha

$Id: article.xml,v 1.6 2003/11/10 12:08:50 joe Exp $

All trademarks and copyrights are owned by their respective owners. All other material is owned by Johann Botha. All rights reserved. Reap the Whirlwind. Articles available at:

The Third World Dog

Stop kicking the Third World Dog. Stop keeping it down.

The Third World Dog is a hard animal. Tough as nails, like a crocodile, only skinny and deceptively scruffy looking. Resilient to disease. It may not be as pretty as pure bred dogs but it has refined its survival tactics to a level no other breed has been able to. A true masterpiece of evolution.

It can go without food for long periods of time, it lives in harmony with the many parasites that suck it dry. The word comfort is not in its vocabulary. Even when it has ample food supply it only eats enough to stay lean, always anticipating the next hungry, cold night.

The Third World Dog's personality has adapted to become humble and shy, avoiding confrontation. Living in fear of the random kick in ribs for its master's sick amusement.

What makes the Third World Dog an exceptionally fascinating animal is its tolerance for abuse. Most animals in this environment would simply have quit, but the Third World Dog just keeps on rolling with the punches. Herein lies the connection. The African Internet today is very much like the Third World Dog.

Stop kicking the third world Internet community. Stop keeping it down.

Internet in Africa

Why do I feel like writing about Africa's Internet problems ? Is there anything new to say ? Maybe it's because I just want some medium to vent my frustration, or maybe it's because I just returned from the IWeek 2003 conference and I'm optimistic that a change is coming.


"The Revolution will Not be Televised, The revolution will be LIVE."

 --Gil Scott-Heron

To get an idea of whats happening in the African Internet world we have to first ask.. exactly how much does the situation suck ? Let me try to sketch you a picture. My view of all this breaks down into three levels: Africa, South Africa and Cape Town. I live in Cape Town, not the worst address along the Information Dirt Road, but there are many things that could be improved along the way. Living on the African continent and being in the technology and communication business is an interesting experience at times.

On some days I think I must be living in an alternate reality. Just like I can't seem to explain the fact that the world is dominated by lame proprietary software I struggle to grasp why Africans make laws to keep themselves down. What truly amazes me on some days is how so many countries in Africa seem to be stuck in the same boat. The big inefficient "Incumbent Telco", the weak independent-non-independent telecommunications regulator, the bureaucratic government department that does not know how to spell the words "change" or "progress". But this is nothing new and not unique to the field of technology in Africa.

On some days I'm amazed by how the First World exploits Africans and how insensitive the developed world is for the problems that Africa face. Worst yet, is how Africans let themselves be exploited. We very rarely get our message right. A message to the rest of the world. A message of change, progress, the right policies and laws, a rational voice supportive of a healthy telecoms economic environment. At the very least we should not have laws that allow others to exploit us.

All we can hope is that with time the people in power wake up... but I can hear a few of you saying, "that's crap,.. things will only change if we make enough noise". And I agree. So let me make this my first message: Africans are meek and easily subdued. Where we have democracies in Africa we need to be more politically active and lobby governments to make the right choices. "Voice over IP is not a crime", it's a technology that is and will be of benefit to all Africans. How can any government make that illegal?

You could argue that the US and Europe are extracting large volumes of money out of Africa, even for Africans to speak to other Africans. This is true. The ITU estimates that each year Africa spends $400 Million to speak to itself.. most of this money flowing to the US because voice and data traffic is routed though the US. Which brings me to my second point. We can address problems like these, but only if we cooperate and get countries in Africa to overlook some political issues to allow the interconnection of African countries.

  • 1. Lobby government, don't be so meek

  • 2. Cooperation

There are 9 Internet Exchanges in Africa. Not bad, especially if you consider that one of them runs completely off solar power, but we have a long way to go, there are 53 countries in Africa. It is estimated that it will take another 7 years for sub-Saharan Africa to have terrestrial interconnectivity using a combination of fibre and microwave.

Back to my second point, to cooperate, overlook some politics and allow interconnection. I'm told there are a number of countries which border on lake Victoria (Kenya, Uganda, Tanzania etc.). What could be better than some fibre cable in a lake carrying traffic between these countries, yet governments don't allow regional telecoms operators to connect countries efficiently, mostly to protect the "Incumbent Telco". I know that there is fibre cable running along power lines that connect South Africa and Mozambique which can, but may not be used to carry Internet traffic. The result is money flowing out of the African economy and inefficient routing.

