Home » Blog » Africa » Eric Osiakwan at Berkman

Eric Osiakwan at Berkman

Eric Osiakwan is a busy man. Near as I can tell, he’s one of the few people I know who has more job titles that I do. And since he’s been to twenty-five African nations in the past five years, he may be one of the unlucky few who spend more time on airplanes than I do. (Joking about this over coffee later, Eric acknowledges that his last girlfriend once told him, “If you like the airplane so much, why don’t you marry it?”)

I last ran into Eric in Grahamstown, South Africa, where he was speaking at the Highway Africa conference. As the executive secretary of AfrISPA (the African ISP operators’ association) and GISPA (the Ghana ISP association), he’s been hard at work on the issues surrounding the proposed EASSy cable, which will complete a fiber-optic link around the African content and, if all goes well, radically reduce the cost of connectivity.

What’s happened with EASSy so far has been pretty fascinating – at its onset, it looked like EASSy would follow the “closed consortium” model that’s helped keep West African bandwidth so expensive. Eric shows a slide that suggests that connectivity in US universities costs roughly $0.12 per kilobit per second of connectivity, while connectivity in West Africa is $8 per kbps – more expensive than satellite connectivity, or connectivity in Central or East Africa. Ironically, the introduction of a cable in west Africa – SAT-3 – hasn’t meaningfully dropped prices in many countries.

Eric describes two approaches to making SAT-3 more affordable. In Ghana and Nigeria, pressure from competitive ISPs strengthened by ISP associations has been able to “push back” on the consortium pricing. As a result, the same E1 circuit that costs $25,000 per month in South Africa costs $1,500 in Ghana. ISPs made satellite connectivity more affordable, forcing the SAT-3 providers to cut their costs much closer to the wholesale cost.

In Mauritius, they’re taking another approach – trying to make the argument at the government level that connectivity is an “essential facility” and using regulation to open access to the cable. Ghana, Nigeria and South Africa are rumored to be exploring this model as well.

The fear has been that the proposed East Africa cable – EASSy – would fall into the same economic traps as SAT-3. But something very interesting has happened around EASSy – a great deal of momentum has developed around the idea that EASSy should be “open access”, that any entity that wants to purchase connectivity from the cable should be able to at a reasonable price without undue restriction. There are forces suggesting that what’s most important is building the cable quickly, and that an open process is bound to be more complex and involved. But Eric and others have argued that the SAT-3 clearly screwed things up and that EASSy has to use a different model, even if it slows down construction.

Eric dreams of a cable where different entities can buy in via different models. In countries where it might be profitable to have access to a fiber cable, like Kenya, the cable should allow for private investment. In countries where private investment in the cable would be at least five years off, Eric sees the possibility of “stretch” funding – public/private partnership to help make private investment in infrastructure more reasonable. In other countries – Burundi, for instance – the cable needs to be treated as a social good, paid for by donors. Eric’s most radical idea is that we could increase African ownership of the cable by floating some ownership shares on regional stock markets, allowing individuals to own a piece of the cable as well.

The conversation broadened quickly into a discussion of communications on the continent, and how communication enables entrepreneurship. Eric suggested that top-down approaches to development miss some of the most exciting innovations on the continent, and that people would be well advised to watch new communication infrastructure in Africa to see what business models develop around it.

Asked about uniquely African innovations in telecoms, I offered four areas where I thought Africans were leading the rest of the world:

– Narrowband – innovative connectivity solutions that use very little bandwidth, like the Ghana “Javelin” project, or Fidonet nodes in Zimbabwe

– Localization – Translation of open source software into a wide variety of languages, especially through the help of organizations like Translate.org.za. Localization of software for challenging environments in projects like Ubuntu.

– Radio – Use of community radio for information dissemination, integration of data and radio in projects like Geekcorps Mali.

– Urban wifi, with huge wifi networks in Accra, Bamako and other African cities.

I wish Eric had an hour to work through his slides – his thinking on the topic is really strong, and I’m hoping he’ll take this dense slide deck and turn it into an article soon for everyone interested in this important project.

22 thoughts on “Eric Osiakwan at Berkman”

  1. Nice piece Ethan and capturing Eric’s thoughts allows some insight into the current mindset of African ISPs that we would not otherwise have access too.


