Electronic Networking for West African Universities

Introduction

The American Association for the Advancement of Science (AAAS) and the Association of African Universities (AAU) jointly held a Workshop on Electronic Networking for West African Universities, in December 1993, in Accra, Ghana. The purpose of the workshop was to focus on opportunities for West African universities to participate in the many academic benefits afforded by electronic networking technologies. Such benefits include fast, efficient, reliable, and cost-effective communications, and, for the researcher, especially, unprecedented access to colleagues and information resources throughout the world. Indeed, in the United States, Europe, Japan, and elsewhere, electronic networking is truly changing the way science is done, facilitating collaborative efforts and putting powerful applications and resources literally at the fingertips of the individual researcher working at a personal computer (PC). The Internet has evolved into far more than a communications medium; it has itself become a vast source of information, an essential academic tool that is increasingly becoming an integral part of universities' infrastructure. One workshop participant put it plainly, stating that "if our African universities fail to connect to the electronic network, we will have to face the fact that we will be awarding degrees of lesser quality."

Impressive strides have already been made in the adoption of electronic networking in Africa, particularly in the eastern and southern regions of the continent. The status of existing African networking projects was the subject of a previous workshop (the 1992 Nairobi Workshop on Science and Technology Communication Networks in Africa), cosponsored by AAAS and the African Academy of Sciences, at which participants detailed the challenges encountered and lessons learned during their network-building experiences. Among the phenomena clearly noted during that workshop was that progress in electronic networking had been significantly slower in West Africa than in the eastern and southern regions of the continent. Therefore this followon workshop was intended to contribute to filling this development gap, in part by facilitating interregional cooperation toward formulating networking strategies for institutions in the West.

Some 35 participants were invited to the workshop in all, including "teams" of vice chancellors and computer experts from eight selected West African universities with active graduate programs, experienced resource persons from existing eastern and southern African networking projects, and representatives of other current and pending network initiatives of relevance to the region. Our goal was to review the technical options for networking from a comparative cost/benefit standpoint, and, more importantly, to identify the human and organizational factors that have been found critical to the success (or, conversely, responsible for the failure) of networking projects to date. The participation by both high-level university administrators and people with technical expertise allowed the workshop to cover key policy issues as well as provide some brief training on actual installation and use of networking software (which was also distributed to university participants). From this collective assessment, West African university participants were able to fortify their plans for the establishment or improvement of electronic networking in their respective home environments.

Technologies and Strategies

It is widely agreed that the ultimate goal in networking is to have access to the "proper" Internet, which uses a protocol suite known as TCP/IP (Transmission Control Protocol/Internet Protocol)<$FThe "proper" or "capital I" Internet, actually a set of interconnected network all running TCP/IP, is often differentiated from the "little i" internet, a broader term that includes connected networks running other protocols, such as UUCP or Fido (explained in more detail elsewhere in the report).>

In order for an institution to serve as an Internet host, however, it must lease a data line to connect to the rest of the Internet. Leasing such a line involves significant costs, especially in Africa; the University of Zambia estimates it will require a minimum of US$75,000 per year (including staff, etc.) to lease a line to the nearest Internet connection, in South Africa (see Neil Robinson's paper). The cost of a leased line is a fixed cost (no additional usage-based charges), which begins to make economic sense only after a certain "critical mass" of users exists to share it. In sub-Saharan Africa (other than South Africa), where there are still relatively small numbers of network users, Internet connectivity has not been economically feasible to date although some nodes (central, mail-routing computers), such as Zambia's, are now approaching the point where a leased line may indeed be cost-effective.

Fortunately, there are lower-cost alternatives to leased lines and the Internet, technologies that allow basic network services like e-mail and provide vast improvements over other, more traditional communications options such as fax, telex, and the postal system. These alternatives are based on "dial-up" access over normal phone lines to computers that collect messages and, in turn, distribute them in a similar manner to other computers as necessary until they reach their final destinations. Networks that send messages "hopping" from computer to computer are known as "store-and-forward" systems, and include UUCP (Unix to Unix Copy Program) and the PC-based Fidonet. It is important to note that these are not "low technology" systems; indeed, the software that instructs messages to be "packed" together, compressed, and sent complete with error detection and correction all features that contribute to the low cost of the systems is quite sophisticated.