Create content in Africa. Host content locally. This has to be my third point. Grow the local Internet market, use our Internet exchange points. I know many South Africans are hosting their websites in the US, we all know it's very cheap to do so and you may argue Africa can't compete with pricing in the US, but give it some thought, move your content back to your own country, your own economy and support your own local industry.

Some Africans like being negative, pessimistic and preach their message of apathy about the African Internet. I can fully understand the feeling of helplessness we feel but it's not very productive just bitching about the situation. I was involved in a local initiative to build a peering point some time ago where we encountered some problems and some politics. The project failed. I think it's very sad if people don't support local efforts with clear benefits to them and the rest of the community.. and then you get those who make an effort to counter such projects, strange.

  • 3. Host content locally

  • 4. Be positive and supportive of local efforts

There are projects to improve the situation and there are organisations working at a number of levels to bring about change. But the best news of all I think, is that there is consensus between the individuals running their Internet businesses, the people building the African Internet, the people who manage and support it on a day to day basis.

Connectivity Africa Workshop : 18 Sept 2003

I attended the Connectivity Africa Workshop at the Iweek conference in Johannesburg. The idea was to talk about PAVIX, the Pan African Virtual Internet Exchange. I listened to a number of really knowledgeable people in the African ISP industry, people from Kenya, Mozambique, Uganda, South Africa, etc. It was evident that many of the people at the meeting had working experience with VSAT satellite technology and knew first hand the frustrations of dealing with their local regulators.

The goals of the project included:

  • Less cash flowing out of African economies

  • Lower cost to ISPs and customers

  • Lower latency, get rid of extra hop to US, more efficient routing

When the meeting started that morning, PAVIX was to be an academic socialist ideal. Do it yourself VSAT hubs connecting existing exchange points in Africa. It would rent its own transponders on satellites and manage all technical aspects. There would be co-operative buying and sharing of bandwidth and to get around the legal issues there would be a time limit to the project and it would be labelled a "test". Basically the idea was that Africans would take care of their own regional connectivity needs.

By that afternoon the idea had evolved to an organisation aiming to encouraging a regional carrier model in a commercial environment. Connection points would still be at IXs, but transit services would be provided between members of IX's and not between IX's as a group. An RFS (Request For Service) would be written by AfrISPA and the IX's. It was agreed that there is no point in labelling the project a "test" and adding a time limit. The alternative is to aim for sustainable regional connectivity and to work within current regulatory frameworks, but actively lobby for change. Where needed, IXs would be encouraged to change their policies to allow members to buy transit at the IX. Since PAVIX was no longer a appropriate name for the project, somebody jokingly suggested we call it something like "".

AfrISPA African Traffic Research Project

If the ITU study I mentioned earlier talks about $400m/y going to the US/EU for Africa to talk to itself, how do we know this figure is accurate, how do we measure the impact of our work and the rate of change and progress. Enter PCH, IDRC, DFID and the AfrISPA African traffic research project.

There is a study being conducted to gather traffic flow statistics at IXs in Africa. By collecting and analysing netflow data we will be able to get accurate statistics on African traffic flow patterns. This information will form a basis to encourage growth in regional connectivity.

ARIN 2003-15 and AfriNIC

One very positive thing that happened recently is that ARIN seems to have been convinced that ISPs in Africa have a different profile than ISPs in the US and the environments they operate in are not even the same sport. Proposal "2003-15" aims to lower the barrier to entry for ISPs in Africa to obtain their own IPv4 address space.

This is a big step for many African network operators, as it will allow them to build more reliable networks, better control their routing policies and most importantly not be dependent on a single upstream ISP which in many cases would be the "Incumbent Telco".

I'm happy to say I was impressed by the support from African network operators for this proposal on the ARIN public policy mailing list.

It's clear to me that we need AfriNIC to become operational as soon as possible. AfriNIC has taken a number of years to become a reality and all indications are that Africa will soon have its own RIR (Regional Internet Registry).

The Internet in South Africa

There is a telecoms monopoly in South Africa. (Telkom)

I just had to make that a paragraph on its own. It's sick. But we have gotten so used to the idea we don't realise how wrong this is.

The SA telecoms environment is a mess. The telecoms sector's contribution to the SA GDP has grown from about 2% in 1990 to 6% currently, but this growth is mostly in the mobile phone business. South Africa's fixed line teledensity is declining. The cost of a local call is about 5 times what it was in 1996.

Internationally there has been a decrease in telecoms cost of about 65% in the last 5 years while in SA there has been an increase of 45%. One out of every 15 South Africans have access to the Internet. According to the ITU South Africa was 13th in the world in terms of Internet users in 1996, but by 2001 it dropped to 26th and falling. South Africa shows one of the lowest growth rates of Internet users between 1996 and 2001.