    The section on where Africa is leading innovation also strikes a chord – I have been seeing lots of stuff happening that looks very interesting in this area including some work out of the people across the river from you. I also posted a piece on EASSy a while back where theorized that the first optical link built
    between Kenya and the global fibre backbone would win it all
    . It would have been great to to see what Eric’s thoughts on the parallel effort to build a cable to Oman from Kenya were and if they sync’ed up at all with this line of thinking.

    – Steve

    – Steve

  2. Pingback: j's scratchpad

  3. Steve, COMPETITION is in the scheme of things the heightest point of change which assures good service and affordable price so YES, am all for more cables on that path.

    Personally i think the Eastern and Southern African economies can support more than one cable in that it would trigger growth and volume if the unit cost are low with expectations of high volume.

    This is the TRUST of the Open Access sermon and am sure the force with which we are moving EASSy would set that precedent for others to follow. For me if EASSy could be built with NO debt as i profess @ http://www.tectonic.co.za/view.php?id=1160 then we would have achive a three prong success of 1. an African re-oriented experiment from a closed access SAT3 to an Open Access EASSy (which is not EASY) 2. which is African lead with support from the international community and 3. African majority ownership with debt.

  4. Ethan said “I wish Eric had an hour to work through his slides – his thinking on the topic is really strong, and I’m hoping he’ll take this dense slide deck and turn it into an article soon for everyone interested in this important project.”

    Yes, Ethan has being on my case for not being so regular on my blog and i admit i do a lot of stuff that could flow out to the community and it is not that i want to keep stuff to myself. Am just too much on the other stuff that i have not being able to cultiavate the habit of blogging like i email.

    Sometimes i cannot help but burst out on the need to avoid an EASSy Debt for Africa which is not my blog but at least i can use ETHAN’s space, what are buddy’s for…..:-) ENJOY.

    On August 29 2006, seven Southern and Eastern African countries signed the Inter-Governmental Protocol of the Inter-Government Authority(IGA) of the East and Southern African Submarine System (EASSy). This is the governmental framework through the New Partnership for Africa’s Development(Nepad) within which the cable is going to be owned, built and operated.

    The protocol, which is the outcome of an Africa-led consultative process, mandates that the EASSy cable has an African majority ownership. The current proposal for the cable is a combination of debt and equity financing of 70 percent against 30 percent for the total cost of US$300 million.

    The question that must be asked is why we want to saddle Africa with another debt if the business proposition of the cable is viable?

    The Nepad E-Africa Commission, which is facilitating this process with the governments’ mandate to have an African majority ownership of the cable based on an open access structure, must consider my proposal not to accrue debt for this project because much of the money can be raised through equity and stocks on the continent.

    The EASSy Special Purpose Vehicle(SPV) must be owned in a public-private partnership with the participation of governments, private sector, educational institutions, network operators, civil society and consumers.

    The EASSy SPV should be listed on the various country stock exchanges so that it works within the stock exchange discipline, which allows its stock to be traded without burdening the company to make huge profits to pay it shareholders. This approach would meet the current “regulated return on investment” clause in the protocol in that the company would not be bent on paying it’s investors huge profits so would price capacity at cost however the investors can trade their stocks in the company on the stock exchanges to make profit based on the performance of the company.

    Governments and public institutions must be able to invest public funds, pension funds etc. into the EASSy SPV. The stock market would serve as a platform to trade these shares later or an exist strategy to recoup the investment.

    For the “indigenous” private enterprises the proposal is to lower the financial uptake for equity from the current US$1-2 million dollars to between hundreds of thousands of dollars and US$1 million. This must include not
    only Eastern and Southern Africa private enterprise.

    Educational institutions who consume a lot bandwidth must also be allowed to invest like the UbuntuNet Alliance which has about US$3 million for the purposes of participation in the EASSy SPV.

    Civil Society and consumers must be allow to purchase shares or bonds of the EASSy cable on the stock market — hence my proposal is for the various governments to guarantee the initial public offering of the EASSy SPV in the various country stock exchanges.

    The trading of stocks of the EASSy SPV on the exchanges would seek to rapidly expand the participation of the African people and create the African ownership, which is the flagship of Nepad.

    The process would also generate long-term activities on the exchanges and create a trading post for a critical regional infrastructure company, which would ensure effective and efficient management of the enterprise.

