Abraham Lakew

PRC Inc. (Advanced Technology Division), McLean, VA

Introduction and Overview.

The Ethiopian communications/networking effort is at a crossroads. The level of awareness and demands have increased and the technology, while far from mature, is much further along today than it was when it was fielded just two years ago. Discussions of varying degrees, product implementation, modest increase in user sophistication, and the rising participation of expatriate computer professionals and networking engineers, should act as landmarks as we proceed from this crossroads into the future.

The need to introduce state-of-the-art communication and computer technology has been the subject of several discussions aboard Cleo/EEDN; the latest being - full Internet access and connectivity and the desire to have Ethiopia linked and be part of the much-talked-about, Information Superhighway. Although there is so much that remains to be done, parts of the discussions have actually reached fruition through the efforts of the first AAU Networking Committee. The networking team, together with the help of the Cleo/EEDN community has already provided AAU with a fully-loaded Unix machine and several modems. As a result, AAU is now enjoying e-mail access to the Internet. Subsequently, individual and group efforts to upgrade the network capabilities have also been attempted.

The strategy for the networking effort should be able to consider first, assessing and evaluating existing networking capabilities (i.e., hardware, software, network topology and protocols, etc.), existing telecommunications infrastructure of the country (with scheduled upgrades, if any), and other logistic and administrative issues. The prevailing wisdom is that cost-effective, and realistically implementable approaches should dictate the progress of the efforts.

To be practical, short and long-term goals should be identified and implemented accordingly. The strategy should also consider alternative communication technologies that are ideal for developing countries; especially, technologies that would help link development and relief- related projects in rural Ethiopia.


The following information (Hardware/Software Configuration) is presented to provide an understanding and to get an overall picture of the computing and networking environment of some of the faculties and departments at AAU. It is also necessary to note that this information is by no means complete, and is about two years old.

Faculty of Technology (Amist Kilo):

PCs: about 25 - a mix of PS/2 Model 30s (21), PS/2 Model 70s (2), HP micros(2).

Network: IBM LAN Server ; was not operational in 1992.

Topology: IBM Token-Ring.

Programming Languages: FORTRAN 77, PASCAL, TORBO C, C++

School of Information Systems for Africa (SISA) at Sidist Kilo: This is a regional school that offers MS degree in Information Systems.

PCs: about 25 (a mix of Tandons - fully IBM compatibles and PS/2s) some with internal CD-ROM drives.

Printers: Dot-matrix, HP LaserJet Series IIs

Plotter: one HP 7550A on the network using up to 8 different color pens. Works on A3/A4 size plotting paper and transparencies and an image scanner: for scanning photographs and text documents.

Application Softwares and Programming Languages.

Spreadsheet: applications such as Lotus 1-2-3.

Database: dBase III Plus, R-BASE, Framework III.

Statistical: SAS, SPCC-PC.

Project Management: Super Project Expert, Harvard Total Project Manager.

Graphic Packages: Harvard Graphics, Freelance Plus, Paintshow, SAS/Graph.

Utilities: Sidekick, Sideways, Norton Utilities, PC Tools, Laplink III, Map Assist, LAN Assist Plus (3.0), Checkit, QT, Panscan, Scan Do, OCR, CC:Mail, Easy Learn, QRAM, Printaplot, and Simulation - Teahing Aids.

Programming Languages: MS Basic 7.0, MS Fortran 5.0, MS Macro Assembler 5.1, MS Quick C Compiler with Quick Assembler 2.1, Turbo Pascal 5.5, Turbo Prolog 2.0, MULISP

This is probably the most organized computing environment. They have a Novell Netware 2.15 LAN (Network Topology: Token-Ring) supporting different instructional PCs scattered within a limited area of their premise sharing printers and disk space. They also have modem connection to PADISnet, a Fidonet node at ECA building.

IDR/DTRC - Demography Dept (Sidist Kilo Campus):

About 15-20 PCs (Gateway 2000s, Compaqs and other IBM compatible clones) mostly 286s and 386s. File Server: a 486/33 MHz with 8 MB RAM; NOS is Novell Netware 3.11 and the topology is bus, using Ethernet cards on each PC.

Applications: most of the office automation packages, statistical, and some more packages that are specifically for research purposes.

Faculty of Business & Economics (Sidist Kilo):

Only two 386s in 1992; expecting about ten 486/33 MHz PCs for the Faculty, and sixteen 386/25 MHz PCs for Economics Department.

Science Faculty (Arat Kilo):

Chemistry and Physics departments are effectively using standalone PCs (current configuration details not available).

Maths Department (Arat Kilo): has several PCs (details and current configuration info not available).

AAU Computer Center (Sidist Kilo).

The Center has a slow NCR mainframe as its main processor. Connected are dumb NCR terminals. Applications range from processing of academic records, administrative, payroll processing, etc. The center has plans to convert and migrate to a PC-based networking environment.

The above information is collected by the author during his TOKTEN assignment to AAU in 1992. TOKTEN (Transfer Of Knowledge Through Expatriate Nationals) is a program that is fully funded and sponsored by UNDP. In order to devise/design and ultimately implement a national or a community-wide networking strategy, various approaches and options have to be explored. To be practical and accomplish this task effectively, however, short and long-term goals should be identified and then, implemented accordingly. The strategy should also consider alternative communication technologies that are ideal for developing countries; especially, to link development and relief-related projects in rural Ethiopia.

Note: implementation of alternative options may not be necessarily in the order they are presented below. It is conceivable that what is listed as long-term may be accomplished in the short-term and vice versa. It is also possible that concurrent implementations are inevitable.

