Internet in Ethiopia - Is Ethiopia Off-line or Wired to the Rim?
First Impressions - What Meets the Eyes
Few things in Ethiopia cause as much frustration and despair than the ongoing and deteriorating dismal state of Internet connectivity in the country. Given how the country continues to struggle with fundamental issues of development for its 80 million people, this lack of progress - and in some cases complete reversal of earlier growth - tests the hope of the most optimist observers and stake-holders.
A case in point is the state of access to the web over the past several years through the numerous Internet cafes that have popped in most major parts of the country. In a test carried over 4 weeks time in July 2007, the average speed of Internet connectivity in the country had dropped to as low as 5 KBps. This is in par with the speed of most 4200 Baud modems the rest of the world was using in the early 90s - almost a century ago in Internet time. Even the handful of Internet cafes that billed their services as broadband of 128 KBps in the new business corridor of Bole Medhane Alem - Kazanchis Road could not manage connectivity speeds better than average of 7-8 KBps. In general, the only time the connectivity speed becomes bearable is usually after 8:00 PM when fewer people log-on lightning up the load on a strained and inadequate infrastructure. As of the fourth quarter of 2007, the number of individual and commercial Internet subscribers in the country stands at a dismal 15-20,000, with an estimated 300 or so Internet cafes serving a population of 80 million plus. This amounts to only 0.025% of the whole population having a regular Internet account. It has to be noted that there is virtually no growth of the individual and commercial subscriber base as these numbers are essentially the same as those dating as far back as the years 1999 and 2000. To add despair to frustration, Internet service is down as much as 10-20% of the time. In a test carried over the same period in July 2007, Internet service was down almost 20% of the time with typical blackout lasting 12-24 hours.
Since the only access to the Internet through the country remains to be the one provided by the country's sole telecom, ETC, this slow connectivity extends to businesses (private companies, multi-nationals, etc), NGOs, private and government colleges and universities as well. By and large, this scenario is what awaits any Internet user in the country these days and could be particularly frustrating and outright discouraging for domestic as well as multi-national entrepreneurs in the ICT sector.
Another factor that demonstrates how badly the unreliable and slow Internet connectivity in the country is negatively impacting businesses is the fact that the majority of functional web-sites of these businesses are hosted overseas, Europe and the US being favorite destinations. While it is rather common to host web-sites remotely in most parts of the world, spending foreign currency (dollars and Euro) for hosting their web-sites is not an economical proposition for most local Ethiopian businesses.
However, even in the midst of these dismal circumstances, Ethiopian Internet users continue to be patient and determined to make do with whatever is available. While there are no formal surveys published so far, there are some estimates on the size of the 'informal sector' putting the number of Ethiopians with some type of occasional to frequent Internet exposure to as high as 200,000. This exposure is considered to be predominantly e-mail usage only accompanied by minimal web-browsing for news and sports. The bases for these estimates are an informal small-scale study conducted in 2004, a report in 2005, and another report in July 2007. The study reported in 2004 seems conservative with regard to its prediction of 14 visitors/day/Internet Café. The July 2007 survey suggests that this number could be as high as 20 in Addis Ababa and some of the larger cities like Nazret and Gondar. To these, if one adds the almost 120,000 students currently enrolled in the government colleges and universities with at least 75% of them having nominal access to the Internet, an estimate of 200,000 Ethiopians with some basic and minimal exposure to the Internet looks like a reasonable estimate. Notwithstanding the skewness of the data towards university students, this translates to a mere 0.25% Internet penetration which is an improvement on the estimates based on commercial Internet account holders only. Nonetheless, this represents a very dismal Internet penetration by any standard.
The View from Government Networks - WoredaNET, SchoolNET, HealthNET, AgriNET, etc.
On the other hand, however, the government argues that the actual number of Internet users and penetration is much higher than what meets the eyes. The argument is that the government has completed and installed two networks that it operates and owns called WoredaNET and SchoolNET. WoredaNet connects government offices in the more than 594 woredas in the country to regional government offices and then to the Federal government. Features include e-mail and video-conferencing as well as some VoIP. Government officials and civil servants are provided with these facilities to conduct eGovernance between the Woreda, regional and Federal levels. In theory, citizens are supposed to benefit from this eGovernance even though there is no documented instance of this. In fact, sources indicate that use of this network by citizens to participate in any form of eGovernment at the Woreda level is non-existent. SchoolNET, on the other hand, is supposed to connect all the 545 or so high-schools in the country to a gateway that provides video and audio streamed educational program. Access to the Internet is available through a downlink-only VSAT (Very Small Aperture Terminal) satellite connection. Further, each SchoolNET node at a high-school is equipped with a large screen Plasma TV. The cost of equipping all the 545 high schools with these high definition TVs is estimated to be around $1.25 million with each Plasma TV costing about $2000-$2500 each. The government continues to plan for similar networks that it owns based on VSAT architecture such as AgriNET and HealthNET to connect agricultural research entities and hospitals respectively to their regional and Federal counterparts. All in all, the government has put at least $40 million in 2005 - 2007 to implement this. In fact, the government itself has said that it will spend 10% of its annual budget in developing and maintaining these networks. These networks are exclusively owned and managed by a new government bureaucracy called EICTDA (Ethiopian ICT Development Authority) which is tasked to oversee ICT developments and policy in the country.
