Africa Research and Education Networking Workshop - Summary Report
18-19 September 2005
Internet2 Member Meeting, Philadelphia
Executive Summary
The meeting opened with a keynote address by AAU Secretary-General Akilagpa Sawyerr. This was followed by three sets of presentations by leaders in African and Latin American networking, focusing successively on the campus, national and continental levels. Each set of presentations was followed by a discussion period, and the workshop wrapped up with an overall discussion of next steps for the coming year. Detailed materials from all the presentations and video of the entire workshop are available at http://international.internet2.edu/resources/events/2005/2005FMMAFRICAAGENDA.html. Several workshop participants also met the next morning to further focus the consensus on what steps to take next.
The main points of agreement were:
Discussion
The main point agreed on in the discusssion was the importance of creating NRENs for infrastructure development, policy advocacy, and capacity-building purposes. NRENs offer much-needed clout in negotiations with local and international carriers; opportunities for united action on policy issues, especially telco deregulation and reduction in cross-border tariffs; and forums for technical learning, most importantly in coordination with continent-wide organizations like AVU and AfNOG. While there are few successful NRENs in Africa so far, Kenya's KENET and South Africa's TENET are exceptions and in many ways models for further developments in this area. Many of the tasks of NRENs are simultaneously tasks for consortia of NRENs on a regional or continental level; Latin America's CLARA has lessons to offer here.
The presentations included overviews of campus networks being built in Uganda, Nigeria, and Senegal. Those involved have had to become masters at doing much with little -- everything is in short supply, including money and reliable electric power. Under these circumstances, having a solid commitment from university officials becomes even more important than it is elsewhere. Fortunately there is growing awareness of the importance of R&E networking, driven by researchers and even more by student demand. Francis Tusubira observed that any African university that doesn't acquire adequate computing facilities in the next few years, won't have any students.
There was much discussion of the problem of developing and then hanging on to technical talent. Kunle Kehinde noted that at his university anyone who has received training is expected to then train other people; he described their general approach as to find motivated people and then give them plenty of latitude. AfNOG workshops (see below) have been a huge help in developing technical skills across Africa. Perhaps the greatest challenge is keeping skilled people on campus; there is strong demand from the private sector, which can usually offer much higher salaries.
There was much discussion of the need to look for non-traditional sources of infrastructure, such as fiber owned by utility companies and used for communications between electric power stations. Americo Muchanga observed that in Mozambique it's less a question of whether connectivity is available, than of whether it's available at a reasonable price. The Mozambican networking community's ongoing discussions with the Ministry of Public Works has led to fiber, or at least fiber channels, being added to road and power projects. Negotiating clout is clearly of great importance in this area; on the other hand, Steve Song observed that fiber availability is often discovered through informal social networks, and there is a danger of over-formalizing the process and disrupting these networks.
Major policy issues for NRENs include submarine cable access and tariff treatment of R&E traffic. Duncan Martin suggested lobbying the owners of the SAT3 cable -- who include major global players like Sprint and Deutsche Telekom -- for access; only 40 Gbps of SAT3's 120 Gbps capacity is currently in use. There is a lot of secrecy around business arrangements for cables; Andrea Johnson stressed the need to negotiate with whoever you can get to first. Martin also emphasized the importance of reducing tariffs on R&E traffic between African countries. Curtis White noted that most telco markets in sub-Saharan Africa are moving toward deregulation, so there is "no better time" to approach governments there about getting some changes. White stressed two such changes in particular: getting a tariff structure that gives non-commercial traffic a break, and funding concessions for things like distance learning. At the same time, F.F. Tusubira stressed that we can't put all the blame on a hostile regulatory environment; there has been a lot of inertia at African universities. Tusubira argued that the focus for the coming year should be on producing a common vision for African university networking.
It was clear from the discussions that Latin America's experience with building NRENs and continental networks -- CEDIA (Ecuador), CUDI (Mexico), and CLARA (continental) in particular -- can be of great value to Africa. Enrique Palaez of CEDIA observed that "even though we are in a different continent...we share exactly the same problems"; Palaez stressed that "teamwork and consortia" have been the leading solutions for R&E networking in Ecuador. Carlos Casasús noted that CUDI started with just seven universities sharing pro-rated costs; even this rudimentary level of cooperation enabled them to negotiate much better prices from Telmex.
At the Africa-wide level, AfNOG is clearly a model for capacity-building activities. AfNOG annual technical meetings and hands-on workshops have trained over 300 Internet engineers and built an active and mutually supportive network engineering community. Immediate priorities for AfNOG are to have more frequent, more regionally-based workshops, and to do outreach to North and Central Africa. Steve Song noted that efforts like AfNOG offer high return for the funds put into them, and are thus attractive to donors.
Song suggested using a 4x3 matrix as a tool for thinking about the various tasks confronting R&E networking in Africa: campus, national, regional, and continental levels would be placed along one axis, and the tasks of capacity building, infrastructure development, and policy advocacy along the other. Song also noted that in this matrix the "continental infrastructure" box is empty; there's really no one working on this. At the same time, there is fear of a hidden donor master plan to create a continental infrastructure. Americo Muchanga emphasized that there is no need for a big initiative to create an Africa-wide infrastructure; he strongly favors the approach of NRENs finding fiber from non-traditional sources as well as telcos, and Africa-wide organizations putting forward a continental vision that will guide connecting all these small initiatives into an Africa-wide network.
Avoiding duplication of effort is a priority in all three of the task areas discussed in the workshop; this would be another function of the projected continental vision. Organizations like AAU and NEPAD are essential here. Sawyerr noted that producing such a "framework for action" is the principal goal of the AAU-sponsored conference planned for 14-15 November in Tunis. Andrea Johnson noted that there is currently no one place you can go for information about who's doing what with African R&E networking; having this would be useful for both funders and African universities. There was general agreement that while a central database with complete contact information would be welcome, this would be a substantial project that could end up becoming worse than useless if not kept up to date.
Finally, there was strong general agreement that regardless of what new understandings and organizations arise from this process, we need to keep talking. Sandra Aluoch singled out the experience reports from various campuses and countries as an expecially valuable aspect of the workshop, and Ed Fantegrossi stressed the need to share the "little tricks" that have worked for each of us. Fantegrossi also observed that all speakers at workshops like this know more than they have time to say.