Task: To provide cheap e-mail connectivity to four remote sites.

Before reading on, stop and think about how you would do this with your familiar technology.

Now, being a good IT manager, you ask "How cheap?" You discover there is commercial broadband satellite access but it costs more per month than the salary of the local manager. In fact, the budget is nearer to single-digit dollars per month than hundreds of dollars per month.

You ask "How remote?" So remote there is no main power supply, only generators, and the fuel for them has to be transported hundreds of miles. Even in the hub towns the remote regions encircle, the power supply is erratic, as is the telephone service.

Such was the challenge faced in Africa in 1998 by the Informatics Development Institute (IDI), a not-for-profit company whose mission is to further the global development of both information and communications technologies and to research their socioeconomic impact. The challenge in bringing communications to remote regions of developing countries is not a simple one of technology transfer. Rather, the problem is that the totality of applied technologies needed to reach the desired objective does not yet exist.

How the IDI Did It

The IDI chose to implement a solution utilizing Low Earth Orbiting satellites (LEOSAT), which are typically low cost (usually US $1 million or less) and available for use at no cost for humanitarian purposes. Unlike the more expensive geostationary satellites that are familiar to most in the developed world, LEOSAT are visible in all parts of the world but only for a few minutes at a time, a couple of times a day. Consequently, they provide only simple radio access and onboard storage to collect and deliver low volumes (typically under 100Kbytes) of e-mail from each ground station they pass over.

The IDI developed in close collaboration with its African partners, over a period of more than three years, a global satellite-based network that linked remote site end users with Internet end users via gateways located in Europe, the US, and New Zealand. It built, tested, shipped, and installed four steerable antenna ground stations (main ground stations, or MGS), and 12 battery/solar-powered portable ground stations (PGS) at African locations.

Lessons Learned

The most important lesson the IDI learned was to question everything and to provide a backup to every fallback contingency. These lessons are similar to those many managers in corporate IT learn over the course of a project. Taking a step back to reflect on those lessons and incorporate them into future projects is invaluable. The IDI learned:

1. So poor was the dial-up phone service and campus network availability, the IDI should have installed PGS at the headquarters as backup for the "normal" e-mail.
    
2. The IDI "saved money" by having the PGS hand built by an individual. Unfortunately, trying to diagnose problems and fix them was all but impossible for anyone but a skilled radio technician. So the IDI learned that a PGS needs to be testable to be serviceable, or must be cheap enough to be able to afford several spares and be easily replaceable as a complete unit.
    
3. Home-built equipment is also too easily accessible. Because exchange spares were not always at hand, the temptation for operators at the local site to "fix" the kit was just too great. As a result, many were rendered unusable and brought the project to a halt in some regions, as there was no budget for new PGSs.
    
4. The PGS was developed to use the same LEOSAT synchronization method as the ground stations in the developed world -- the weekly automatic or manual update of the satellite pass schedule from Kepler data. However, the Keplers downloaded from the satellite itself were not always correct; and the difficulty of communicating revised data to local operators for manual updates made the IDI realize that a more adaptive system is needed.
    
5. The IDI knew that training was important, and budgeted for bringing staff to its headquarters in Ireland for an initial train-the- trainers course, as well as for a site tour by an expert to reinforce the training locally. The IDI learned:
   1. Training in Ireland took place too far in advance of delivery of the systems to the regional sites. As a result, some trained staff had either changed jobs or forgotten some knowledge by the time the system was ready.
   2. The training was delivered in classic academic classroom fashion following a fixed syllabus. There was no written examination or hands-on testing to continuously validate that foundation items were fully grasped before proceeding.
   3. The IDI should have accredited or verified the ability of the local trainers to perform the training, and also verified that the trainees had actually learned the material.
   4. The IDI had not anticipated the high rate of turnover of local staff, and learned that it's necessary to have backup trainers both in the initial session, and to have further fallback replacements nominated.

Conclusion

The IDI intends to exploit the field experience gained in developing, refining, and implementing the specifications for the next-generation PGS, appropriate for use in remote regions in developing countries. It has made such a proposal and while it has not yet found a project in the European Research Sixth Framework Programme (FP6), IDI is seeking people who are interested in helping with the push for widespread dissemination of this technology in developing countries around the world.