On paper, this ambitious plan looks very exciting and ground-breaking. So much so that even Cisco which won the project that lay the basic infrastructure for these networks in 2005 along with Hughes Network Systems put out what turned out to be an embarrassing and un-informed press release which prematurely predicted Ethiopia was set to leapfrog development. The problem with this is, however, many-folded. The most fundamental issue, however, is that the Ethiopian government seems adamant about carrying out an Internet operation model based on complete government control and totally inconsistent with ICT technology practiced in the rest of the world. The Internet was created to be a highly decentralized self-governing, fluid, chaotic, and most importantly 'access-to-all' network. Furthermore, this model pursued by the very economies that created this network is based on the sense of 'ownership' that average individual stake-holders espouse. In a model that is practiced everywhere in the world including such places like China, anyone or business requiring an Internet connection of any speed that it can afford is provided for a fee. However, in the Ethiopian government model, no business can walk straight to ETC or EICTDA (the new bureaucracy that oversees ICT activity in the country including ETC) and pay for access to the VSAT based networks WoredaNET, SchoolNET or HealthNET. The government continues to maintain an absolute and tight ownership control leaving no room for the private sector.
A case in point is a recent
attempt by researchers and entrepreneurs to utilize the WoredaNET infrastructure
for an independent entrepreneurial eMedicine initiative in the country
connecting Ethiopian private and government clinics and hospitals with
the outside world. "Ayachalim', "ayfeqedim' ('Impossible'
and 'not permitted') were the responses that were commonly heard by
entrepreneurs while enquiring with ETC and EICTDA officials about the
commercial use of these government networks. As is common in any government
controlled enterprise, these kinds of decisions are made by a single
person with no check and balances and due considerations of the bigger
picture and its devastating impact on national economic interest. In
countries that promote successful models, these types of initiatives
are not only encouraged but treated with priority. In short, such practices
point to the main flaw in such command economy model that is being pursued
in the country which continues to have massive negative impact on ICT
and knowledge-based economy in the country.
A comparison of the different networks in the country in terms of coverage, cost and access is provided in Table 1. The table shows that the government-only networks seem to possess a higher bandwidth than the commercial access for individual and business account holders, cost substantial amount of money, and suffer from no-access to the public. Further, SchoolNET has only a one-way link with a complete downtime in the Kremt season.
Broadband, Fiber Optics and Access to the Sea
One bright area that seems to offer an encouraging sign is the achievement of ETC in laying as much as 4000 kilometers of fiber optics network across the breadth and depth of the country. The network currently stretches in six directions of the country following the main highways, i.e., Addis-Jimma, Addis-Lekemte, Addis-Awassa, Addis-Gondar, Addis-Dessie, and Addis - Dire Dawa. Along these routes, a total of 78 towns have been linked with optical fiber having a 155 MBps (megabytes per second) capacity and another 46 towns with a capacity of 2.5 GBps (gigabytes per second), and nine towns with a capacity of 10 GBps. One needs to note that if and when these networks become operational, the increase in Internet connectivity speed from the current average of 5 KBps to 10,000,000 KBps translates to a theoretical 2 million fold increase!
The eventual plan is to connect the national optical fiber network to the rest of the world through the planned under-water cables surrounding the continental shelf of East Africa called the Eastern Africa Submarine Cable System (EASSy) . This under-water cable system is planned to run from South Africa to Port Sudan with landing points in six countries (South Africa, Tanzania, Kenya, Somalia, Djibouti, and the Sudan) connected to at least five landlocked countries - who will no longer have to rely on expensive satellite systems to carry voice and data services. EASSy has been plagued by a lot of wrangling and disputes and the earlier optimistic completion date is the end of 2008. Realistically, this would suggest the availability of this network to be delayed as late as 2009 and 2010.