    This would also have an impact on the stock markets in that trading of an “unusual” entity would create innovation, ensure that our financial sector is able to re-engineer to scale with development interest and ensure that private interest is at par with development goals to create a win-win situation.

    The stock market serves as the platform for trading the stocks of the EASSy SPV so that should the company be doing well then the investors can make money by trading their shares; otherwise the stock market is a good exit strategy for those who want to dispose of their shares if the company does not do well in their opinion.

    Why should we saddle Africa with an EASSy debt when the viability of the project can guarantee raising equity for its implementation ensuring that an African led process, is African financed WITHOUT DEBT?

  5. Eric,

    Lots to read and think about here. Thank you so much for sharing your thoughts on this. I will print this out, read, think and be back.

    And Ethan is right, we would love to see more on your blog as well as an RSS feed. I check on your every couple of months and had not seen anything new for a while apart from the latest burst of EASSy content.

    And thank you Ethan for letting Eric put this up here.


  6. Pingback: …My heart’s in Accra » Wireless versus wireline in Bamako, Mali

  7. I am not sure if Eric is still checking on this thread once in a while but I hope he is as I have a couple more questions and some thoughts. It is kind of hard to think and write and impossible to see what one has written before submitting it in this post window so bear with me.

    On the economic model, making lots of simplistic assumptions (which I ask that Eric fix when he runs across them) and depending on my high school math, if we assume that the cost of EASSy is to be recovered on a cost + 20% simple interest basis for the current debt funded model, for the $300m cost over a 30 year loan, total cost of the project is $360m for a 4 gbps pipe ( I will postphone the task of looking at the economics of running the link to another time )

    Considering 2 scenarios:

    Scenario I
    – assuming $75/mbps/month is available to service infrastruture debt over the life of the loan, this means that income available for servicing debt is

    4000 * 75 * 12 * = app. $3.6m p. annum => It would take 360/3.6 = 100 years to pay for the infrastructure

    Scenario II
    – assuming $600/mbps/month (8 times as much as scenario one) is available to service infrastruture debt over the life of the loan, this means that income available for servicing debt is

    4000 * 600 * 12 * = app. $28.8m p. annum => It would take 360/28.8 = 12.5 years to pay for the infrastructure

    It seems that that pricing would have to lie such that $600/megabit/month is available to pay for infrastucture if a loan were to be paid back in a 10 year period implying a cost of about .49 US$/kbps. This is still way way less than the current $4.38/kbps ($4485.12/mbps/month) but much more than the $.12/kbps ($122.0/mbps/month) American colleges pay for access.

    Based on this math, the equity/debt model that Eric is proposing starts to look very very interesting and workable from a financial perpective since it is not fully dependent on income return to be successful as institutions putting money into the project are very likely not going to be looking for a monetary return. Still, this is essentially break even pricing and the question that has to be asked if the capital markets would go for this model at this price UNLESS a sufficient number of institutional investors can be found who would improve the DEBT/Equity ratio for the project and make it more appealing to those looking to the project as a investment vehicle.

    The question then for Eric is if he believes that there are enough institutions with the kind of financial resources required who are willing to partitcipate in EASSy (is this what the slides call “Social Funding”?). (Eric specifically named UBUNTU Net that has 1% of capital costs available to throw in the pot, EASSY would still need maybe 30% of this type of financing to make the numbers work).

    As far as competition goes, Eric replied that as far as AfriISP is concerned, “COMPETITION is in the scheme of things the heightest point of change which assures good service and affordable price so YES, am all for more cables on that path.” Bearing this in mind, it seems that the obvious question that needs to be asked is if AfriISP is persuing any other backbone or intra-regional connectivity projects (i.e. has moved on from proposed to “In Planning” or “In Implementation” under the Open Access initiative. 70,000kms of proposed network is A LOT of infrastructure to build. As an aside, I especially like the fact that the Open Access project is using a combination hybrid/wirless design that will greatly reduce infrestructure cost.

    On a completely different front, I wonder if Eric can tell us if AfriISP is doing any collective work on behalf of its member ISPs as a continental body on dealing with the issues of last mile access with regards to EASSy or if they are leaving that to the individual ISPs.

