Is the Great Firewall a simple on/off switch that either blocks a site or doesn't?

No. The article says this is a common but incorrect assumption, and that the system's results are uneven—some sites are consistently unreachable, some load slowly, and some work at certain times or networks but fail at others.

If a website won't load from inside China, does that mean it's blocked?

Not necessarily. The article says failure and blocking are different things; a site may be deliberately blocked, throttled, affected by distance and routing, or experiencing site-side problems.

Does the Great Firewall use only one method to filter traffic?

No. Academic researchers describe it as implementing many different types of censorship and content filtering rather than a single mechanism, and filtering can happen at several points along a request's path.

How the Great Firewall Works: Partial Blocking Explained

Q: What is the difference between the Great Firewall and the Golden Shield Project?

The article states that the Great Firewall serves the propaganda system, while the Golden Shield is a tool of the public security system, and that the two are often lumped together despite having different purposes.

Q: Is the Great Firewall's enforcement static, or is it changing?

The article says the system is not static, citing a leaked Shaanxi Telecom directive that points to a possible escalation in how censorship is applied at the network level, though it remains a leak rather than a confirmed policy.

A common assumption is that China's internet filtering is a simple switch: a site is either blocked or it isn't. Measurement-focused reporting on the system, including a recent explainer from Index in China, points to something messier. Understanding how the Great Firewall works means accepting that its results are uneven, and that a site that fails to load from inside China is not necessarily a site China has blocked.

How the Great Firewall blocks traffic

The Great Firewall is a nation-scale system. Academic researchers describe it as implementing many different types of censorship and content filtering, not a single mechanism. The Chinese government has been controlling access to online information for more than 25 years, and the system has changed a great deal over that time.

A popular analogy compares it to a mailman who tosses away any letter addressed to Google or Facebook. It is a useful picture, but incomplete. A real request passes through your device, your internet service provider, and a series of networks before it reaches a website. Filtering can happen at several of those points, and each method can behave differently.

It also helps to separate the Great Firewall from the related Golden Shield Project. One explainer notes that the Great Firewall serves the propaganda system, while the Golden Shield is a tool of the public security system. The two are often lumped together, but they have different purposes.

Why blocking is partial and inconsistent

Because the system is layered and operates at national scale, outcomes vary. The source article frames its analysis around measured data on how likely a given website is to be blocked, rather than treating blocking as a given. The picture that emerges is one of degrees:

Some sites are consistently unreachable.
Some load slowly, a pattern usually described as throttling.
Some work at certain times or on certain networks, and fail at others.

This unevenness matters. Network operators inside China run their own infrastructure, so a result measured on one connection may not match what someone sees on another. A leaked directive reported earlier suggests enforcement may be changing too: a leaked Shaanxi Telecom directive points to a possible escalation in how censorship is applied. That would shift attention from individual users to the network level. It remains a leak, not a confirmed policy, but it shows the system is not static.

A site that fails isn't always blocked

The most useful lesson from the measurement-driven approach is that failure and blocking are different things. If your website loads poorly or not at all from inside China, several explanations are possible:

Deliberate blocking: the site is actively filtered.
Throttling or congestion: the site is reachable but slow enough to appear broken.
Distance and routing: cross-border connections can be slow or unreliable without any censorship involved.
Site-side problems: page elements, scripts, or third-party services your site depends on may fail, even if the main domain is fine.

The last point trips up many site owners. A page can fail to render properly because one embedded resource is unreachable, which looks like a block on the whole site. Diagnosing the real cause requires testing from inside the country over time, not a single check.

For website owners, the practical takeaway is to avoid assuming the worst. Treat a failed load as a symptom to investigate, not as proof of a block.

What this means for users relying on encrypted connections

The same unevenness shapes the experience of people inside China who rely on VPNs and similar tools. If filtering is partial, then circumvention tools may work well one day, slowly the next, and not at all on a different network. Users often read this as a tool being broken, when the network conditions are what changed.

This is where obfuscation comes in. It is a technique that disguises encrypted traffic so it looks like ordinary internet traffic, making it harder for a network to detect or block VPN usage. How well any such approach works depends on how the filtering around it is being applied at that moment.

China is not the only country moving toward tighter control over network infrastructure. Russia, for instance, is advancing plans to tighten operating requirements for internet service providers, which would reshape who can legally run a network there. In both cases, the trend is toward control at the infrastructure level.

What This Means For You

If you run a website with an audience in China, test from inside the country, repeat the tests over time, and check separately for slowness, partial rendering, and outright blocks before drawing conclusions. If you live in or travel to China, expect inconsistency. A tool that fails once has not necessarily been permanently blocked, and one that works today may not work tomorrow.

Actionable takeaways

Don't assume a failing site is blocked. Check for throttling, routing delays, and broken third-party resources first.
Test repeatedly and from more than one network, since results can differ.
Remove or replace site dependencies that may be unreachable from inside China.
Learn how obfuscation works so you understand what VPNs do to avoid detection, and keep an eye on the leak report on possible overseas access blocking to see how enforcement may be escalating.

How the Great Firewall works is less a story of a single wall than of a shifting set of filters. Treating it that way leads to better diagnosis, better expectations, and better choices about the tools you depend on.