UUCP (and its PC-based, Unix-emulating version, UUPC) and Fido protocols represent the two most popular store-and-forward systems throughout the world. Fidonet is the more widespread of the two in sub-Saharan Africa, probably because of the relatively greater diffusion of PCs as opposed to Unix machines on the continent, and the fact that the PC version of UUCP, i.e., UUPC, has only been made available relatively recently. There are significant and growing numbers of Fidonet nodes and network users throughout Africa, particularly in the eastern and southern regions (again, for details see Robinson's paper, and Lishan Adam's, as well as the User's Guide to Academic Networks in Africa, available from AAAS). In several nations, mostly those in close proximity to South Africa, a UUCP-compatible "point" (end-user) software known as SNUUPM (see Mike Lawrie's paper), developed at Rhodes University, has enjoyed recent gains in popularity.

The Fido- and UUCP-compatible systems are fairly similar in terms of what they are basically able to do and with respect to the user interface, regardless of their underlying technical differences. Nonetheless, each system has its proponents, and some important differences between them may be gleaned, although these depend to an extent upon which particular version of either system one is considering. Because of the variation among versions of software, and the fact that new versions tend to continually and rapidly proliferate, it is difficult to say with any authority what particular features password protection, ease of installation, number of users supportable, and so forth one "camp" boasts over the other. The papers by Lawrie, Lishan, and Robinson included in this volume explore much of this territory in some detail.

However, two key points, at least, can be made with confidence:

1. Fido software is superior in terms of handling poor phone lines, because of its sophisticated data compression and error detection/correction features.

2. UUCP software is compatible with and upgradable to Internet standard upon acquiring a leased line, whereas Fidonet uses an entirely different protocol that is not compatible with the Internet standard. Therefore the jump to Internet from UUCP would require very little retraining for system operators.

Neither point makes an ironclad case for either software. One may argue that Fidonet is better suited for Africa, given the still shaky telecommunications infrastructure in many countries and the relatively greater amount of available expertise and support on the continent. One may argue equally well that it is preferable to start in the UUCP world, given its promise of a smooth transition to the inevitable Internet once the number of users justifies a leased line.

Which costs less? Usually Fidonet, depending on such variables as the particular modem used, because of its highly efficient transmission of data, but not by very much. There is public domain (free) software available for both systems. For both systems, a computer, a modem, and a node with which to exchange messages are all else that is required to start communicating with the internet.

There is also one additional requirement for an institution setting up a new network node (as opposed to connecting as a point or points off an existing local node). Certain nodes operate "gateways" whereby e-mail can flow between different types of systems in a way that is invisible to the end user. For example, a Fidonet user in Ghana can send a message to an Internet user in Washington, DC, via GreenNet, a London-based organization that serves as a major international gateway/mail router. Thus for international mail, particularly that which is destined for recipients on systems that differ from the one the sender is using, it is also necessary for a new node to make some mail-exchanging arrangement with an organization like GreenNet, or Rhodes University, which operates a similar gateway .

Ultimately the choice of a particular technology will probably be determined by convenience and local conditions, such as what is most readily available in terms of technology and expertise. A general rule of thumb was offered by one of the workshop's technical resource people: UUCP is preferable where a Unix host is already in place, and Fido is preferable otherwise.

But whether one chooses Fidonet or UUCP is really not nearly as important as creating an environment in which networking can flourish. This enabling environment involves such elements as ensuring proper training (for end-users, system operators, and trainers themselves), encouraging system use, providing reliable service, marketing to attract new users, and, sooner or later, developing a workable method of charging for services (for non-university users, an issue that is discussed in greater detail below).

Discussions among participants during the Ghana workshop yielded a "minimum checklist" of necessary conditions for a university e-mail connection point . This checklist is intended to provide some basic guidelines concerning the basic elements without which a university e-mail connection will not be possible. Note that the minimum level of e-mail connectivity is as a point dialing into an existing node, which usually means that the point relies on the node to transmit messages (the exception is that point software often includes the capability to dial a final destination directly, bypassing the node).

Some of the requirements listed in the checklist imply a somewhat greater level of connectivity. For example, it is not strictly necessary to have an international direct dial (IDD) line for e-mail, if one routes one's mail through a local node. The IDD line allows direct connection to a gateway or other remote destination, necessary for a node, but really just an additional (useful) option for a point.

Minimum conditions notwithstanding, workshop participants were in agreement that universities would do well to adopt a more ambitious goal e-mail access, and eventually full Internet access, for the entire university community. The ideal set-up to provide community-wide access for the university would involve the establishment of a local area network (LAN) through which network services could be provided. Therefore installing a LAN is an important and worthy goal that would without a doubt greatly facilitate the provision of network access for the entire university, although it is not by any means a prerequisite for establishing a viable university node. It was also noted at the workshop that e-mail capability, LAN installation, and full Internet access can and should be pursued in parallel, not sequentially.