  8. Joshua, i just flew into Oakland and was picked up my colleague who told me he is going to Guinea on Tuesday to establish our long distance wifi links (am part of the team doing this in Ghana as part of the TIER Group @ UC Berkeley) to connect radio stations so they can share information and this is funded by USAID Washington (Brian King, really nice fellow).

    Ethan could not emphasis this more in our talk, imagine community radio stations with seamless wifi connectivity. Even without that i can say that the radio community in Africa is largely responsible for the democratic tenats we have and Ghana is a typical example.

    Once such a cheap and inexpensive link is in place in a place like Guinea when war breaks (am not saying it should happen) it can serve as a major resource for combat and when the war is over it can serve a useful resoruce for rebuilding because this would take care of the major communication needs in terms of voice and data.

    Post 9/11, Barry Greene and a Cisco team build a massive VoIP network which was used in the disaster recovery. When Kathrina came, Jim Forster, another Cisco friend told me how they immediate deployed a wifi network with voice and data capabilities that was used to restore the situation.

    So my answer to you is, worry less about the skeptics they would always be there when African is mentioned. Am an optimist and a pragmatic one at that who believes that the availability of these resources could be used in the positive good as well as negative bad.

  9. Ntwiga, your analysis is on the money but my figure was not wrong because assuming that the Development financing Instittutions (DFIs) who are all for “regulatored return on investment” on the EASSy cable would not ask for commercial terms. More so that if Africa has a capital flight of $500million annual due to transiting of local traffic International which is because the same institutions consider the building of Telecom infrastructure the domain of private enterprise and hence it has not happened for years. Cant i ask them in the name of reducing poverty in Africa to offer that as a “grant”? Dont get me wong because these institutions fund road, water, electricity on the common good basis so why not Telecom infrastructure?

    Moving right along, am now even saying that if you wont give Africa that grant dont saddle it with another debt @ http://www.tectonic.co.za/view.php?id=1160 so you would ask me, how then do we raise money for EASSy buildout. My radical propositon is lets go to the stock markets in the various African countries and raise the money there – in otherwords the discipline of the stock market does not put pressure on the operating enterprise to provide dividends because once it is doing well and the stock market is active, the shareholders would be trading their shares and making money. Devidends then become secondary.

    I personally represent two ISPs who have more than a million dollars to invest in EASSy but guess what my recent meetings in Joburg and in DC reveals that they require those companies to have an international gateway license to make that investment. The companies am representing do have but the question must be asked why do they need a license to make an investment when the investment is de-linked from their access to capacity on the cable?

    If we minimise the requirement from the millions of dollars i can guanrantee you more ISPs who would be interested in making this investment and the incentive for that is the cable would be owned and managed by Africans. This also has links to the stock market activity in that they can exit through the markets and recoup their money – generating more action on the stock markets. So the EASSy cable can actually be used to stimulate major financial interaction on the continent and not only the bandwidth needs.

    The 70,000km of fiber with wireless hybrid was done with Rahul Tongia and colleagues from Cornegie Mellon and the conclusion is, that amount of fiber can be build under less than a billion dollars at FOB. However you can see from the slides that much of that already exist so really we can raise the rest of the money to complete the assignment within Africa.

    On your final question, most ISPs in Africa have built the wireless networks you are talking of because the copper in most places is nothing to write home about but guess what, the regulatory policy environment is stifling innovation. Most ISPs cannot do VoIP and even some of the innovation around data raise eye brows in some countries.

    By the way if the ISPs have cheap access to international connectivity from EASSy dont you think they would use the rest of the money to build the lastmile wireless network to deliver service? And if the lastmile network exist already dont you see them using their money to do more R&D so they can innovate on the edges of the network?

    Dont worry, keep coming because am watching this space.

  10. Eric,

    Thanks for clearing up many issues and questions (some of which I had not even gotten to posting) in your reply. I like many of the ideas that you have passed across including the concept of “public common good”, the building of a continental hybrid network and the need/how-tos of raising private funding for African communications infrastructure projects.

    Some more quick questions that bear asking while I think about your last round of comments if I may. You will note that my outlook on all of this is so East-African centric as to seem myopic: please forgive me for that.

    1. On the issue of KDN’s announcement of a plan to build a competing link to Oman. Looking at the financials for EASSy, this proposed link more and more begins to look like “vaporware” at the $150/mbps/month pricing they claim to be aiming at (0.14 US$/kbps which is pretty much what academics are paying the US according to your slides). I cannot figure out how it would be possible to price at this point and still have a commercially viable project.