Short-Term Options:

1. Integration and expansion of local area networks (LANs).

From what we know, (and as stated under Background, above) there are an appreciable number of standalone PCs scattered in the university community and some major cities of Ethiopia; and that LANs are not widespread and that their benefits are not fully explored. Hence, in the short-term, it would make so much sense, to encourage the continuous installations of LANs. This will not only allow users or workstations to share files, disk space, and peripherals, in the short run, but will make future connectivity/internetworking to other LANs, WANs, and remote networks probable.

2. Setting up bulletin board services (BBSs).

BBSs present end-users with little or no experience a unique opportunity on what electronic messaging is really about. BBSs are not only very practical and economical, their applications are also limitless. Setting up a BBS is the simplest and very cost-effective means of electronic messaging with file transfer (download and upload) capabilities. Most BBS systems are designed to be intuitive and menu-driven; with full message screen editors, built-in internal file transfer protocols (such as Xmodem, Ymodem, Zmodem, Kermit,and many more), and the ability to attach files to messages. In the long run, the file transfer feature of BBSs will also develop to a substantial level where the demand for a more powerful FTP will be imminent.

A typical BBS server configuration will only require a PC, modem, a telephone line, and a BBS software. At the subscriber's end, a PC, modem, and a communication software is all that is needed to access the BBS. Future installations and usage of BBSs will also complement Fidonet's configuration which is nothing but a series of BBSs tied up together creating a global network. Currently, Ethiopia is connected to the Internet via Fidonet.

3. File transfer (sharing) using external storage media.

Transferring files and programs using CD-ROMs, write-once, read-many (WORM) optical disks, removable magnetic hard disks, back- up tapes, diskettes, etc., can also be quite useful as a temporary solution. At first glance, this alternative solution may seem very trivial; but in a country with limited technology and resources, it can be an effective ad-hoc solution.

In addition, other possible moves that need the attention of everyone concerned are:

--utilization of the Unix machine (Faculty of Technology at Amist Kilo) i.e., using it to its potential capacity not only as a mail server, but as a gateway to other LANs, WANs, and the outside world-the Internet.

--securing a leased line if cost is not prohibitive.

--studying the possibility of using a commercial Internet access provider.

Long-Term Options.

1. TCP/IP (Transmission Control Protocol/Internet Protocol) connectivity.

The first phase of this strategy considers TCP/IP connectivity. TCP/IP (the defacto standard protocol in) is a networking protocol that provides communication across interconnected networks made up of computers with diverse hardware architecture and various operating systems.

This emerging technology has made it possible to interconnect several disparate physical networks and make them function as a coordinated unit. This methodology overrides the details of network hardware and permits computers to communicate independent of their physical network connections.

TCP/IP is no more a Unix-only protocol; actually, it s emergence on desktop PCs has been one of its biggest commercial breakthroughs. Adding TCP/IP to PCs puts a whole new spin on the desktop PCs role in networking. A TCP/IP-based PC now has direct access to Unix applications and data. It is believed that the number of PCs talking TCP/IP is increasing by the day, at an alarming rate. One may ask, why TCP/IP? The answer simply is: that it is the only widely available standard protocol/technology that lets different vendors computers and network devices communicate. It is a layered architecture, with protocols for linking up computers and letting machines exchange e-mail, transfer files and emulate terminals, etc. This move of course, assumes that the Ethiopian telecommunication infrastructure supports wide area networking.

2. Unix integration and training.

Similarly, there is no escaping that Unix will remain as the dominant multi-tasking operating system for sometime to come. Hence, in a country where Unix operating system is almost virtually unknown, the focus should be to acquire and install more Unix systems. This effort should also include end-user training and incorporation of Unix related courses as part of the curricula in colleges and universities.

3. Utilization of ISDN.

The long-term strategy should not only consider TCP/IP connectivity between LANs, WANs, and remote networks, but should be able to integrate the next generation in communication technology, i.e., ISDN (Integrated Services Digital Network). If indeed ISDN becomes a reality on a global level, and Ethiopia can afford to implement it, then, its application is by far, the best alternative. This is because ISDN switched networks offer high-speed digital switched services for internetworking. Once implemented, they provide wide area connections that are cost- effective, flexible, reliable and easy to manage. When using switched networks, you pay only for the time the line is in use.

On the contrary, internetworking is often accomplished through routers or gateways connected by leased lines. The user pays a flat rate for the leased line service, whether the line is in use or not. Leased lines can be economical if the connection time between locations is long enough to cover the cost of leasing. However, data transfers are done in short bursts, in which case a switched network would be more economical.

4. Field Communication Systems (FCS).

FCS technologies such as FM radio, satellite, cellular radio, meteor bursts, microwave, and packet radio are digital technologies that can be used to provide either general communications or specific purpose-oriented communication services for a variety of applications. FCS are very ideal for rural communications and could have wider applications in developing countries like Ethiopia. Their basic attributes include a low per circuit mile cost, and flexibility, both in terms of media (i.e., data, video, audio) and expansion capabilities. FCS can be implemented at lower costs relative to the more traditional telecommunication systems and can serve individual, group, and regional communications needs.


The transition from standalone PC to multi-platform networked system environment(s) should be gradual, systematic, less painful; and the migration performed incrementally. At the same time, implementation of such systems and the changes associated with them should also be targeted towards the long-term betterment of the country s educational, technological, and above all, economical survival in a cut-throat competitive global market; and not just to suddenly overwhelm users with state-of-the-art computer and networking technologies.

Can there be a networking strategy that will answer all the questions? Probably not all; but the biggest challenge is in establishing a practical, feasible, and easily implementable plan. There is no doubt that the willingness and vision, know-how, and technology are all there. What needs to be done is, to formulate a working strategy and act on it.