Currently, lacking its own natural access to the sea, the first attempt to connect Ethiopia to the outside world was done through a fiber network that stretches from the capital Addis Ababa to the Red Sea port of Djibouti. This was completed in 2006; but the line between Djibouti and Dire Dawa was rendered useless through looting and sabotage allegedly by the ONLF and OLF rebel groups operating in the area. The same year, the government then approached the Sudanese to lay a safer route between Metema from the Ethiopian side to Port Sudan on the Sudanese side. An existing microwave connection to Verizon's network in Saudi Arabia from Port Sudan is planned to carry the broadband voice and data traffic to the global network. This project was reportedly completed in July of 2007 at a cost of $12 million by Siemens. However, this project remains - for all practical purposes - inconsequential. Engineers familiar with ETC suggest that the laying of the network was not accompanied with the upgrading of routers and switches of the core network operated by the telecom thereby introducing a major bottle-neck that has made broadband access in the country still an unrealized dream. The upgrading of these routers and switches should not have been a show-stopper as it could have been achieved with relative ease and modest expenditure of money. Why this problem still persists remains to be a puzzling and frustrating aspect of Internet connectivity in the country. In August 2007, Cisco was reportedly offering to upgrade these bottle-necks free of charge. As of November of 2007, the outcome is yet to be seen.
In the bigger picture, the laying of these physical wires through neighboring countries raises some inconvenient questions of geopolitical, strategic, and national security concerns that require careful deliberations from Ethiopia's side. As a result, the whole issue of optical fiber network seems to have escalated to the office of the Prime Minister. Currently, there are ongoing talks of connecting the national optical fiber network grid to the Somaliland port of Berbera. But this requires digging the lines and operating in the currently volatile region of the Ogaden - once again exposing the infrastructure to uncertainty. Other alternatives that are being considered and that have gained some currency among engineers and policy makers are connecting either the Addis-Jimma grid or the Addis-Awassa grid with the Kenyan port of Mombassa and subsequently to EASSy. The Addis-Jimma grid extension could turn out to be an expensive choice given the fact that the terrain of the virgin Gojeb and Kaffa region will pose a lot of physical difficulties dropping an altitude of almost 2000 meter from the Jimma area down to Northern Kenya. However, economic activity in the region is bound to increase with the ongoing work on at least three massive hydroelectric power projects which the Kenyans have expressed interest to link to. This could help reduce cost if the project follows some of the engineered roads currently being built for the hydroelectric project in the Gibe and Gojeb areas. The Addis-Awassa grid, the author believes, offers a less-expensive and safe alternative. In general, it seems that the most prudent approach is to work with the Sudanese or Kenyans or both with whom Ethiopia is slowly developing economic leverage in terms of the planned electrical power grid extension to these countries.
As of now, however, there is a limited (and insanely expensive) access to the growing national fiber optics network. As expected, ETC and EITCDA assign these access case by case. As of the end of November of 2007, only two international companies seem to have an access allowed by these government bodies. For example, the recently introduced eMedicine and eEducation initiative in Ethiopia funded by the Indian government has an optical fiber network component running on the Addis-Dire Dawa grid from which it is connected to an under-water cable in Djibouti through a microwave connection . The underwriters for this project are reportedly paying as much as $8000/month to ETC to keep their networks running. Interestingly, the cost of renting access to the underground cable running through the Sea of Arabia and Red Sea from Mumbai in India to Djibouti is a fraction of this cost. The Chinese company Huawei which maintains a presence in Ethiopia is reportedly paying as much as $7000/month to get this access to the only existing broadband in the country. These prices are, of course, mind-boggling and insanely expensive indicative of the kind of chaos, across-the-board confusion and lack of affordability that government ownership results in. At this point, therefore, broadband access through optical fiber network will remain elusive. An optimistic estimate puts 2010-12 as the year when subscribers in Ethiopia could expect a reasonably stable and affordable broadband access.
Even then, however, it is only fair to mention that the country's achievement in laying the infrastructure for the optical fiber network needs appreciation hoping that some of the ownership and management issues will be settled in the long run. The site of these modern day technology life lines in some of the remotest parts of the country is - in most accounts - a symbol of national pride about possibilities.
One of the interesting areas to watch in Internet and mobile telephony technology in today's Ethiopia is the role played by the new powerhouses, India and China. The Chinese presence in Ethiopia - at least in its modern form and shape - extends to the late 90s starting with the completion of the ring road in Addis Ababa. Since then, the Chinese have established a reputation of hard work, discipline, on-time completion of projects, and in some cases under-cutting of the competition. Similar to other places in Africa, the Chinese presence has inspired admiration as well as resentment. In the ICT sector, the big presence of the Chinese is in the current build-out of a 10 million subscriber capacity WCDMA wireless network by ZTE covering a large segment of the country. The optical fiber network expansion from 4000 Kms to 10,000 Kms currently being carried out by CITCC (China International Telecommunication Construction Corporation) is also considered a very important and fundamentally consequential project in the country.
