Understanding Undetermined results in Australian country reports

Net Neutrality Monitor is a beta, volunteer-supported project developed and powered by AirVPN. It tracks and analyses Internet censorship and filtering by ISPs and DNS servers, then publishes country reports using blacklist, probe, DNS and statistical data. The results are offered for indicative research rather than as definitive legal or network-wide findings.

In an Australian country report, an Undetermined result means the available observations do not support a confident classification. It is neither a clean bill of health nor a declaration of censorship. It marks a point where technical evidence is missing, inconsistent or impossible to interpret with confidence.

That distinction is essential in Australia, where the same service may behave differently on an NBN home connection, a business fixed-line service and mobile data. A URL that loads in Sydney may time out in Perth, while a domain with an Australian identity may resolve to servers located overseas. Country reports must therefore be read as measurements, not snapshots of every Australian connection.

What Undetermined results communicate

Undetermined is best understood as an evidence-status label. A blocked verdict needs repeated, meaningful evidence of interference; an accessible result needs successful probes from relevant vantage points. When observations are too sparse or mixed to distinguish filtering from an outage, DNS failure or an endpoint problem, the honest result remains unresolved.

The label does not mean a URL is harmless, legal, safe or unimportant. Nor does it mean every attempt has failed. It says the project cannot tell from the recorded data which explanation is better supported. This matters especially in a beta report, where coverage can change as probes, submissions and analytical methods improve.

Why a measurement can stay inconclusive

Internet access involves several steps: a name is resolved, a route is selected, a connection is established and an HTTP response is returned. A DNS lookup that returns no answer, a timeout, a connection reset and a 403 response are not interchangeable. Treating any one error as proof of blocking can turn a local fault into a false censorship claim.

Timeouts can also come from overloaded servers, captive portals, faulty middleboxes, poor mobile coverage or congestion on an international path. Australia’s geography makes this visible: even traffic between Adelaide and Perth can take different routes, while traffic to overseas hosting points may cross submarine cables and distant transit networks. A probe’s own location can therefore shape the outcome.

Coverage is another constraint. The project’s country measurement coverage shows why frequently probed countries can yield firmer patterns than places with fewer observations. A quiet or intermittent signal in a sparsely tested country may reflect missing probes rather than deliberate filtering.

Why Australian network differences matter

Australians commonly talk about “home internet” and “mobile data” as separate experiences, and they can be separate networks even when they use the same phone or laptop. Telstra, Optus and other providers operate fixed and mobile infrastructure, while many fixed-line services are delivered over the NBN and resold through retail ISPs. A failure reproduced only on one provider’s mobile network cannot automatically be assigned to the Australian Internet as a whole.

Geography adds another layer. A website may be reachable from Melbourne but intermittently drop out on a regional service in Western Australia. DNS resolvers can return different answers by location, and IPv4 and IPv6 paths may diverge. The familiar .com.au suffix suggests an Australian registration or presence, but it does not prove that the content is hosted in Australia or available to every Australian network.

Comparing evidence without overclaiming

A useful comparison separates what an observation supports from what it cannot establish. The same caution applies whether the report refers to a submitted domain, an IP address or a broader country pattern.

Repeated success, repeated failure and mixed outcomes answer different questions. Reading them beside the underlying DNS and probe records prevents a label from carrying more certainty than the evidence warrants.

Evidence pattern	Supported reading	Important limitation
Repeated successful probes	The endpoint was reachable at those times and from those probe locations.	It does not guarantee access on every ISP, device or route.
Repeated failures with consistent filtering indicators	Interference becomes a plausible explanation.	An outage, server error or configuration issue must still be ruled out.
Some probes succeed while others fail	Access may be selective, route-dependent or intermittent.	DNS information, timing and probe context are needed before naming a cause.
Missing, sparse or conflicting observations	Undetermined is the defensible status.	It is not evidence that censorship is either absent or present.

Checking a result from an Australian connection

If a reported site “won’t load”, the simple ISP test offers a practical way to separate a local browser problem from filtering or throttling. Compare name resolution with the response from the same destination, and test again at different times. A useful record includes the ISP, general location, connection type, time and exact error message.

Next, compare home internet with mobile data rather than assuming the two share a route. When someone says a site is “dropping out”, that description can cover anything from a crowded train station to a congested international link, so a second observation is more valuable than a dramatic screenshot. Checking the target service’s status page and its DNS response can provide useful context before trying another network.

The service accepts URLs and IP addresses for monitoring, but submissions need careful context. Redirects, expired pages, rate limits and region-specific responses can resemble filtering, while an IP address may behave differently from its hostname. A page can also be reachable over one protocol but not another. These distinctions make a user report evidence to investigate, not a verdict in itself.

Responsible use of monitoring data

Country reports are most valuable when their scope is stated plainly. Net Neutrality Monitor is volunteer-supported and presented as a beta, so results can be partial, revised or superseded. The student expression guidance comes from a different policy area, but it illustrates a shared analytical rule: an institutional rule or automated classification should not become a broader claim without examining the context.

Privacy deserves equal weight. Submitted URLs and IP addresses can reveal browsing interests, affiliations or location, even when no page content is collected. Retention should be tied to a defined research purpose and kept as limited as practicable. A related concern appears in drone data retention guidance, where public-safety UAS records are discussed as a case for clear retention limits. The technical contexts differ, but both show why collected data needs a defined purpose and an end point.

Corrections are part of the method, not an afterthought. People with reproducible observations can contact the project so those details can be checked against existing records. Clear evidence helps distinguish a genuine pattern from a one-off local failure and keeps interpretation tied to observations rather than expectation.

Undetermined should therefore be treated as a bounded statement: the current data cannot settle the question. In Australia, that boundary may reflect DNS behaviour, route variation, sparse probing, endpoint errors or the complexity of a market in which home broadband and mobile networks do not behave alike.

The key point to remember is that a country report classifies evidence, not the character of a country or the legality of a service. An Undetermined entry preserves uncertainty honestly; it becomes misleading only when readers replace it with an unsupported claim that the Internet is either open or censored.