---

Satellites have for many years been the mainstay of telecommunications links into Africa, with terrestrial and submarine fibre covering only a small part of this vast continent. Over the past decade, almost all new phone lines - including those in many urban areas - have been established using GSM technology and there is increasing interest in the use of wireless technology for Internet access. So, do satellites have a future in this underserved part of the world? Jeremy Rose explains why satellites are now even more important in providing critical infrastructure.

Africa continues to be the underserved continent as far as telecommunications facilities are concerned, but the good news is that demand is burgeoning for Internet access and similar services, as well as GSM backhaul. Satellites are able to provide instant infrastructure to service much of this demand but, at least in the short term, there is a very significant shortage of capacity which has resulted in very high pricing. Over the next five years there are a number of satellites due to be launched and, hopefully, this will result in a return to more reasonable capacity pricing. But COMSYS estimates indicate that demand will continue to outstrip supply for the foreseeable future.

There are also a number of submarine fibre projects that are planned to become operational over the next few years, and these will provide significant additional capacity in those areas within easy reach of the places on the coast where the cable lands. But for much of Africa, the lack of connectivity to areas located away from the costal regions or the larger cities and towns means that a substantial proportion of the population has no real access to telecommunications links. It has been argued by some that the centres of economic activity are in the well-populated areas, but this simply ignores the fact that rural medical facilities need access to the Internet, and that mining and agricultural businesses may be located well away from urban areas.

In summary, telecommunications in Africa can be characterised as follows:

• Significantly lower teledensity than other regions of the world, with the split between urban and rural teledensity exacerbating this effect. For example, in many African countries the "average" teledensity (as reported by the ITU) is less than 2-3 lines per hundred population (compared with >50 in many European countries). However, Africa compares even more poorly with respect to rural teledensity and in many countries there is effectively no real telecommunications network outside of the cities and larger towns.
  

  COMSYS

  COMSYS is a satellite communications consultancy firm based in the UK, but with global clients. The firm has been active in the African market since the early 1990s and over the past year has conducted four studies of demand for satellite services in Africa. COMSYS also publishes The VSAT Report, the tenth edition of which has recently been issued and which provides extensive data on VSAT operators and users in Africa, as well as the rest of the world. See www.comsys.co.uk for more information.

• Teledensity levels are being increased through the use of GSM networks and most new phone lines in Africa are now being established by cellular operators.
  
• An increase in offshore data processing applications. For example, most of New York's parking tickets and traffic violations summonses are processed by a company in Accra.
  
• Increased use of the Internet and a significant increase in ISPs and Internet cafes providing low cost e-mail and web access. Most of the Internet cafes in Africa also offer VOIP service, providing long distance calls at relatively low cost.
  
• Increased fibre capacity linking major cities in a number of countries, and providing increased connectivity between South Africa and some other coastal Africa states with Europe and Asia.
  
• Increased demand for centralised processing for corporate users. For example, producers of fruit and vegetables for the UK market based in Kenya and Tanzania are required by their UK customers (primarily the major supermarket chain) to establish and maintain access to their Extranet servers. The trend over the past twenty years towards decentralised processing has been reversed over the past 3-4 years as telecommunications costs become much lower.
  

• Relaxation of the regulatory environment in essentially all African countries, resulting in increased competition for cellular and other wireless services, and for basic international resale. There are still a very small number of countries where licensing costs and administrative procedures are quite restrictive, but in almost all cases governments are now realising that the relaxation of telecommunications monopolies results in increased national efficiency, with concomitant increases in GDP.

It is particularly interesting to note that most of Africa's new line infrastructure is being established via GSM / wireless technology, although most of the infrastructure is within closely defined geographical limits and that almost the entire continent has extremely low teledensity. For example, teledensity in Chad is 2.3 lines per hundred population - but this is an average over the entire country, with almost all lines being in urban areas and essentially none in rural areas.

In the African continent, much of the terrestrial connectivity dates back to the Panaftel (Pan-African Telecommunications) network built between 1975 and 1990. The bandwidth of these circuits tends to be measure in hundreds of circuits or in hundreds of MBit / second rather than in the common units of fibre today (Gbit / second or even Tbit / second).

If the expenditure on fibre in established markets during the "boom" years has yet to achieve a return on investment and with industry views suggesting that significant capacity upgrades or new systems are not expected to be seen until 2013 onwards, it appears extremely unlikely that the current African terrestrial connectivity will see any significant improvement.

