Summary
- UbuntuNet is a member-driven non-profit organisation registered in Malawi for national research and education networks in East and Southern Africa; its documents use both “association” and “Trust”.
- Regional capacity has been built from carrier contracts, PoPs, member routes and AfricaConnect programmes, including 10 Gbit/s core links and higher-capacity national connections.
- The portfolio now extends beyond transport to identity, cloud and storage, open science, cybersecurity and NREN development, with services at varying levels of maturity.
- Bundled demand and expertise create value; but weak campus connections, arrears, supplier power, conflicts, funding cycles and incomplete reporting can prevent capacity from reaching researchers.
Network 3 brought the core into the 10 Gbit/s operating class
The 2023 report describes the shift from approximately 2.5 Gbit/s to substantial 10 Gbit/s links. At the same time, at least 20 Gbit/s total capacity is mentioned through redundancy. These are different levels, not a universal speed figure.
A plausible architecture is one with redundant 10 Gbit/s links on relevant core segments, not 20 Gbit/s on every path or for every member. Rates vary by route, carrier and national connection.
The move was significant for pricing and headroom, but remains far below networks with hundreds of Gbit/s or Terabit capacity in wealthier regions. Any comparison must explain regional procurement realities, not just rank speeds.
Lower unit prices were an infrastructure outcome
The annual report cited price reductions of 50 to 71.5 per cent compared with 2022. Depending on the NREN, services rose from around 130 Mbit/s–4 Gbit/s to 500 Mbit/s–8 Gbit/s.
The reduction came from bundled procurement, long-term contracts and more efficient core utilisation. The goal was not merely more capacity, but more service per budget unit.
Passing savings on to members leaves less buffer for currency movements, outages or late payments. A low price is sustainable only if ongoing costs are transparent and regularly covered.
BotsREN and ZIMREN show how regional upgrades reach national networks
BotsREN was first connected at 10 Gbit/s in 2023 and upgraded to 20 Gbit/s in January 2026. ZIMREN joined at 10 Gbit/s in June 2025. These are member connections, not the speed of the entire network.
They nevertheless show that core investments translate into tangible national capacity. The benefit then depends on connected institutions, domestic circuits, campus equipment and operations teams.
A national 20 Gbit/s access can end behind campus boundaries at 1 Gbit/s, unreliable power or overloaded firewalls. Connection announcements should therefore be supplemented with usage, loss, availability and institutional reach.
The regional layer begins where national research networks hit their limits
Students and researchers rarely see the name UbuntuNet Alliance. What is visible first is the campus or institutional network, then a national research and education network such as KENET, RENU, TENET, ZAMREN, TERNET or MoRENet. UbuntuNet becomes relevant at the next boundary: when a national network needs to exchange traffic with another country, reach an international research facility, connect to GÉANT or use a jointly supported service.
This position explains why the Alliance can matter without acting like a consumer ISP. It does not normally sell residential connections, does not connect every university directly and does not control the local Wi‑Fi. It operates the regional routing and services layer atop legally and operationally independent national systems. Other regional networks, carriers, clouds, internet exchanges and the destination system complete the path.
To the user, everything feels like a single session; responsibility, however, is divided. UbuntuNet can change routes, expand core links, monitor regional services and coordinate incidents. It cannot fix a campus power outage, correct an outdated identity directory, force a national carrier to repair fibre or guarantee the performance of the target application. The task, therefore, is to make independently controlled layers work together reliably.
Federation is the operating model, not a temporary substitute for centralisation
UbuntuNet Alliance is an alliance of national networks, not a holding company. Each member retains its legal personality, staff, budget, domestic infrastructure, university relationships, political embedding and pricing. The regional layer exists because the members jointly procure certain capacity, exchange traffic at shared locations, share services and advocate common interests.
That fits a region with widely differing telecommunications markets, currencies, higher-education systems, procurement rules, languages and levels of maturity. A centrally owned academic network spanning all countries would require political authority that UbuntuNet neither possesses nor needs. Federation combines national autonomy with regional scale and enables knowledge transfer from established to emerging NRENs.
