Summary

  • APNIC and ITU helped establish a Timor-Leste exchange test case in 2016. UNCTAD’s 2024 assessment later said TLIX did not appear functional, while stressing that operational data were limited; it separately described a new budget, tender and tests with three operators.
  • APNIC says ANC is establishing TL-IXP in 2026. That is a useful status report, not proof of commissioning. The decisive public test is whether operating authority, support, funding and continuity can be traced after deployment assistance ends.

A launch is not a handover

The 2016 record is brief. APNIC Executive Council minutes say APNIC Technical Assistance, working with ITU’s Bangkok regional office, helped establish a test-case internet exchange for Timor-Leste, TLIXP. They do not describe a later operating model, traffic, members, uptime or an exit plan.

Eight years later, UNCTAD’s eTrade Readiness Assessment offered a bounded account: TLIX had been established in Dili in 2016 but, given the limited operational data available, did not appear functional at the time of the assessment. The same report says a budget for another IXP had been approved in 2024, tendering was under way and tests were being conducted with three operators. Those are two dated observations. The report does not establish why the first arrangement stopped, or whether the newer procurement was a direct continuation of it.

That distinction matters because an exchange is not made durable by the presence of a switch or route server alone. It needs someone with authority to maintain the fabric, restore a failed service, set participation rules and fund replacement equipment. It also needs a record that lets users tell whether the service still works. The documentary gap is not proof of a technical failure. It is a reason to ask what a successful transfer would look like.

What the new project says—and does not say

A 2025 presentation hosted by APNIC Foundation describes ANC as the planned host and names Timor Telecom, Telemor and Telkomcel among major ISP stakeholders. It sketches shared governance, free peering, route servers, IXP Manager, mirrors and monitoring. The presentation is a design artifact; it cannot establish that the design was accepted or deployed.

ANC’s 2025 technical-adviser terms are more operational. They ask for route-server and IXP Manager deployment, RPKI validation, monitoring, policies, onboarding manuals, staff training, a sustainability and handover roadmap, and 60–90 days of support after deployment. The vacancy closed in August. The reviewed record does not confirm who was hired, which outputs were delivered or whether that support period began.

APNIC’s 16 October 2026 profile says ANC is establishing TL-IXP and lists route servers, DNS root mirrors, monitoring and RPKI validation among the plan’s components. APNIC is a knowledgeable project partner, but this is not an independent acceptance test. Internet Society Pulse, meanwhile, displayed no active Timor-Leste IXP entry under its rule based on apparently active PeeringDB listings with at least three autonomous systems. That is a discoverability signal, not proof that no facility or private interconnection exists.

Even a live route server would answer only part of the question. RFC 7947 specifies that a route server distributes routing information; it does not forward user traffic. Session counts and accepted routes can show participation in the control plane, but not how much traffic crossed the exchange, whether paths stayed local, or whether latency and resilience improved.

Make continuity observable

A useful handover record would name the operating organization and the role responsible for routine maintenance, incidents, software updates and equipment replacement. It would state the support window, escalation route, funding source and the point at which local staff assumed each task. If responsibilities are shared, the record should say which decisions belong to ANC, members and any external technical partner.

Operational evidence can be published without exposing sensitive configurations. A dated service report could distinguish members, route-server sessions, accepted routes, traffic volume and measurement periods; report uptime and planned maintenance; and summarize incidents and restoration times. Each metric needs a denominator and a method. A route count is not a traffic count, and an online dashboard is not proof of service continuity unless it covers an identifiable period.

The sequence also deserves a public explanation. The record can describe the 2016 test case, the limited evidence available in 2024, and the separate procurement and establishment effort now under way without inventing a cause for the gap. If the current project is a successor, say so and identify the decision. If it is a new arrangement, state what was carried over and what was not. That history is a governance choice as much as a technical one: it determines who can be asked to answer for past commitments and who owns the next maintenance decision.

The system risk is conditional. If deployment support ends without a named local maintainer, budget and escalation path, ordinary faults and upgrades may again depend on ad hoc outside help. If those responsibilities are assigned and reviewed, the exchange gains a more credible route to continuity—even before anyone can claim lower prices or better user experience.

Timor-Leste does not need to publish every configuration or promise a quantified national benefit. It does need enough evidence for participants to distinguish installation from operation and operation from impact. Until an acceptance record, operating owner and periodic service evidence appear, “establishing” remains the most precise public description.

Sources