Other principles of university network building also became evident during the workshop discussions:

o The objective of e-mail at the university is not to allow just for a few messages to be transmitted, but to establish strong, regular links between the university and the global academic community; all students and staff should have access to e-mail on a system that allows fast turnaround time for messages.

o To fail to establish network connectivity is to ensure an ever-widening "information gap" between unconnected institutions and the networked world.

o E-mail must be free to the university end-user or else it will not take root and grow. This requirement has policy implications:

     
	* cost control:  universities have communications budgets,
	and e-mail reduces, not increases, communications costs,
	as long as e-mail displaces other, more expensive forms
	of communication.

	* usage:  encourage, don't restrict, use of the system.
	People will need to play with the system in order to
	familiarize themselves with it and its capabilities. The
	fewer regulations, the better.  Increase internal access. 
	Do not charge on-campus individuals or departments.
	
	* outside access:  providing e-mail service for a fee to
	private-sector organizations, and to governmental and
	nongovern-mental organizations, can be a good way of
	subsidizing the operation of the network for academic
	purposes.

	* administration:  a system manager who has links to senior
	university administrators can be useful in establishing/
	promoting a university network.  Neither system manager
	nor system operator needs to be a full-time job, but can
	be additional functions of existing positions.

o E-mail needs to be promoted, through such means as disseminating information about its desirable features (speed, cost-effectiveness, and so on), allaying fears (privacy, difficulty, newness), and locating key users, or "champions," to spread the word. Above all, the system must be made reliable and easy to use. If these conditions are met, communities of common interests will grow, and the "will to communicate" (see Robinson) will increase. The inevitable result will be the expansion of academic resources colleagues and information and the knowledge base. The promotion of e-mail on campuses is a subject that is covered in depth in the background papers, particularly those by Robinson and Lawrie, both of whom have considerable experience with networking in a university environment (Zambia and Rhodes, respectively).

o High-level leadership is essential to bring networking to the entire university community and integrate its use into ordinary research and communications. Many Vice Chancellors at the workshop echoed the belief that network access has become a defining characteristic of the modern university, and that it is therefore imperative that connectivity be considered a priority of the highest order.

o The time frame for establishing a network node is a matter of months, not years. Indeed, installation and configuration of the necessary software generally takes a half an hour or less, so given a computer, a modem, a phone line, and a potential message recipient, it is no exaggeration to say that e-mail service can be rendered within an hour. The longer estimate takes into account further training and the setting up of additional connections to the system.

o Donor support has been, and will continue to be, extremely useful in helping to get networking projects off the ground. It is often difficult, even in the best of circumstances, for any new technology to gain acceptance, and assistance with start-up and operational costs is critical to reducing resistance to changing old habits and learning new skills. Sustainability is primarily a function of a sufficient user base; continued donor support in the short term will open the door for many people to discover a technology that they will likely soon find indispensable.

West African Universities

The eight West African universities represented at the workshop submitted in advance brief summaries of their respective overall telecommunications capacities and prospects. With some variations, a general picture of a "typical" university emerges from these reports: Relatively few computers, concentrated in a handful of departments; few phone lines, often of poor quality, and some radio links; scarcity of modems; very limited, if any, networking a few universities reported small LANs (e.g., two or three computers linked in a computer center), and a couple of universities have a small number of e-mailboxes through RIONET (a global network of the French research organization ORSTOM, which has nodes in several West African countries); in some countries, digital data networks exist, but are usually expensive to use. Each university, however, has some computer and communications capability the seeds of growth and a demonstrated commitment to initiate or expand network connectivity.

By the end of the workshop, most had formulated or solidified plans to move toward this goal of greater connectivity. As one of the university representatives put it, one measure of the success of the workshop is that participants had quickly "gone from asking `why do we need this?' to asking `how do we get this?'." The workshop's technical resource people did a superb job of offering answers to both questions.

Some examples of specific workshop outcomes from the universities themselves:

o University of Ghana-Legon reports plans to hook up with GHASTINET (a Fidonet node in Accra) using Fidonet; possibility of three Ghanaian universities forming network with one satellite uplink

o University of Ibadan reports phone lines are good, modem will be acquired, hopes to have Fidonet set up within months; high-level assurance of promotion of e-mail at university

o University of Cape Coast in midst of plans to connect to HealthNet (a satellite-based network for health information); considering Fido link in Ghana or UUCP link to RIONET in Abidjan; will advertise e-mail in campus newsletter

o University of Sierra Leone has LAN in place, but plans to set up a stand-alone Fido host at first because of available expertise; considering later UUCP connection via RIONET; good phone lines, digital phone system in place; need for modems and system operator training noted

