The South African Government has launched a drive to provide Telecentres to communities and Internet access to schools. The Telecentres are normally centrally located with respect to clusters of schools and other community services. In the context of this drive, a Telecentre was established in Manguzi, a remote town in the KwaZulu Natal province in South Africa. The surrounding schools did not benefit from this centre due to the inappropriate distance between the schools and Telecentre. In addition, the schools could not be connected to Internet directly due to the absence of telephones. In this case study we will show that existing "off-the-shelf" technologies were not applicable to the specific situation and hence there was a need for a new solution.

There are unusual challenges in providing Internet connectivity to a "sparsely" populated rural community separated by vast distances from nearest urban development. This case study details how we combined existing Internet access technologies to overcome various obstacles such as the lack of existing telecommunications infrastructure, remoteness of area, as well as political and economic issues. Furthermore the solution implemented had to be cheap, suited to the specific regulatory and geographic environment, robust and suitable for a particular application, namely Web browsing and e-mail.

We used the asymmetric nature of the data requirements of the specific applications to our advantage, using radio links and satellite broadcast technology to provide the required connectivity. We will discuss the expected merits of the new solution and its implementation. We will also present our practical findings and discuss how it compared to our expectations.

Similar needs and situations exist in other parts of the world, especially those that have a lack of telecommunications infrastructure, very remote rural areas that are very sparsely populated. We hope that the outputs of this paper can contribute to the technology decisions of people responsible for rolling out Internet infrastructure in similar environments.

Contents

Introduction
Background
Problems and Requirement Specifications
Methodology
Implementation
Findings
Conclusions and Further Work

Introduction

Providing a remote rural community with Internet access can be a challenge at the best of times. If the intended target audience does not have access to telephones or any of the other traditional telecommunications infrastructures,the challenge becomes even greater. The Information and Communications Unit of the CSIR (Mikomtek) did just that in a project in Manguzi, a rural community in South Africa's KwaZulu Natal province. The initial part of the project consisted of the establishment of a Telecentre in the centre of town. The community's desire was that the facilities offered at the Telecentre should be available to the largest possible audience, including the students. However, walking (cars are an extreme luxury, there is no public transport such as busses and trains, most people don't even own a bicycle) the five kilometres to the Telecentre on a regular basis was not practical. At a community workshop we were approached by one of the headmasters with a request to connect his school to the Telecentre, in order for his students to have access to the facilities from his school, eliminating the need to walk to the Telecentre. In this paper I hope to show why the "normal" solutions weren't appropriate in this situation. The challenge lay in devising a cheap, robust and legal solution.

Background

Manguzi

Manguzi is a rural community in the Maputaland region of the KwaZulu Natal province of South Africa. It is situated about 15 kilometres south of the Mozambique border en route to the Ponto Do Ouro border post.

Figure 1: Map of Manguzi and environments

The area is 60 square kilometres in size with about 100,000 inhabitants. The people are very poor, most are subsistence farmers. Maputaland offers subtropical and tropical climate zones, and embraces an ecological diversity - both terrestrial and marine - that makes it one of the most a popular ecotourism destinations in South Africa. It has a largely pristine coastline with coral reefs, estuaries, lakes systems, forests and rugged mountains. In the past it has been sheltered from major human intrusion by the tsetse fly and malaria and has therefore remained relatively sparsely populated and unknown (and underdeveloped). The biggest tourist attraction is the Kosi Bay nature reserve, which is renowned for its remoteness and unspoiled beauty. During the last couple of years, ecocultural tourism has become an important means of sustainable rural development in the area. A number of projects are underway in the area to promote ecotourism among which is the introduction of Information and Communication Technology (ICTs) in the form of a Telecentre.

Telecentre

The community runs various projects, one of which is the Telecentre. The Telecentre was established with the help of the CSIR's Information and Communications Technology Unit (Mikomtek) and has been operational since September 1998. A big problem in the area is the near complete absence of a communications infrastructure. Most homes and businesses don't have telephones and an Internet Café is unheard of. The Telecentre is situated in the centre of town and its purpose is to address this lack of telecommunications infrastructure by providing IT and telecom services to the community. It consists of two parts:

• A phone shop with five telephone booths and a fax machine; and
• The IT component,which consists of a local area network with eight Windows 98 personal computers and a FreeBSD file server. It is connected to the Internet with a dial-up analog connection, which dials on demand, as a user on the network requests Internet access. They also provide services such as word processing, scanning, printing, photocopying, etc.

