Summary

  • LACNIC’s April 2026 article names five root-server identities in Panama, while five operator records retrieved in October sum to six submitted instances.
  • The sum is not an audit of physical equipment, and the published response-time discussion does not map each Atlas probe to a route scope or site.

Five root-server letters and six listed instances can coexist without contradiction. LACNIC’s April 2026 country article names D, E, I, K and F. The public site records maintained by the operators use another field: instances. Retrieved on 10 October, those records list one Panama City site row for each letter and add up to six.

The two figures come from different sources and dates, and they count different units. The useful question is whether a reader can follow the transition from identity to site to instance—and then from a measurement probe to the server instance it actually reached.

Identity, site and instance

A root-server letter identifies an operator service in the root-server system. It does not mean that service has only one physical server. Anycast lets an operator announce a service address from more than one site, with routing bringing a query to an instance selected by the network path. A site record is another layer: operators describe where their service is offered and how many instances are represented there.

The Root Server Operators’ operational-information draft makes that distinction explicit. Several instances at one location can be grouped in a row with an Instances value above one; locations with different parameters can instead appear as separate rows. The Type field describes route scope. So neither the number of letters nor the number of rows is automatically the number of instances.

LACNIC’s own reporting also needs dates attached. Its April 2024 article named K, I and F and reported 35 RIPE Atlas probes. The April 2026 article named five identities and said 21 probes were connected, including an anchor at InteRed, with probes concentrated in Panama City. Those are dated descriptions with different wording. They do not establish that fourteen probes disappeared or that two deployments were added between reports.

What the operator records add

The five root-specific JSON records each contain one Panama City row. D reports Instances 1 and Local scope; E reports 1 and Local; F reports 2 and Local; I reports 1 and Global; K reports 1 and Global. The submitted counters therefore total six.

This is a reproducible sum of operator-maintained fields, not an independent physical audit and not necessarily the count used by LACNIC’s April chart. The records were retrieved six months later. Their Updated timestamps are metadata-change times, not deployment receipts. Three rows have an empty Status field; that absence cannot responsibly be translated into either “operational” or “failed.”

The D record is a useful warning against counting by eye: it includes two identifiers while its Instances field says one. The F record, by contrast, reports two instances in one site row. Summing visible hostnames or rows would answer a different question from summing the published counters.

A city address is not a measurement path

The site records classify D, E and F as Local, and I and K as Global. The operators’ FAQ defines these as routing scopes. A Global route can be announced across the Internet, yet a source may still use another instance. A Local route is announced only to nearby networks; a distant source cannot query that local instance. Neither label grades latency, reliability or quality.

That distinction matters for the Panama chart. A probe located in Panama City does not, by location alone, reveal whether its query reached a Local site, an instance elsewhere, or a particular root letter’s site. The accessible prose in LACNIC’s 2026 article says it presents response times from Panama, but does not identify a probe-to-site mapping, route class, source-network cohort, address family, distribution statistic or observation window. The chart may show useful observations; the prose does not support assigning a value to D, E, F, I or K.

InteRed’s 2022 post says its members could reach I and K at its exchange and reports round-trip times below five milliseconds for those member networks. That is an operator statement about its own network context, not an independent country-wide measurement. The current records’ Global labels for I and K further illustrate why exchange location and route visibility are separate facts.

Put the denominator beside the result

LACNIC’s 2026 article already acknowledges a sampling limit: its 21 connected probes were concentrated in one city, and broader network and geographic coverage would improve the view. A future country snapshot could state whether a count refers to unique root letters, site rows or the sum of operator-submitted Instances. For each site it could show the identifier, submitted count and Local or Global type.

A measurement note should give the probe’s network and geography, anchor status, address family, observation interval and summary statistic. Where the selected site cannot be observed, mark that attribution as unknown instead of implying that geography supplied the answer.

That would not turn an Atlas sample into a census of every resolver or network in Panama. It would let readers distinguish a measured result from an inventory and understand which part of the network the sample represents. Five identities, six listed instances and a set of response times can all be true at once. They become decision-useful when the unit, date and path travel with each figure.

Sources: LACNIC, Internet in Panama (15 April 2026); LACNIC, country internet measurements (12 April 2024); Root Server Operators’ FAQ; operational-information data model; Panama records for D, E, F, I and K; InteRed on I and K at its exchange (16 September 2022).