It's clear from these figures that something is very wrong. South African citizens are paying way more than they should for telecoms services and the sector is not growing as it should. Something is keeping it back.

Not only the end user is suffering, the current environment is very unfriendly towards ISPs and Value Added Network (VANS) operators. 70% of ISP's costs are directly to Telkom. Legislation forces VANS to use Telkom services, yet Telkom has the biggest share of the VANS market. This is a major conflict of interest. Telkom frequently abuses its monopolistic position with anti-competitive actions and the regulator has very little power.

Everybody loves to hate Telkom. Public Enemy Number One. The monopoly telco that's not supposed to be a monopoly, but that still is a monopoly. But Telkom is probably but a symptom of an underlying disease.

The three pigs of the South African telecoms market are, Telkom, ICASA and the SA Department of Communications (DoC). The little piggy seems to have went to market. The middle piggy seems confused. And the big piggy has one word on its mind.

Convergence. Nice word.


My ignorance of science is such that if anyone mentioned "copper nitrate", I should think he was talking about policeman's overtime.

 --Dr Donald Coggan

There are some new ideas flowing around about how to change the telco market's legal framework. The idea is to change from a model where we depend on one PSTN to one where we have many telco companies that can all offer the entire range of services, from providing copper lines to value added services. The problem seems to be that government has no reason to change the legal framework soon. There is no incentive. Government has a stake in both Telkom and the new kid (Sentech). The current, broken environment is good for both.

Somewhere around 1996 the telecoms act was amended. In this amended act the concept of a multi-media license was introduced. A bit later (2003) government sells off a chunk of Telkom. The multi media license is pretty open to interpretation, but it does not allow for circuit switched voice services.

At this point you may think that this is a good plan to introduce some competition in the market. It turned out an entity called Sentech was given the license, a parastatal, owned and controlled by the government.

If you like conspiracy theories you may like this. Government gets a nice amount of cash selling a slice of a monopoly. Soon after a parastatal company with the somewhat limited license gets a nice cash injection to build a new network. Then government starts to talk about "convergence" which would remove the circuit switched voice restriction, leaving it with a second entity with all potential the monopoly had. Slick. But hey, I don't like like conspiracy theories.

Problem is, Sentech is still a government organisation, clumsy and inefficient about a year late with its plans for rolling out its new network and services. They sponsored connectivity for this year's IWeek wifi hotspot and managed to get it working about half a day out of the five days. I recently met somebody that had some amusing horror stories about Sentech after being involved with installing billing software for them. Sentech is a laugh.

The one event this year that really left me with a total sense of humour failure was when ICASA and the DoC announced that they would not license a Second Network Operator this year, even though we should have already had an SNO more than a year ago. If government constantly shifts the policy framework there can be no safe investment in the SNO, yet one of the major reasons for not licensing and SNO is because ICASA believes the current candidates are not backed by enough investors.

Current thinking seems to be to create and protect big clumsy telco companies, yet allow them to directly compete against the small and medium sized companies and VANS using anti-competitive strategies. Maybe it's just me, but I think this sucks.

As you may have guessed my loyalty lies with the VANS. The little guy. VANS are entrepreneurs, we manage complexity, we solve problems, we innovate. VANS are the key to entering the knowledge economy, yet South Africa creates an environment that is not supportive of VANS.


The ISP business in Cape Town is like SCUBA diving in Cape Town, there is a bit less action and a bit more inconvenience, colder water, low viz and not so much to see. But if you learn to dive here you can dive anywhere.


Voice over IP (VoIP) and Wifi (Wireless Networking) are technologies that will change the way we communicate. They will change the way we view "the phone company". This is rapidly happening all over the world. Africa and South Africa has the opportunity to "leapfrog" old PSTN technology by moving directly to the new telco model. The sad thing is that South Africa will be at least 5 years behind the rest of the world in skills and experience with these new technologies because of South Africa's legal framework.

The really amusing thing is that ICASA agrees the current legal framework is a mess, pity they don't have more of a voice and they are are the DoC's lap dog.

Let me just solve all South Africa's telco problems in one paragraph. Wifi, bidirectional VSAT and VoIP should be 100% legal. License the SNO and a few other PSTN operators. Support the VANS and medium/small business in the telco market.


A few correct choices by the governments of Africa could easily create a much better environment for the African Internet to grow and prosper. What makes me optimistic is that the solutions are so clear and that there is consensus about what needs to change. There are a number of projects working to improve the situation with support from the international community.

Give the Third World Dog more freedom. The Third World Dog is your friend.

References and Credits