o University of Cheikh Anta Diop in Dakar already connected to RIONET; excellent telecommunications system, X.25 (digital data) lines in use; notes that university still far from leased line/full Internet connectivity, but hopes to extend existing access to more researchers and the wider university community

o University of Ouagadougou also already connected via RIONET, but limited to few individuals; plans to generalize access; need for more modems and "sensitization"

o University of Yaounde computer center connected to RIONET and IRISA of Rennes through an X.25 link; plans to connect all university departments, then advertise/connect to other five universities in Cameroon, as well as international organizations located in the country. Yaounde reports that progress has been made since the workshop towards connecting university departments, through a French-sponsored program called PAESRUC; optical fiber cables have been installed and other necessary equipment has been ordered.

o Other organizations represented at the workshop reported on their activities and noted that various forms of help are available to universities and other institutions wishing to achieve connectivity. Examples include:

o ORSTOM's RIO network has already been noted, for example, and is open to allowing access to its well-established system. RIONET already has some 270 network users at its West African sites, including Bamako, Brazzaville, Dakar, Ouagadougou, Yaounde, and others. RIONET uses UUCP technology.

o RINAF, a project of UNESCO and the Italian government, has been working to bolster existing nodes in Africa and try to move them towards full Internet connectivity. A UUCP-Fidonet-Internet gateway will be set up in Pisa, and even countries/nodes that are not an "official" part of the RINAF project will be able to dial the gateway or get polled on a cost-recovery basis.

o HealthNet, also noted previously, is a network based on digital radio and a low-earth-orbiting satellite, and is focused exclusively on the exchange of health-related information. It is administered by a Massachusetts-based organization called SatelLife. Several groundstations serving as hosts to institutions in Africa are reported, although estimates of numbers of users are not available.

o The project for Capacity and Infrastructure Building in Electronic Communication in Africa (CIBECA), a joint effort by Canada's International Development Research Centre (IDRC) and the Pan African Development Information System (PADIS) of the UN Economic Commission for Africa, is currently being launched and will provide training, documentation, equipment, and support for selected Fido-based network sites in Africa.

o South Africa's Foundation for Research Development (FRD), which administers UNINET and has for years provided support for several nodes in the region, continues its commitment to furthering electronic networking in Africa.

The Association of African Universities (AAU), cosponsor of the Accra workshop, has a strong interest in electronic networking as well, both for its own communications needs and as a part of its mandate to encourage communications among African universities, and to strengthen research and graduate training. AAU notes that its own communications bill, including large expenditures on faxing, could be substantially reduced by greater reliance on e-mail. Also noted by AAU at the workshop is the encouraging sign that telecommunications systems in Africa are improving steadily, and are greatly improved over the situation as recently as five or six years ago in many places.

The AAU has some funding to set up a small pilot networking project, comprising a number of West and Central African universities and the AAU Secretariat at the outset, and is beginning a feasibility study of its various options regarding details of connectivity. AAU also has the capacity to provide technical training and other support services, other than supplying equipment, for member universities.

Conclusion

Electronic networking, and in particular the Internet, cannot be ignored. The Internet has already profoundly altered the way millions of people in the world communicate, personally and professionally, and the way they gain access to information. Its expanding variety of powerful services and rapidly growing user base make it overwhelmingly likely that it will become an ever more integral part of everyday life for professional users and providers of information. Researchers and others in academia will continue to have a particularly compelling need for Internet access.

Store-and-forward networking technologies like Fidonet and UUCP provide a low-cost alternative to full Internet access. Universities in Africa will most likely wish to consider these as interim technologies, ways of participating in some of the basic and significant advantages of electronic networking, while building up the user base to the point where an Internet-enabling leased line becomes cost-effective. Internet access can be actively investigated even while interim systems are implemented.

The AAAS User's Guide to Academic Networks in Africa contains more detailed information on what already exists in terms of networking on the continent, and how to participate. The contact people listed therein represent an excellent starting point for those seeking expert advice and assistance from nearby colleagues.

The primary lesson from the Accra workshop, as noted by Professor George Benneh, Vice Chancellor of the University of Ghana-Legon, in his concluding comments, is that African universities "cannot afford not to get connected" to the world's electronic network. He further urged his fellow university administrators to support efforts to actively expand the network user base, to obtain funding for the necessary equipment, and to solicit government support to facilitate this usage of the telecommunications infrastructure by universities to strengthen the knowledge base that is so essential to meeting development goals. There have been few, if any, similar opportunities to reap rewards of such magnitude for such a small investment.