Complexity does not disappear; it resides at the handoffs. A slow transfer can be influenced by the campus firewall, a national connection, a regional PoP, an international carrier and the destination system. Each operator sees only a slice. Clear service boundaries, shared telemetry, escalation paths, change processes, contribution rules and real exit options are therefore part of the infrastructure.
The legal structure is usable, but public terminology remains inconsistent
Current UbuntuNet documents describe the organisation as a non-profit Association registered in Malawi. The 2023 annual report and historical transitional documents also refer to a non-profit Trust with members and Trustees. The current constitution was not available in the Research Pack; the discrepancy should therefore not be silently resolved.
The practical governance is clearer. UbuntuNet has members instead of ordinary shareholders. Documented are a Members’ Assembly, a Board of Trustees and a Secretariat headed by a CEO. The organisation employs staff, receives grant funding, enters into contracts, holds or leases facilities and operates services. There is no indication of distributable corporate shares or a commercial valuation.
The choice of words shapes expectations. “Private telecommunications company” would imply investor control and returns; “voluntary community” would downplay contractual and fiduciary duties. The most robust description is therefore: a non-profit membership organisation of NRENs, registered in Malawi, whose own sources use both Association and Trust.
The founding problem was cost and routing, not a total absence of internet
Universities already had internet access before UbuntuNet was formed. What was missing was a lasting regional structure that bundled academic demand, exchanged research traffic and reached global R&E networks on terms tailored to science and education. Individual institutions bought small international capacities at high unit prices; traffic between African countries could traverse Europe.
A commercial connection offered general access, but it did not automatically create direct routes between neighbouring NRENs, predictable capacity for data-intensive research, federated identity, shared security structures or collective bargaining power. New NRENs could struggle to demonstrate value as long as they lacked members and capacity.
The Alliance changed the demand unit. Campus networks aggregate nationally, national networks aggregate regionally. That larger community can plan PoPs, sign long-term contracts, organise operations teams and negotiate with global partners. UbuntuNet does not replace the market; it acts as a coordinated buyer and mission-driven operator.
Five NREN initiatives formed the first coalition
UbuntuNet Alliance was conceived in the second half of 2005 by initiatives in Kenya, Malawi, Mozambique, Rwanda and South Africa. The founding base is associated with KENET, MAREN, MoRENet, RwEdNet and TENET, in their institutional forms at the time. Their starting points varied.
TENET and KENET brought operational experience and established higher-education communities. MAREN anchored the later Secretariat in the host country. MoRENet represented a government-backed model. RwEdNet later demonstrated both potential and fragility: collapse and rebuilding showed that a regional link does not replace a sustainable national institution.
That very mix was useful. Mature networks alone might have created a narrow interconnection club; weak initiatives alone would have had too little traffic and credibility. The coalition combined productive build-out with institutional development.
Francis Tusubira helped turn a network proposal into an institution
Francis “Tusu” Tusubira is closely associated with the founding period and the African NREN movement. The task was not purely technical. Universities, ministries, regulators, carriers, funders and engineers had to be persuaded that a regional research network should exist as an institution, not a time-limited project.
That required legal capacity, membership rules, accounting, contracting capability, a recurring forum and a shared articulation of why an NREN is more than a cheap ISP. Tusubira’s contribution belongs within that collective build; it does not replace the roles of the founding networks, boards, technical communities and international partners.
The lasting result was a body that could fund programmes, procure long-term capacity and represent a region. That institutional capability made AfricaConnect feasible in the first place.
Amsterdam provided the first legal vehicle; Malawi became the regional home
Registration documents were filed in Amsterdam in 2006. For a young organisation, a Dutch structure offered a framework familiar to European partners for contracts and projects, before a stable legal base existed in the region.
As it matured, a European domicile fitted the mission less and less. A new constitution was adopted in Malawi in 2012, registration followed in 2013, and historical documents describe the transfer of assets and obligations. The Secretariat is now based in Lilongwe.
The move brought governance closer to the members, but left the association/trust terminology in place. The correct chronology distinguishes conception in 2005, early formalisation in the Netherlands and later relocation to Malawi.
