May 27, 2013
As a continent aspiring to bridge the digital divide and to catch up with the rest of the world, we have some major challenges to overcome. Connecting a continent with an area of 30 million square kilometres and sparsely populated (10 times the size of India which has a comparable population) is a daunting task. Africans have to be more innovative in their approaches to wire the continent if we should overcome some of these challenges. Fibre Optic infrastructure sharing is one important approach that operators can embrace to lower their CAPEX and afford to bring fibre to more Africans.
In most cases, the model for deploying optical fibre has been vertically integrated – where one operator deploys, owns and operates their own fibre optic infrastructure. It is not uncommon for three or more network operators to deploy individual fibre infrastructures along the same routers such as; Harare to Bulawayo in Zimbabwe, East London to Cape Town in South Africa or Lagos to Abuja in Nigeria.
The problem with this model is that it could cost the operator up to $200 million for a 1000km link, based on recent projects in Africa. A large part of the cost is in civil works, or trenching. In the case of cable deployed underground it could cost up to 80% or $160 million in civil works alone on a 1000km link. If you have three operators trenching along the same route, the total cost of civil works alone could go up to $480 million. Can we really afford this in Africa for every 1000km?
There are a number of infrastructure sharing models that could significantly lead to savings on fibre optic infrastructure deployment including shared trenches, shared ducts, shared cable, shared fibre and shared wavelength or packets.
In shared trenches, operators can come together and share the cost of civil works but each of the operators or service providers can install their own ducts and cable. This model can significantly lead to cost savings in civil works while maintaining each operator’s independence and flexibility to install whatever cable and optical fibre of their choice. Moreover they can use system electronics of their choice and will not worry about possible infringement on their data by their competitors. Using the cost assumptions above for a 1000km link, three operators could end up contributing only about $53 Million to the cost of civil works.
The shared trench model is already being implemented by a number of operators, including MTN, Vodacom and Neotel in the National Long Distance (NLD) in South Africa. A consortium of operators in Tanzania, Airtel, Tigo and ZANTEL, have used a similar model to deploy metropolitan fibre optic networks. The consortium had plans to deploy a national backbone but the Tanzanian government, which obviously underestimates the benefit of ICT to national development, is denying the consortium rights of way. More operators should adapt this model if the wiring of the continent should be a reality.
In shared ducts, each company installs its own cable in a common duct. In this model, they can realise additional savings in the cost of ducts, although this may take away the flexibility to blow in additional cable when future requirements so dictate.
The shared fibre model is usually implemented by service providers who build their own networks and offer dark fibre, or passive fibre that is not lit, to other service providers. Instead of building their own networks, service providers can lease fibre pairs on an already existing network and use their own electronic systems to light the fibre. This is the model that has been implemented by Dark Fibre Africa in South Africa, Liquid Telecommunication in Zimbabwe and Kenya Data Networks which is now owned by Liquid Telecom. This model may be more suited for the small service providers without enough financial muscle to enter into a consortium with others to build a network. For large operators, the longer term cost of leasing fibre may justify building their own network.
For smaller service providers who can neither build their own network nor lease fibre strands, their option is to lease either a wavelength or packets within a wavelength of a lit optical fibre pair. This model can be contentious when competitors have to transmit their data on the same network because of fear of industrial espionage. Service providers adopting this model need to take additional security measures to protect their information.
In order to bring terrestrial optical fibre to as many Africans as possible, operators have to be creative in the way they approach network deployment. Infrastructure sharing can significantly lower CAPEX in fibre optic network deployment.