The Indians have also invested as much as $12 million in Ethiopia in the past few years from a pool of $120 million they plan to invest in Africa in introducing tele-medicine and distant education. The Indian initiative known as "Pan-African e-Network" is ambitious and is a strong indication of India's desire to project its technical prowess across the globe, particularly in the developing countries. The project is planned to provide tele-medicine, tele-education, Internet connectivity, video-conferencing and VoIP services to support e-governance and e-Commerce. The total allocated fund is $120 million for build-out of a satellite and optical fiber network that links 53 countries in Africa. A total of 169 VSAT terminals, out of which 3 will be in Ethiopia are planned and are in the implementation phase. In Ethiopia, tele-education centers have already been built by the Indians at the Addis Ababa University and Alemaya (Haramaya) University with a link to a university in New Delhi, India. Further, tele-medicine centers at Black Lion Hospital and the Nekemte Regional Hospital were recently commissioned with a connection to CARE hospital in Hyderabad, India . Table 2 summarizes some of the important ICT projects carried out by Chinese, Indian as well as European and American companies in the past several years. The table demonstrates the dominance of Chinese companies with total project commitments in excess of $3.11 billion that translates to a market share of almost 95%.
While some people are not comfortable with the rise of China and its influence as well its strong economic relationships with developing countries, a closer look reveals that the Chinese involvement over the past several years has in fact produced very significant positive developments in Ethiopia unmatched by any other country around the globe. In a discussion with a senior VP of one of the Chinese telecom companies building the fiber optics network, the author was convinced that it is not only favorable and presumably cheap financing that the Chinese companies bring (as the executive admitted), but also a deeper commitment to the project that very few companies in any part of the world can match. For example, it will be almost an impossible task to find US or most European companies sending their own engineers, technicians, and project managers in some of the most difficult terrains in the word with malaria and tse-tse fly prevalence such as the Gibe valley in South west Ethiopia for stays as much as 6-12 months. The only other multi-national companies that have shown such a level of commitment in Ethiopia over the past several decades are the Italian companies like Salini (builders of Gilgel Gibe I-III projects with 2 GWatts of energy), Yugoslavian engineering companies (road building) and in rare circumstances French companies. The Italian company, Salini Construction, needs a special mention here as it has established a track record that matches and - in some cases - surpasses the record of the Chinese companies themselves in promoting Ethiopia's progress .
Table 2. Summary of major contracts awarded to foreign companies in the Ethiopian ICT sector between 2000-2007.
In a recent international event, for example, the local software companies like OmniTech, CyberSoft, and CDS were out in full force demonstrating their talents and products. OmniTech, for example, writes software for enterprise resources planning such as payroll, human resource, and procurement for governmental as well as non-governmental entities. The company has built a strong reputation inside the country around its core expertise in developing some successful products in the areas of financial, human resource and operations management systems. Their clients vary from Gondar and Bahir Dar universities (university admissions and roaster systems) to Metahara sugar factory (resources planning). Looking closer, it is striking to note that the codes - which are written from the ground up locally - are actually as powerful and feature-rich as the expensive enterprise resources management programs like SAP, BEA, and Oracle. OmniTech was, in fact, one of the major sponsors of this international event that drew hundreds of people from all corners of the world. I talked with its executives Ato Deres Tesfaye and his business manager, Ato Mignote Kassa regarding the challenges they have. The common thread that one observes is that these entrepreneurs and technology leaders built their companies from the ground up and built successful business ventures with strong sense of discipline and optimism.
Another interesting company with impressive record and talent is CyberSoft. Talking with the founder and CEO, Ato Tekeste Berhan, one senses an enthusiasm and that creative restlessness common to successful entrepreneurs. Digging deeper, it even becomes more impressive to note that some of the technical challenges the company is working on like a 'conflict early warning system' are actually in par with the research that is being done in top universities in Europe and the US. This model called CEWS uses a combination of genetic algorithm, neural networks, and heuristic models with data such as natural resources, weather patterns, man-made and natural disasters and events and predicts the possibility and probability of conflict in a geographic region that results due to competition for resources. This specific technical research area, one will find out with quick search, is currently being pursued by top US and European academic research groups with backings of millions of dollars. To see an Ethiopian company and Ethiopian software engineers develop such a computer model with direct application to African and global scenario is truly inspiring. The system is being considered, the author was told, for implementation by UN-ECA and AU for conflict predictions in Africa.