Early GÉANT transit demonstrated value before a full regional backbone
Historical sources place an R&E transit service to GÉANT in 2008. Even at limited scale, it opened paths for members into European and global research networks and showed that the regional organisation could deliver operational services before a full backbone existed.
The relationship brought more than bandwidth: procurement know-how, project management, identity and security experience and access to EU programmes. That collaboration accelerated the build, but also created dependency on external programme cycles.
The reality lies between imported infrastructure and full independence. UbuntuNet used global partners to build regionally controlled capability. What mattered was which elements could transition to steady-state operations after the project ended.
Long‑term carrier contracts turned bundled demand into infrastructure
The 15‑year contracts with WIOCC in 2013 and SEACOM in 2016 gave UbuntuNet more planning certainty than short‑term, small internet contracts. Long terms can lower unit prices, secure particular routes and enable expansion.
They also create obligations. Provider quality, technology changes, maintenance, currency risk and regular payments remain relevant. An international invoice falls due even if a member pays its contribution late.
The contracts show how federation works technically: one organisation bundles demand, signs for multiple members, installs equipment and coordinates restoration. It does not thereby become the owner of all fibre, landing stations or transport networks.
The 2014 launch turned promise into production operation
In July 2014 UbuntuNet Alliance and DANTE announced regional connections at 622 Mbit/s. The figure feels small today, but it was a major step for networks that had much less and more expensive capacity.
From that point there was real traffic, outages, configurations, on‑call duties and expectations. An error was no longer just a project problem; it affected universities and research.
The first production operation also created PoPs, NOC processes, carrier relationships and billing mechanisms that could later be expanded.
AfricaConnect2 expanded the programme and strengthened eastern resilience
AfricaConnect2 was announced in 2015 as a €26.6 million programme covering several regions, involving UbuntuNet, WACREN, ASREN and GÉANT. That sum is not UbuntuNet’s revenue; it is a multi‑regional programme envelope.
For the Alliance the phase brought additional links and members, as well as the SEACOM path Kampala–Dar es Salaam–Amsterdam. This increased eastern route diversity and reduced reliance on a single path.
AfricaConnect2 also cemented the federated pan‑African model: independent regional networks develop their infrastructure within a shared programme. Outcomes remained uneven because of differing markets and national maturity levels.
AfricaConnect3 made services as important as the backbone
AfricaConnect3, with a total programme value of €37.5 million, supported network, pricing, training, identity, cloud and open science until April 2025. The mission was no longer measured by circuits alone, but by whether members could use the shared infrastructure.
Network 3, new PoPs, faster member connections and lower unit prices were technical outcomes. In parallel, identity federation, repositories, Utafiti Africa, AfricArXiv and cybersecurity grew.
The expansion is logical: fibre without storage, identity and skilled people can go unused. It also increases the maintenance burden, however. Every service brings software, data, security responsibility and ongoing costs.
AfricaConnect4 opens a new investment window and sets a sustainability deadline
AfricaConnect4 began in 2026 with an additional €40 million of EU support over four years for regional networks in sub‑Saharan Africa. The amount is programme‑wide, not an exclusive allocation to UbuntuNet.
It foresees national and campus components, cybersecurity, climate sensing and data, EUMETCast, federated computing, GPUs and open science. This recognises that a fast backbone achieves little if institutions cannot use it.
By 2030 it must be clear who will run stations, repositories, identity services, clouds and clusters after funding ends. Every service needs a permanent operator, budget, billing model, staff and a migration or decommissioning plan.
AS36944 defines UbuntuNet’s public routing boundary
UbuntuNet uses AS36944. PeeringDB and BGP observations confirm the identity, some peers, prefixes and valid RPKI states for observed origins. They show the public edge, not all private links.
PeeringDB reports a self‑declared traffic category of 5–10 Gbit/s and 1,200 IPv4 and 150 IPv6 prefixes. bgp.tools sees fewer directly originated prefixes. The sources measure different snapshots.
None of these figures replaces core capacity or member access. The ASN proves participation in the BGP system, but not a specific availability, peak load or physical diversity.