Since its establishment the community has actively used the services offered by the Telecentre.

Problems and Requirement Specifications

Recognizing the potential of the Internet as information source, Mikomtek was requested by a headmaster of one of the schools in the area to connect his school to the Telecentre to enable his students to utilize the facilities there, specifically the Internet.

In a well-developed telecommunications infrastructure the solution would have been trivial: install a modem, open an account with an Internet Service Provider and you're connected. In a remote rural area, far removed from urban development and telecommunications infrastructure, this request was not so simple, for the following reasons:

1. Underdeveloped telecommunications infrastructure:

• The vast majority of homes and businesses in the area do not have telephones. A couple of public telephones are available in the town itself, but due to the absence of transport, are not freely accessible to anybody outside the town. Imaging having to walk many hours or even days to make a telephone call!
• Although cellular coverage is available, it is very patchy and unreliable. Due to the high cost associated with ownership it is not an option for most of the local inhabitants and exists mainly to serve tourists.
• High bandwidth services such as ISDN, VSAT and leased lines are not available.

2. Of the 71 schools in the vicinity of Manguzi only three have electricity and none have telephones available to connect to either the Internet or the Telecentre.

3. The funds available for installation of the required infrastructure were extremely limited. The reasons for this were that unemployment stands at 85% and the area is predominantly agriculturally-based and under a tribal authority.

4. The solution therefore, had to be cheap and preferably not involve recurring monthly costs.

5. For a variety of reasons, which will be explained later, none of the traditional telecommunications infrastructures were suitable.

6. Telco (Telkom) monopoly.

7. In South Africa, rural tribal authority politics combined with our particular legislation and historical inequalities in access provision makes for an interesting and risky mix.

Methodology

1. Obtain community buy-in and co-operation.
Even more important than choosing the most appropriate technology is the co-operation with and buy-in from the community. Together with the community a pilot project was launched with the aim of exploring the various options available to provide Internet access to the schools.

2. Identify schools as partners for the project.
Two schools - Shayina Secondary School with 1,002 students and Maputa Senior Primary School with 450 students - were nominated to participate. They were ideal for a couple of reasons:

• Both have electricity;
• They are close enough to the Telecentre (3 - 5 km) to make for easy access to both sites during the installation and testing phases; and,
• Despite being relatively close to the Telecentre, the schools did not have line of sight to the Telecentre, enabling us to properly test the solution installed.

3. Identify possible solutions and test suitability.
In deciding how to connect the schools to the Internet, we explored the various traditional methods available (more detail later in the paper) and came to the conclusion that none of these were suitable. The first choice would have been to install access technology directly at the schools. This was not possible either because of limited financial resources or because the technology was unavailable. The Telecentre had a dial-up link to the Internet and the required infrastructure (account with an ISP, modem, file server, trained manager capable of providing the necessary support) and it was decided that the easiest option would be to utilize this link to provide Internet connectivity to the schools. We then had to find a way to connect the schools to the Telecentre network.

4. Establish relationships with relevant partners.
This was largely dictated by the technology chosen.

• Maputaland Development and Information Centre (MDIC): Community organization and owner of the Telecentre;
• WBS: Existing Mikomtek partner. Black empowerment company (necessity to conduct successful business in post apartheid South Africa) in possession of a license roll out a wireless network.
• Siyanda: Satellite Broadcasting ISP. Already involved in providing Internet to rural schools.

5. Technology options and considerations.
We investigated various options to connect the two schools to the Telecentre. These options with associated comments are given in Table 1.

Option	Comment
Telephone lines	Not available
Cellular telephone	Coverage not ubiquitous. Reception very unreliable.
Two-way VSAT	Installation and monthly costs too expensive.
Spread spectrum radio solutions	Requires line-of-sight but the two schools are not visible from the Telecentre. License required.
ISDN	Not available
Satellite Internet Broadcast	High-speed Internet downloads via satellite uses telephone line for the back channel to the Internet. No license required.
Low Frequency Radios	Normally used for telemetry - are limited to very low bit-rates. Attractive option because a partner company had a license and line-of-sight isn't a problem.

