Summary

  • KIXP opened in November 2000, was closed by the regulator the next month after a Telkom Kenya complaint, and resumed in February 2002 after CCK approval and new procedures.
  • Three ISPs exchanged traffic at the restart; reported performance and cost gains are useful evidence, but the figures come from provider claims and a contemporary OECD account, not a controlled national study.

A local route, then a stop order

The first Kenyan Internet exchange did not begin as a government project or a carrier's international gateway. A group of Internet service providers created the Kenya Internet Exchange Point, KIXP, to exchange traffic between their networks inside Kenya. The exchange began operating in November 2000. The following month, the Communications Commission of Kenya (CCK) ordered it closed after a complaint from Telkom Kenya, the incumbent telecommunications carrier.

That sequence exposed a gap between two kinds of competition. ISP services were open to competitors, but the OECD's 2002 account says Telkom Kenya held a monopoly over fixed infrastructure: local, national and international facilities, including leased lines. An ISP could compete for customers and still depend on the incumbent to connect its network. A local exchange changed where some traffic moved; it did not remove that dependency.

The exchange's value was practical. Before KIXP, traffic between two Kenyan networks could travel abroad and return, sometimes through the United States. For domestic messages or a locally hosted site, that path added international links to an exchange that could have happened nearby. The OECD reported that Kenyan ISPs used satellite links for international connectivity at the time. The providers argued that exchanging local traffic at KIXP would reduce delay and the cost of carrying domestic traffic.

The early operating record is striking but should be read with its source attached. The OECD reported average latency of 800–900 milliseconds before local exchange and 60–80 milliseconds during KIXP's initial operation. It also reproduced providers' price comparisons between domestic and international leased lines. The report says the ISPs behind KIXP estimated that traffic between Kenyans made up 80% of their total traffic. These figures show why the participants wanted a local path; they do not establish an independently audited national baseline or prove that KIXP alone caused every improvement.

The restart required an institutional step as well as working equipment. A contemporaneous message from ISP Kenya's chief technology officer says that, after extensive dialogue and the establishment of procedures, CCK approved the exchange in October 2001. KIXP went live again on 14 February 2002. ISP Kenya, SwiftGlobal and UUNET were the first three networks to connect and exchange traffic; ten ISPs had applied. The message described the mission as keeping local traffic local and anticipated benefits for service quality, international-circuit congestion and local applications.

Those were the organizers' goals, not a report of measured outcomes.

The disagreement was about the boundary of the service. The ISPs said the exchange was a closed user group and did not breach Telkom Kenya's international monopoly because international traffic still crossed the incumbent's links. That was the providers' interpretation, not a court judgment. The documented result is narrower: the first KIXP was stopped, later dialogue and procedures preceded approval, and a three-network exchange restarted.

KIXP's history makes interconnection look less like a switch that engineers can simply turn on and more like a layer of operating rights. A peering point can be technically small and locally useful while depending on rules for access, facilities and permission. Its first shutdown did not prove that local exchange was impossible; the restart did not prove the market had been liberalized. The two dates reveal which conditions had to change for the same function to operate.

Sources