PoPs and peering create options, but a map is not an asset register
The 2023 report lists Cape Town, Mtunzini, Maputo, Dar es Salaam, Nairobi, Amsterdam, Gaborone and Johannesburg; Djibouti and Mombasa were planned or being built. Older pages mention London and other historical locations.
Differences may arise from timing, varying PoP categories or changed handovers. A dot on a map does not reveal who controls the router, fibre, rack, port or contract.
Johannesburg and NAPAfrica created new southern options. Resilience, however, depends on genuinely independent paths, providers, power, datacentres and control planes. Two drawn routes can share the same physical bottleneck.
Lagos–Cape Town made pan‑African federation more tangible
Developments in 2025–2026 linked WACREN, via ZAOXI, SANReN and TENET, to the South African environment and onward to UbuntuNet. Partners described this as the first effective direct connection of African regional networks for research on African soil without transit via Europe.
The claim must be attributed and restricted to RRENs. It does not assert that no commercial or private north‑south links existed before. What is new is the institutional and operational research‑network relationship.
The route can keep more research traffic on the continent and reduce detours. It remains dependent on carriers, exchanges, BGP policy and national partners. No one party controls the entire path.
Membership figures describe different categories and points in time
The 2023 report cited 15 active members, more than 3.5 million users and over 1,000 institutions in 13 countries. An audit tender of March 2026 listed NRENs from 18 countries, while the website displayed 16 profiles. Differences between legal, active, affiliated and published membership are possible.
Lesotho was welcomed in November 2025; trustee elections took place in 2026. Every figure needs a date and a definition.
Membership does not prove an active connection, and a national connection does not prove equal service across all universities. In conflict areas, legal status and operations can diverge significantly.
Campus infrastructure determines whether regional investments are used
A fast backbone does not repair undersized campus fibre, outdated Wi‑Fi, unstable power, overloaded firewalls or missing staff. The weakest segment defines the experience.
That is why AfricaConnect4 includes national and campus components. Regional capacity can go unused if institutions lack storage, servers, transfer tools or adequate LANs.
Responsibility is distributed: UbuntuNet coordinates regionally, the NREN builds nationally, the university runs the campus, and governments, carriers and funders set the conditions. Impact must be measured all the way to the researcher.
Lilongwe carries governance, Kampala concentrates network operations
The Secretariat is based in Lilongwe, Malawi; organisational documents place the NOC in Kampala, Uganda. Such a separation is plausible for a distributed regional organisation.
The Secretariat manages members, contracts, projects and finances. The NOC monitors links, raises tickets, coordinates providers and supports NRENs. The model needs secure access, clear responsibilities and business‑continuity plans.
Documentation, on‑call rotas, access controls and succession are critical. A small organisation must not concentrate critical knowledge in a single person.
Shared operations should strengthen national capabilities, not replace them
Not every NREN can sustain a 24/7 NOC. UbuntuNet can provide monitoring, preventive maintenance and first‑ or second‑level support. The national organisation retains customers, policies, providers and authority.
The service is valuable when it closes gaps and transfers knowledge. It becomes risky if the member no longer understands its own configuration or cannot leave. Action rights, maintenance windows, logs, escalation, costs and handover must be documented.
The regional NOC can diagnose and coordinate, but cannot repair a national fibre without the carrier. Its value lies in correlated evidence and shared operational knowledge.
Federated identity is a second network built on institutional trust
The physical network carries packets; federation allows a service to trust a person authenticated by another institution. The university manages the account, the national federation manages rules and metadata, and regional and global layers link the trust.
UbuntuNet does not store all passwords. It runs or supports proxies, federations, metadata and access to eduGAIN. Quality depends on local directories, account lifecycle, MFA, certificates and attributes.
The benefit is access to libraries, clouds, repositories and computing using home credentials. The risk is the amplification of a compromised IdP. Federated trust demands federated audits and revocation processes.
eduroam makes campus identity mobile
eduroam enables credentials to be used at a visited institution. Authentication can traverse local, national, regional and global RADIUS proxies and return to the home institution.