Table 1: Link options investigated to provide Internet connectivity

An important consideration to keep in mind in the choice of solution was that in South Africa, the Telco (Telkom) has a monopoly on fixed telephony until 2003. This results in most independent and innovative solutions (especially wireless solutions) running foul of legislation.

Implementation

As explained earlier, the Telecentre's dial-up analog connection was the only available "entry point" into the Internet. In some way the schools had to be connected to the Telecentre in such a way that they could access the Internet on demand. The solution we implemented combined a variety of "off-the-shelf" solutions, which on their own were not suitable.

Radio only solution

Mikomtek has a relationship with a black empowerment company with a South African Telecommunications Regulatory Authority (SATRA) approved license to operate a mobile data network in the 420 MHz frequency band. We decided to investigate the feasibility of using their radios to connect the schools to the Telecentre. If someone at the school is interested in browsing the Web or reading e-mail, the Unix server at the Telecentre dials on demand. We provided each school with a PC, radio and antenna (Yagi). At the Telecentre one of the Windows PCs acted as router. This PC was connected to the Telecentre Local Area Network and also to a radio. Figure 2 provides a diagrammatic representation of the network. The radios emulate a network card, enabling us to run TCP/IP over the link. We managed to connect the schools to the Telecentre and also access the Internet.

Figure 2: Configuration with radios only

We ran into a problem however. The radios used were telemetry radios, with a maximum bandwidth of 4200 baud. This configuration was suitable for e-mail but not for Web browsing. We had to find another way of downloading Web pages to the PCs at the schools.

Radio combined with satellite broadcast

The asymmetric communications requirements of Internet applications could be used to solve the problem. The eventual solution deployed combined Satellite Internet Broadcasting with the radio network. The radio link via the Telecentre is used for the uplink path (in place of a telephone line) and the satellite is used to download Web content directly to the PC at the school. Figure 3 provides a diagrammatic representation of the network combining the two technologies. Satellite receivers are usually capable of receiving data at much higher rates than what is possible via normal telephone lines. Siyanda was selected as the Satellite ISP. Siyanda makes use of the PAS-7 satellite whose Ku-band footprint covers the whole of Southern Africa. Their data receiver is compliant with MPEG-2 DVB digital TV. The ordinary 90 cm Digital Satellite Broadcasting (DSB) satellite dishes are used for reception. Requests from clients are sent to Siyanda via Virtual Private Networks (VPNs) over MWeb's (a local ISP) terrestrial infrastructure. Distribution of data is this fashion is permitted under the Broadcasting Act of 1999.

The equipment required at each site is as follows:

1. Schools	2. Telecentre
Radio	Radio
Yagi antenna	Omni-directional antenna
Satellite receiver card	Connection to Telecentre LAN
DSB dish

Figure 3 depicts the equipment used and provides a short description of each.

Figure 3: Equipment used

Figure 4: Combined radio and satellite broadcast network

Here is an explanation of how the system operates. Refer to figure 4 above.

1. A pupil at Shayina Secondary School wants to access a Web page.
2. The PC forwards the request to the PC acting as radio base & router at the Telecentre, using the low speed radio link.
3. The radio base & router forwards the request to the Unix file server.
4. Unix server makes a dial-up connection to the closest ISP POP: Siyanda makes use of Virtual Private Networks (VPNs) to tunnel their private network's traffic across the public network of M-Web (a local ISP).
5. The Unix file server forwards the request to a VPN server on the Siyanda network. At the same time it loads a virtual adapter that connects to the VPN server on the Siyanda network. The Unix file server needs to become part of the Siyanda network before it can forward data requests to the Siyanda network.
6. The VPN server receives the request from the Unix server and forward the request to the Siyanda proxy server.
7. The proxy server requests the Web page, somewhere in the Internet, on behalf of the PC at Shayina.
8. The Siyanda proxy server receives a response from the Web server, and forwards the response to the requesting PC at the school. The response delivered using the high-speed satellite link, directly to the PC.