Further, a recent development that is pointing to the success of these companies is that they have started competing internationally for projects reminding one that the now formidable Indian software industry started from a similar humble origin. Equipped with bright software engineers and very ambitious and young entrepreneurial technology leaders, these companies may very well point out how big an ICT-based economy the country could generate given encouraging regulatory environments.
In the backdrop, however, our talk with these technology people demonstrated that the biggest hindrance to further progress for these companies seems to be cultural more than anything else. For example, the mind set in the country from customers is that anything foreign, particularly from Western Europe and the US, is far superior than anything produced in the country. As these Ethiopians market themselves to potential customers by presenting far superior technology and far more economical and easily accessible technical support, the average customer (particularly the government) still places premium trust and inclination towards foreign companies. Even if the Ethiopian companies win the project, the offer they get seems to be much lower than the foreign companies. "We need to fight this perception through some innovative means if we want to grow to the next level", said Ato Deres when we discussed this issue.
It seems what the country needs is a critical mass of such successful companies to turn the tide in their favor so that the nascent but capable ICT industry in the country reaches a formidable position that earns respect in the country and the accompanying willingness to offer projects.
Between the frustration of Internet users in the cities and the proliferation of government owned networks in the country, a time for reflecting where ICT in Ethiopia is heading and then making the necessary corrections is due. Time and History - as has become a part of modern Ethiopia's identity - continue to accord the country with so far un-seized opportunities after opportunities. With Ethiopia's population rapid and unparalleled growth and still unrealized potential, the country continues to be seen by international technology companies and development policy makers as a possible significant market for technology, particularly ICT and telecom. If the country had encouraged foreign and local entrepreneurs and investors in ICT and telecom industry, one can only imagine the amount of capital investment the country could have enjoyed.
A discussion of this major national issue with Ethiopian technology leaders, entrepreneurs, engineers, policy-makers and a multitude of stake-holders, the author believes, reveals with no shades of doubt that proponents of freeing the ISP (Internet Service Provider) industry for competition continue to face formidable and unrelenting resistance from the government establishment. Therefore, ten years after the introduction of Internet in the country, the need for arguing and fighting for the removal of government control from this industry continues. The main argument that we put is a choice between well-tested and successful economic models in the ICT sector and one based on a failed command model.
What current practices lead to: The most powerful indication of what current government ownership practices of Internet connectivity have resulted is given by the Internet connectivity chart for Africa shown here. At less than 0.02%, only dysfunctional Liberia surpasses Ethiopia in the distinction of having Africa's and possibly the world's worst Internet penetration rate. Comparison with similar economies (in size as well as structure) like Kenya, Rwanda, Uganda, the Sudan, and Namibia suggests that the reasonably expected Internet penetration in Ethiopia should be between 5-10%. But at 0.25% penetration rate, the country's expected Internet penetration is stunned by a factor of as much as 40! As a result, the country and its people continue to pay huge prices in terms of lost opportunities.
Ethiopian stake-holders as well as independent observers are left to wander how much of a negative effect slow connectivity has on potential economic growth and job creation in the country that could have been enabled by ICT technology. Some of the real concerns that arise due to this serious handicap are:
The Remedy Resisted So Far: While there are a multitude of consequences of current Internet connectivity ownership, the remedy is very simple and straightforward. As experience everywhere in the world indicates, the only proven model is that of a free-market in the ISP space. This simple solution that has been argued in-favor of over the past several years is the sole solution that is on the table.
The reality is, of course, that proponents of government ownership have their reasons for continuing such a model. The arguments often cited are that Ethiopia's conditions are so unique that the farmers which form the majority of the country's population will be left out in a free ISP market. This argument assumes Ethiopia's conditions to be so different from any other country on the face of the earth that it almost places the country as something out of this natural world. Conditions on the ground show that this argument is, of course, outdated, inaccurate and ingenuine. The truth, as any independent observer could see, is that there is a mistrust and fear of wide Internet access that could allegedly be used for political agitation purposes. However, this fear - as shown in the rest of the world including even China - is unsubstantiated and almost paranoid. No government - as conditions on the ground testify - has ever lost power because Internet access is widespread.
In a nutshell, therefore, attitudes have to change - particularly at the government level - where access to Internet by the average citizen is considered something to be feared. If the history of the country itself and the rest of the world is any indication, limiting an access always results in more damage than in any good that may come out of it. Looking forward, the country needs to make a decision between following a path that has so far proved unsuccessful and unsustainable and a correction of path that enables the proliferation of a dynamic sector that could add 0.5-1% points to the country's GDP as demonstrated by other progressive countries around the globe.