    What are your thoughts (if any) on this?

    2. Switching of intra-African telecomm traffic – this is currently an $500 sinkhole on African resources which makes EASSy and the rest of the projects that you listed in the pipeline commercially appealing even without a link to the rest of the internet backbone. Bearing this in mind, why isn’t more progress being made on this interconnectivity front on a smaller private level by ISPs and telcos with gateway licensing?

    3. What do you think that the future of VoIP is in Africa in the short term (

  11. 1. The business model here is you lower the price and trade the volume, i did a personal spreadsheet calculation for EASSy and came out with $125 so EASSy can still compete with the KDN offer.

    2. I told you of the EARPTO crossboarder link which essentially means they are going to link a regional carrier to connect one country’s traffic to the other. Until the regulatory policy framework allows’s that the ISPs and operators can operate illegally.

    3. Very very cheap communication even in the rural areas.

  12. Eric

    I re-read my comments and now see the ambiguity in my last post: when I talked about “vaporware”, I was talking about KDN’s proposed link, not EASSy.

    EASSy can and will have a $125/mbps/month price point but that would be under the “Open Access/Common Good” model that you are proposing , not a commercial one. On the other hand, it now seems to me that KDN cannot price at $150/mbps/month with making significant losses. Even US educational institutions which have the best rates possible (aside from wholesale buyers like search engine services and consumers in Far East nations like Korea) are only just able to get this type of pricing.

    The question I was really asking is if you think that the KDN proposal is realistic giving the economies of supplying bandwidth in Africa.

    – Steve

  13. KDN is a purely commercial enterprise and would not do this for a lose not even charities do so am sure they may be using a model that you and i may not know. My best guess is that they are going the volumes way and am pretty sure that the commercial enterprises and other elements in Eastern and Southern Africa are enough to support the volumes approach even on two cable.

    I may be wrong but….

  14. Eric,

    Thank you very much for taking the time to answer all the questions that I posted up here and sharing some of the work that you are doing with us. You have given me lots of stuff to think about.

    – Steve

  15. Eric, comparing the pricing structure for bandwidth in the US to Africa is not a good one because for a start the US has quite a lot more telecoms infrastructure than (I would say) all African countries put together and has had the capital to build it out whether it was through US govt subsidy, huge telecos or stock market boom of the late 90s.

    The US has gone through all the issues from monopoly to duopolies to a relatively competitive telecoms market while African countries for the most part are just getting started when it comes to infrastructure.

    Either way the returns have to be there to justify the investment whether it is in the US or in Africa.

    In terms of bandwidth beyond returns it might be a good idea to change the ownership structure of the fibre cables. Cables like SAT-3 are highly priced not only because of the ROI built into them but also because of monopoly ownership.

    Reducing prices may be achieved by requiring monopoly owners to resell bandwidth competitively to service providers. But it has to be determined whether prices are higher than they should be in the absence of regulation and with investment returns taken into account.

    About Kenya’s link with the UAE. I think EASSY, SAT-3 are good initiatives from the standpoint of providing access for the continent as a WHOLE. I think they would make intra-African communications cheaper in the LONG RUN not just for Kenya but for other countries on the East African coast and the interior. I think that they would help African telcos get into the game of MANAGING international network infrastructure instead of just outsourcing it to an external party. These objectives should be kept in mind over and above the debates over access.

    But the emphasis should not be on access models at the start but getting the infrastructure built first and foremost.

    Access is a secondary issue for regulators to deal with and SAT-3 is where the focus should be placed to explore alternative ways of bringing down prices whether its through bringing in new longer-term capital or competition or what not.

  16. Our U.S.-based nonprofit organization, Order of Kush International, is interested in developing an international cable/satellite network, called the Sons and Daughters of Africa (SDA)network, and would like to tap into your EASSy system. Do you have a pricing system established for such a network and, if so, what is the monthly cost?

  17. It’s cheaper to fly to the China from South Africa,and download something big, than to download some qtty of the traffic in SA. I have huge experience about it(
    SA – incredibly expensive. Even hosting have small packages, and always over traffic and getting surprised by hosting providers with those fantastic invoice(

Comments are closed.