The total cost including personal computers are estimated at $2,300 for the Telecentre and $3,000 per rural site. Refer to table 3 for a breakdown of costs. The advantage of this solution is the fact that the bulk of the costs are a once off. The recurring monthly cost is minimal ($40). In the case of Manguzi, an arrangement was made between the schools and the Telecentre that these costs will be funded from the profit made by the Telecentre.

Item	Unit Cost	Comment
DSB Dish	$200	At least one required per site
Usage cost (100 Mbytes download)	$40 / month
Radio modem	$900	One required per remote site and a shared one for the Telecentre. (Adding more sites does not require additional equipment at the Telecentre)
Installation	$400 - $1200	1 - 3 man-days per site needed to determine suitable sites for antennas, positioning and fixtures, etc. Variable "rural" surprises may also feature

Table 2: Cost of installation

Findings

• The main reason for the introduction of Information and Communications Technology (ICT) to Manguzi was to facilitate access to opportunities and information. It is succeeding in this mission. The members of the community are actively using the facilities offered by the Telecentre to do word processing, desktop publishing, send and receive e-mail, surf the Web and perform numerous other activities.
• In the configuration as for the pilot project - which consisted of one PC at each school connecting to the Internet - the performance can roughly be compared to the experience of an Internet user with a 9.6 kbps dial-up connection. This is much slower than the 30Kbit/s expected and can probably be attributed to the extremely slow uplink.
• Training teachers in the utilization of ICT in education accompanied the introduction of the technology to the schools. This was done in cooperation with the KwaZulu Natal Education Department and SchoolNet SA (a NGO with a Government brief to introduce Internet into South African schools). Teachers, who had never seen a PC before, moved to basic Web literacy over a week. However, skills degenerate if there isn't regular use and therefore, professional development of teachers using the Internet has to go hand in hand with establishing professional networks and support groups between teachers in Manguzi and elsewhere.
• One of the biggest hurdles to overcome during this project was not technological in nature, but related to community support and the politics surrounding access provision. In South Africa, rural tribal authority politics combined with our particular legislation and historical inequalities in access provision makes for an interesting and risky mix. It is important to introduce technology with the full buy-in and understanding of the community. Everybody concerned should be realistic about what can and cannot be achieved with the introduction of ICT; it is not an instant cure all ills.
• Technical support for this project was a major challenge. Because of the remoteness of Manguzi (minimum 8 hours by car, or 2 hours by airplane) it was not possible to "jump in the car" and go to the site when something went wrong. The project had to be planned in the finest detail and thoroughly tested before being rolled out at Manguzi. Furthermore it is important to have a person on site that is able to do at least first line support. The solution (hardware and software configuration) must be extremely robust and able to stand up to the rigors of rural life.

Conclusions and further work

This project proved to be a bigger challenge than initially expected. The team learnt a lot about innovative use of technology, but even more about the social and cultural aspects that accompany the introduction of ICT in rural South Africa. We also saw the difference access to information can make in peoples' lives. This project was limited to one PC at each of two schools. The schools are also quite close to the Internet access point. Developments are underway to increase the number of PCs supported at the remote site, the distances that can be covered and performance. Rollout will be in a cellular fashion with "base stations" - which will serve as the Internet access point - each with their own "outstations".

About the Author

Ronel Smith is the Manager of Connectivity Business at e-tek (Pty) Ltd, the commercialization vehicle of CSIR Information and Communication Technology. The aim of this group is to be a "vibrant hothouse for new digital start-up businesses". Ms. Smith has ten years of experience in IT and telecommunications with a specific interest in broadband communications. She has been involved in projects ranging from the design and implementation of a wireless rural network to a large ATM network at one of South Africa's largest universities.
E-mail: ronel.smith@mikom.csir.co.za

---

Editorial history

Paper received 8 September 2000; accepted 27 September 2000.

A version of this paper was presented at INET2000, Yokohama, Japan, in July 2000.

---

Copyright ©2000, First Monday

Overcoming Regulatory and Technological Challenges To Bring Internet Access To a Sparsely Populated, Remote Area by Ronel Smith
First Monday, volume 5, number 10 (October 2000),
URL: http://firstmonday.org/issues/issue5_10/smith/index.html