Eric Osiakwan visited the Berkman Center yesterday from Ghana, by way of Stanford University. He’s an old friend from the Ghana ISP scene, one of the founders of GISPA (the Ghana ISP association) and currently is acting as executive secretary of AFRISPA (the African ISP Owners Association), the leading trade group for ISPs on the African continent.
Eric is a Reuters Digital Vision fellow and is focusing on questions of fiber and wireless internet and telecommunications access in Africa. Learning from the successes and failures of internet and telecom projects in Africa over the past decade, Eric is advocating for a model of infrastructure development in Africa that favors small business ISPs and NSPs over formerly state-owned telecom incumbents.
Some of Eric’s key points:
There’s a widespread perception that telecom privatization in Africa has failed. This is true, but it doesn’t mean that private infrastructure is a bad thing – just that incumbent operators have been disastrously managed and are crippled by overstaffing. Eric said more than once, “Supporting the incumbents is a dead end.” Eric doesn’t consider them “worthwhile investment vehicles”.
Any model that works in Africa has to be a shared access model. African telecom was built around phone shops, where multiple users shared one phone. As a result, Africans were well prepared to accept the idea of shared access in cybercafes. This same idea needs to be embraced as we think about building infrastructure for the continent.
The startup costs of models need to be considered as well. Eric points to BusyInternet, one of my favorite success stories in Africa – while Busy has been hugely successful, it required a $2m investment to start it up, a condition that’s unrealistic for most African internet operations.
Demand for communication services in Africa is a benefit and a challenge. While GSM services have grown at a stunning rate in Africa – 20 -35% annual growth in the past few years – operators are now having difficulty handling their customer base. Eric points to SpaceFon, another one of my favorite companies, and mentions that, while they’ve now got an amazing 1 million subscribers in Ghana (a nation of only 18 million people!), their network is starting to experience prolonged failures. Eric thinks that GSM operators are going to need a strong IP backbone to link their points of presence to be able to succeed on a large scale.
Eric’s a big supporter of Internet exchange points, both nationally and regionally, and points to an IDRC-sonsored study, prepared by Russell Southwood, which suggests possible voice and data savings of $400m a year across Africa if IXPs were properly implemented.
Based on lessons learned, Eric’s recommendation is for a “horizontal layering of communications infrastructure”, which allows SMEs to compete with larger providers on the network. Basically, he’s supportive of a network that’s technology neutral, transparent, interconnected and allows small players to rent infrastructure from larger players to offer services. This said, Eric feels like Africa needs fiber – lots more fiber – to satisfy the demands of consumers at a high quality level. He’s interested in a number of regional and cross-continental cables interconnecting the SAT-3 and EASSy cables which link Africa to the outside world.
Here, I think Eric’s being a bit conservative. One of the great successes of African telecom has been the growth of wireless services, both for voice and data. We spend some time talking about interesting technologies for running data over 50km via terrestrial wireless and about new low-cost phone technologies like CorDect and CDMA 450. That said, Eric also points out that fiber has gotten much cheaper, and that there are a number of situations, like the Volta River Authority situation in Ghana where fiber exists and is unused.
The critical unanswered question in Eric’s proposal is “who’s going to pay for this?” While laying fiber is much cheaper than it used to be, the costs are still substantial and it’s unlikely that most African SMEs will be able to raise capital to lay fiber. Eric suggests a possible model in which the World Bank lends money to SMEs to purchase fiber, requiring companies to take their companies public and repay the loans in shares, which then return to a funding pool. It’s an interesting proposal, but one fraught with possible complications – does SBC or France Telecom get to lean on the World Bank for this sort of financing? Who determines how much money is neccesary to lay cable, and how big the loans should be? I’m looking forward to hearing Eric’s answers to these questions as he develops the idea further.
Eric Osiakwan visited the Berkman Center yesterday from Ghana, by way of Stanford University. He's an old friend from the Ghana ISP scene, one of the founders of GISPA (the Ghana ISP association) and currently is acting as executive secretary of AFRISPA (the African ISP Owners Association), the leading trade group for ISPs on the African continent.
Eric is a Reuters Digital Vision fellow and is focusing on questions of fiber and wireless internet and telecommunications access in Africa. Learning from the successes and failures of internet and telecom projects in Africa over the past decade, Eric is advocating for a model of infrastructure development in Africa that favors small business ISPs and NSPs over formerly state-owned telecom incumbents.
Some of Eric's key points:
There's a widespread perception that telecom privatization in Africa has failed. This is true, but it doesn't mean that private infrastructure is a bad thing – just that incumbent operators have been disastrously managed and are crippled by overstaffing. Eric said more than once, “Supporting the incumbents is a dead end.” Eric doesn't consider them “worthwhile investment vehicles”.
Any model that works in Africa has to be a shared access model. African telecom was built around phone shops, where multiple users shared one phone. As a result, Africans were well prepared to accept the idea of shared access in cybercafes. This same idea needs to be embraced as we think about building infrastructure for the continent.
The startup costs of models need to be considered as well. Eric points to BusyInternet, one of my favorite success stories in Africa – while Busy has been hugely successful, it required a $2m investment to start it up, a condition that's unrealistic for most African internet operations.
Demand for communication services in Africa is a benefit and a challenge. While GSM services have grown at a stunning rate in Africa – 20 -35% annual growth in the past few years – operators are now having difficulty handling their customer base. Eric points to SpaceFon, another one of my favorite companies, and mentions that, while they've now got an amazing 1 million subscribers in Ghana (a nation of only 18 million people!), their network is starting to experience prolonged failures. Eric thinks that GSM operators are going to need a strong IP backbone to link their points of presence to be able to succeed on a large scale.
Eric's a big supporter of Internet exchange points, both nationally and regionally, and points to an IDRC-sonsored study, prepared by Russell Southwood, which suggests possible voice and data savings of $400m a year across Africa if IXPs were properly implemented.
Based on lessons learned, Eric's recommendation is for a “horizontal layering of communications infrastructure”, which allows SMEs to compete with larger providers on the network. Basically, he's supportive of a network that's technology neutral, transparent, interconnected and allows small players to rent infrastructure from larger players to offer services. This said, Eric feels like Africa needs fiber – lots more fiber – to satisfy the demands of consumers at a high quality level. He's interested in a number of regional and cross-continental cables interconnecting the SAT-3 and EASSy cables which link Africa to the outside world.
Here, I think Eric's being a bit conservative. One of the great successes of African telecom has been the growth of wireless services, both for voice and data. We spend some time talking about interesting technologies for running data over 50km via terrestrial wireless and about new low-cost phone technologies like CorDect and CDMA 450. That said, Eric also points out that fiber has gotten much cheaper, and that there are a number of situations, like the Volta River Authority situation in Ghana where fiber exists and is unused.
The critical unanswered question in Eric's proposal is “who's going to pay for this?” While laying fiber is much cheaper than it used to be, the costs are still substantial and it's unlikely that most African SMEs will be able to raise capital to lay fiber. Eric suggests a possible model in which the World Bank lends money to SMEs to purchase fiber, requiring companies to take their companies public and repay the loans in shares, which then return to a funding pool. It's an interesting proposal, but one fraught with possible complications – does SBC or France Telecom get to lean on the World Bank for this sort of financing? Who determines how much money is neccesary to lay cable, and how big the loans should be? I'm looking forward to hearing Eric's answers to these questions as he develops the idea further.