Summary

  • BOCRA’s 2024 IPv6 guidelines set a 2026 target of 100% for government, named operators and several other groups, but the target table does not define its denominator or verification test.
  • AFRINIC’s Botswana accelerator report records training, certification and project charters. Those are important inputs to migration, not measurements of how much of a network or service is reachable over IPv6.
  • BOCRA already publishes other dated ICT indicators. A compact, aggregate IPv6 progress receipt could make the plan legible while preserving the confidentiality the guidelines promise to provider submissions.

The number arrives before its denominator

In Botswana’s IPv6 migration plan, the last column is simple: 2026, 100%. The row is less simple. BOCRA’s Appendix 2 calls the measure “Target % of IPv6 Network/Services.” It assigns government, BoFiNet, BTC, Mascom, Orange and operators a 25% target for 2024, 60% for 2025 and 100% for 2026. ISPs, content-delivery networks and data centres, enterprise customers and end users have a 20%, 60%, 100% schedule.

That is a clear timetable. It is not yet a self-explanatory score. The table does not say whether the denominator is interfaces, sites, customer lines, applications, core and access network elements, or some weighted combination. Nor does the appendix specify how a counted unit is tested: an IPv6 address somewhere in a network is not necessarily equivalent to a service that a user can reach end to end. Those questions may be addressed in stakeholder plans or regulator records; the public table alone does not answer them.

The distinction matters because percentages borrow authority from arithmetic. A result of 60% is comparable only if the counted population and test are stable enough to be understood. If one operator counts enabled access nodes and another counts customer services, both could report the same percentage while describing different states. A national total could then look precise without being comparable. This is a measurement-design question, not evidence that any party has misreported progress.

Training is a start, not a rollout ratio

There is substantial evidence that preparation took place. AFRINIC Academy’s 2025 report describes the first minimum-viable edition of its Government IPv6 Accelerator in Botswana, hosted by BOCRA for ministries, departments and agencies alongside private network operators. The programme included a three-day strategy and planning workshop for executives and policymakers and a five-day certification event for engineers. AFRINIC says 54 engineers were trained and 38 passed, a 70.37% pass rate.

Its account also records a session with 31 CTOs, 24 government ministry project charters, seven operator charters and playbooks for seven private operators.

Those are concrete outputs. A trained engineer is not a deployed prefix; a project charter is not a production service; a playbook is not proof that every dependency has been tested. They are the work that can make deployment more likely and better managed. Calling them something else would diminish the programme while overstating what its report measures.

BOCRA’s guidelines themselves recognize this chain. They assign government a leadership role, call on service providers to assess readiness, plan implementation and declare deployment status, and give the Authority responsibilities for monitoring and audit. Providers are to keep rollout reports and submit progress at agreed times. At the same time, the guidelines say information or documentation supplied by providers to BOCRA will be treated confidentially. A public status view must respect that boundary; the existence of a confidential reporting duty is not evidence that reporting has failed or that the regulator holds no data.

A public reporting pattern already exists

BOCRA’s March 2026 State of ICTs in Botswana offers a useful, limited comparator. It covers a period ending September 2025 and says its tables draw on service-provider data. In the section on universal-access work, the report says 23 of 26 targeted highway sites were complete and operational by September 2025. Its SADC broadband scorecard compares targets with Botswana’s reported values across ten indicators. BOCRA also explains why two indicators were not reported: measurement could not be applied consistently across member states because the method was unclear.

That is the right kind of disclosure for a target that cannot yet be measured cleanly: retain the target, show the available result, and say why a value is absent. The report contains no IPv6 progress chapter or IPv6 search match. That establishes only what this particular public report covers—not that BOCRA has no IPv6 submissions, that no deployment has occurred, or that no other publication exists.

The comparison is not one-to-one. Highway sites are countable physical assets. “IPv6 network/services” spans technical layers and organizations, and publishing unit-level operator information could reveal more than a public reader needs. The lesson is not to copy the highway metric. It is that a regulator can give readers a dated numerator, denominator or scope, and a candid account of missing or non-comparable data while still relying on provider information.

An external snapshot cannot substitute for that scorecard. AFRINIC’s Botswana IPv6/DNSSEC monitor is dated 21 May 2026 and lists selected public-facing domains and their DNS, mail and web service-interface observations. Some government domains in that sample show zero IPv6 interfaces for the listed services. That observation is worth checking, but it is neither a census of government networks nor a measure of the guideline’s percentage denominator. It is months old at this article’s cutoff and cannot establish national attainment or non-compliance.

Publish the method, not private submissions

A useful aggregate receipt could answer six questions without naming individual provider results: What reporting period is covered? Which target class is being measured? What counts in its denominator? What is the numerator and the exclusion rule? Which test or record supports the figure, and who validated it? What share is reported, not yet verified, or not comparable with the previous period?

Where a single percentage would flatten unlike systems, the report could separate network core, access, public services and customer-facing applications—or publish a range and explain its scope. It should keep “not reported,” “not measured” and “zero” distinct. Those labels describe different evidence states. A value withheld to protect a provider is not the same as a test that returned no IPv6 service, and neither should silently become a national failure claim.

The public-interest case is practical. Government procurement, service design and network investment depend on knowing whether IPv6 is only planned, available in selected parts, or supported across a defined class of services. Operators also benefit when readiness, deployment and service reachability are not collapsed into one ambiguous score. A stable method lets a future report be compared with the last one; a number without a method cannot do that work.

The bounded conclusion is therefore not that Botswana has or has not met its 2026 target. The published record reviewed here shows a formal timetable, a real capacity-building programme, regulator duties that include progress reporting, and public examples of BOCRA publishing other indicators—including an explanation when measurement is unclear. What remains unclear in the public IPv6 record is the denominator and a dated aggregate result. Publishing those two things, while protecting provider-specific submissions, would turn the 100% target from a headline into something a reader can actually audit.

Sources