The simple user experience depends on certificates, realms, secure Wi‑Fi and well‑managed accounts. UbuntuNet can provide the regional layer, but cannot fix an expired password or a misconfigured access point.
Success should be assessed by active institutions, successful authentications, incidents and resolution times, not merely by installed servers.
eduID.africa and eduGAIN create a route into larger trust frameworks
eduID.africa is operated with WACREN and ASREN as a catch‑all federation and template for NRENs that do not yet have a mature national federation. This allows participation to begin before every country has built the full stack itself.
MAREN and TERNET joined eduGAIN in 2025, connecting Malawi and Tanzania to the global interfederation. Admission demonstrates institutional interoperability, not identical maturity on every campus.
The regional layer should accelerate national capabilities and allow a clean migration. A permanent opaque substitute would weaken autonomy.
Research4Life turns federation into practical knowledge access
A 2025 pilot enabled institutions in Kenya, Malawi, Uganda and Zambia to access Research4Life remotely using their home credentials. Follow‑up reporting cited 66 participating institutions in 15 countries.
This reduces shared passwords that are hard to revoke and attribute. Access can be tied to person and affiliation without creating a separate account for each resource.
The figure is time‑specific, not universal coverage. Institutions must maintain identities; Research4Life retains eligibility rules. Even so, the case shows the direct benefit of the trust infrastructure.
Cloud services link regional infrastructure and commercial procurement
One pillar covers community‑run or hosted infrastructure: UbuntuNet Open Science Cloud, file transfer, Drive, hosting, collaboration and experimental computing. The other opens commercial offerings, especially via the 2024 Redington/AWS agreement.
A regional cloud can keep control and expertise closer to members; a hyperscaler offers broader products. UbuntuNet bundles demand and support, but does not own AWS and does not control pricing, availability or roadmap.
The 2023 report documents an installation in Lusaka and equipment for Kigali, but no current total capacity, utilisation, tenant count or shared SLA. Installed hardware is not yet a mature multi‑region cloud. Orchestration, identity, backup, security, support and funding are what count.
Research support addresses everyday problems as much as big‑data questions
Utafiti Africa consolidates funding and training opportunities and passed 1,100 active users in 2024. It does not award grants; it reduces search effort.
File transfer, Drive, Meet, Learn and hosting can avoid duplicate procurement and ease collaboration. Maturity and operators differ; older pages mix productive offerings with plans.
Transferring a large file is not the same as moving continuous telescope, climate or genomic data. Such workloads need tuned endpoints, storage, observability and sometimes dedicated circuits. Every service needs a clear domain of use.
Open science became a second infrastructure layer
Connectivity does not automatically make an output findable, citable or durably available. Repositories, metadata, identifiers and people also rely on hosting, storage, identity, certificates and backups.
UbuntuNet and Access 2 Perspectives have hosted the AfricArXiv infrastructure since late 2023. AfricArXiv remains community‑governed and is not owned solely by the Alliance. UbuntuNet is also the Southern Node of the Africa Open Science Platform.
Repositories have a long lifecycle: software, rights, data, security and preservation require ongoing maintenance and export capability. A neglected scholarly archive can suffer lasting damage.
Repositories and persistent identifiers demand continuous institutional work
The Research Repository Stewardship Programme trains librarians and data stewards. In July 2026 the first Portuguese‑language workshop was held in Mozambique; localisation is thus part of operations.
Institutions need deposit policies, metadata, rights clearance, support, backup, updates, preservation and budget. Multi‑tenant platforms reduce duplication but create a shared failure domain. The Network Adoption Fund aims to turn NRENs into sustainable service providers.
DataCite, DOI, ORCID and the 2026 DOCiD partnership with TCC Africa add identity for people, institutions and entities. Identifiers do not fix bad metadata. UbuntuNet can coordinate procurement and integration without obscuring global governance or local responsibility.
Cybersecurity capability remains distributed across people and institutions
The Alliance organises bootcamps, national workshops and identity and system training. There is no evidence of a central SOC for all members. The model strengthens local teams, contacts and shared procedures.