So, in general, satellites continue to be used for international connectivity in Africa:

• Where no fibre is available.
• For land-locked countries.
• On those trunk routes where the capacity needed is quite small and the cost of laying optical fibre is not justifiable.

The use of satellites for trunk links has meant some increase in demand for connecting international voice switching centres and ISPs in remoter regions to the Internet backbone. As Internet usage grows worldwide, a point may come where optical fibre could be justified, so it is not expected that the satellite trunk link market will grow substantially. On the other hand, VSAT networks enable enterprises to operate networks covering many remote sites under their own control and with a uniformity of service quality which they would probably not be able to achieve if they were reliant on local public networks.

As a result, VSAT services can be differentiated adequately from terrestrial offerings and so they have experienced steady growth. COMSYS research over the past year has shown that there is currently a very substantial lack of useful satellite capacity serving Africa, and that this situation is likely to continue until at least 2010. We expect there to be several new satellite launches over the next two years, and these will bring significant additional capacity to the market. But even then, we forecast that the increase in demand will continue to outstrip the increased number of available transponders.

Part of the reason for this problem is historical, with satellites linking Africa having traditionally been low power C-band with global coverage, and operating to very large earth stations. Most African countries routed essentially all of their telecommunications traffic through a gateway earth station to a satellite in the Intelsat system and this is still true for a majority of links. Over the past decade, however, the advent of Ku-band satellite capacity and higher power C-band transponders has meant that the market for Direct to Home TV broadcasting and VSAT / corporate networks has burgeoned, and in the past five years demand for GSM backhaul links has resulted in tens of transponders being used for this application.

The problem for the satellite operator in dealing with this additional demand is that it takes at least two years for them to design, build and launch a new satellite. Whereas we do not forecast that growth in demand will decrease it is possible that some mismatch in transponder type and coverage and sources of demand will continue. For example, most users in Africa currently want to terminate traffic into Europe, for interconnection to the PSTN or the Internet backbone, and most Pay TV services are uplinked from Europe. But this situation may change over the 10-15 year lifetime of a satellite, and the spacecraft design needs to take account of this factor. There have been calls within Africa for satellites to be launched that provide intra-African coverage that does not rely on interconnection to Europe, or North America.

But the fact is that most Internet traffic is accessing web sites in those regions and will continue to do so for the foreseeable future. And even if that change happens quicker than our prediction, the satellite operators need to plan for satellite designs which will cope with unexpected changes in traffic patterns over the next two decades. By way of example, the United States Federal Communications Commission (FCC) decided in 1995 to invite applications for Ka-band satellite services within the USA. There were 17 filings to the FCC, from established satellite operators, defence and aerospace giants, and start-ups - and each listed the applications that would be served, from telemedicine to distance learning, videoconferencing and CAD-CAM file interchange. What was most surprising was that none of them mentioned Internet access as a major driver of demand and the word Internet is barely mentioned, with most filing documents not mentioning it at all. The lesson for Africa is that satellite operators cannot launch satellites that will not be able to cope with these changing patterns of demand, and cannot provide an easy fix to the supply shortage.

The other significant problem in Africa is the currently very high cost of space segment capacity, which has in some cases quadrupled in the past 2-3 years. What was very low priced inclined orbit capacity selling at about $1,000 / MHz / month a couple of years ago may now be the only available option at $4,000 - and some premium C-band capacity is selling for up to $8,000, which is similar to that charged for premium Ku-band capacity in Europe. As new satellite capacity comes into the market, one would expect that pricing will fall - but that assumes that the demand curve will grow at a lower rate than the supply curve, and COMSYS studies over the past year found that this is unlikely to be the case.

So, satellite operators bringing new transponder capacity online in Africa will have little incentive to reduce pricing, even if they take the position that some users will find it difficult to afford the new capacity. Some of these users will migrate onto terrestrial facilities - but in many locations this option may not be available to them. Hopefully, the indigenous African satellite operators will be able to buck this trend and offer lower pricing that their non-African competitors but, at present, the pricing levels they are quoting are broadly in line with the industry standard. There are many new satellite operators about to enter the African market (see sidebar) and existing operators will be providing significant levels of additional capacity. The African continent is burgeoning with new demand from Internet users, corporates and Pay TV operators - but whether the end user will be able to afford access to this new market has yet to be seen.