Laws, data and incident authority remain national or institutional. The regional layer can share tools, but cannot automatically remediate a compromised campus server.
A 2026 finance training with entities from 15 countries shows that resilience also covers budgeting, currency, compliance and contribution collection. Administration is part of the infrastructure.
Governance ensures continuity, although public disclosures have gaps
The Members’ Assembly is the supreme body, the Board of Trustees oversees, and the Secretariat executes. Professor Madara Ogot was appointed CEO for four years from 1 February 2022 and continues to be referred to as such in current 2026 sources; a separate reappointment was not found. Professor Hellicy Ng’ambi became the fourth Chairperson on 1 January 2024.
Miriam Chahuruva and Sabelo Dlamini were elected as trustees on 29 May 2026. Sources do not thereafter publish a fully reconciled board list and do not name all departing members.
This is not evidence of vacancies, but it limits journalistic precision and illustrates the value of a synchronised public governance register.
The 2023 numbers show a large mission with a very small financial buffer
UbuntuNet reported US$2,404,779 in revenue, US$2,393,911 in expenditure and a surplus of US$10,868. Connectivity contributed US$1,902,318 or 79.1 per cent; network operations cost US$928,032.
The buffer is small given the mission and risk. A currency swing, an invoice or a late payment can consume it. Passing savings to members fulfils the mission, but reduces reserves.
The CEO reported that late operations and maintenance contributions affected the ability to meet obligations. More recent audited accounts were not found; programme values do not permit an estimate of the current financial position.
Funders, carriers, currencies and conflicts remain part of the operating model
AfricaConnect finances major expansions; carriers provide physical infrastructure; clouds control platforms; contracts are often in hard currencies; and some members operate under conflict conditions.
RwEdNet collapsed in 2018 and was rebuilt in 2024. This shows that a national institution can fail despite the regional network. Membership of a conflict‑affected country does not prove uninterrupted operation.
These dependencies define the work: diversify, build national capacities, document responsibilities and not confuse sovereignty with isolation. UbuntuNet creates more options but does not remove the global ecosystem.
Public evidence is strong on programmes and weaker on ongoing operation
There is rich documentation on contracts, connections, elections, trainings and launches. Missing are a current route register, multi‑year availability, complete traffic series, link utilisation, a public incident database, today’s staff count and accounts after 2023.
PeeringDB, bgp.tools, core capacity and member connections measure different things. A PoP can mean technical or contractual presence without disclosing the physical asset. No independent SLA can be calculated from them.
No public outage does not mean no outage; absent data does not automatically mean poor performance. Greater disclosure of availability, restoration, usage and service adoption would improve accountability.
Regional autonomy comes from options, not from eliminating all dependencies
UbuntuNet gains control through joint procurement, more local African routes, engineering build‑up, and regional identity and repository services. Global standards, commercial cables, vendors, hyperscalers and research partners do not disappear as a result.
WACREN, ASREN, SANReN, TENET and GÉANT show a pan‑African federation without a sole owner. Lagos–Cape Town improves route localisation, but still depends on carriers, exchanges and coordinated policy.
Every option has a trade‑off: AWS simplifies access but increases platform dependency; a regional repository strengthens governance but needs software and staff; a long‑term contract lowers prices but raises carrier and currency risk.
The test is portability. Can a member switch providers, export data, move workloads and keep identity? Autonomy grows when the exit is real.
UbuntuNet turns separate networks into a usable research environment
The most important achievement is institutional and technical. Independent national networks bundle demand, share operations, connect globally and develop services that would be expensive or unavailable singly. Backbone, identity, repositories, cloud and training are different layers of the same answer to fragmentation.
The organisation remains unfinished. Capacity is limited when measured against climate, astronomy and AI; public information is stronger on milestones than on performance; national finances are unequal; providers, funders and services at varying maturities remain part of the system.
It is precisely these limits that make the profile meaningful. UbuntuNet is neither a separate African internet nor a finished continental platform. It is the shared layer that enables a larger system. Success means that members can use, fund, observe and leave services without losing their autonomy.