I hesitated having to do this for quite some time, but the fact is that the situation reached unsustainable levels.

Every business has operational costs and Winstep is no exception. Server hosting and bandwidth costs money and so do the GeoIP queries required to determine user location - on the website so that the price of the software can be displayed in the proper currency (USD, EUR or GBP), and on the software so the Weather module can automatically default to the user's city and country when you run it for the first time.

Winstep also does basically no marketing, relying mainly on word of mouth and on the popularity of the free version of Nexus to do the marketing for itself. This works because a percentage of the users trying the free version of Nexus will want to enjoy the benefits of the paid versions, and it's those sales that cover the operational costs and provide the income necessary to put food on the table.

This has worked well for decades (literally decades, Winstep Software Technologies has been in business since 1999) - that is, until recently, when the Nexus dock apparently started to become popular in China.

Just to give you an idea of what this means, the free version of Nexus had over half a million downloads last month alone, consuming over 14TB of bandwidth! That's over 500,000 new users of the free version of Nexus in a single month! Unfortunately the vast proportion of these downloads originated from China.

Had it been almost any other country this would have been extremely good news and I would be laughing all the way to the bank (remember, for every 100 downloads of the free version of Nexus, a small percentage will update to one of the paid versions) and China, with a population exceeding 1.4 billion, has what is apparently a huge POTENTIAL market.

Unfortunately, as so many Western businesses now pulling out of the Chinese market found out, potential is not the same as effective. For various reasons related to China alone which I will not get into here, the vast majority of those half a million downloads basically translated into zero sales - as if that wasn't bad enough, ever since China discovered Winstep, bandwidth usage basically quadrupled and so did GeoIP query usage (which has a fixed price per query that comes directly out of my pocket).

Over time traffic from China showed no signs of stabilizing or even decelerating, quite the opposite in fact, making the Winstep server go over the contracted bandwidth limit for the first time this month. Once that happens I start paying by the GB - now imagine what would happen if I released a new update under these conditions (which also exponentially increases bandwidth usage for over a week)!

Unfortunately the only practical solution I have at this point is to block the entire range of Chinese IP addresses at the Firewall level. Even though I only had 120 or so sales to China in over 25 years (!) - and no idea how many of those were renewals instead of new sales - I still hesitated to resort to such an extreme measure - until the server going over the contracted bandwidth limit forced my hand.

So, at this point all traffic coming from China is blocked. Any legit Chinese customers can still use a VPN to access the Winstep web site and download updates.

A VPN that gives a Singapore (and others, I think) IP when connecting is Proton VPN, which is also available in a free version. Also, China-based users can use TOR browser and the TOR network to connect to Winstep.

Just to illustrate the problem, here is a graph of the past 24 hours of bandwidth usage. It's now clearly visible the point in which Chinese IP addresses were blocked and the impressive effect it had on the outgoing traffic.

I'm usually competent in this sort of thing, but I don't understand why using a VPN instead of not using one makes a difference in this situation.

If the same people switch to a VPN, why does that make a difference? Please explain.

Normally VPN servers are located in different countries and you can choose which one to use. If you are in China but use a VPN in, say, Sweden, to the Winstep server the traffic appears to be coming from Sweden (because the IP address of the VPN server will resolve to Sweden) although it is actually coming from China (but being "intercepted" and "redirected" by the VPN server).

Also, in the case of people based/living in China, VPNs and/or TOR - which works similarly to a VPN - are even more important to many users, both ordinary users because so many western sites are blocked/banned by the Chinese government, and people like journalists, activists etc. to circumvent surveillance for obvious reasons.

Proton VP - which now is the largest and fastest, as well as the most secure and private one there is - works slightly differently to most if not all other VPNs in that it not only offers a wider range of protocols, inc. their own 'Stealth' protocol on all platforms though Linux is still to follow, but also offers a connection called 'Secure Core', where you connect to a special S.C. server in Switzerland, Iceland, or Sweden which then connects to a second server in your chosen 'exit' country. A further option is connecting via VPN to the TOR network. There are also connections available that are specially meant for streaming to bypass many streaming services VPN block, and others specially suited for torrenting and other peer-to-peer stuff.

There's also a basic free version available, which is faster than others in my trials. (I use an 'all-in' paid package which worked out cheaply enough.)

I'm still not clear. Is it because there is too much traffic coming from one location? What if almost all of them use the same VPN location other than China? Would that create the same problem?

Or is it in the hope that they'll spread out the bandwidth to several countries?

It is illegal to use a VPN in China (although many people still do) as that bypasses Chinese government controls over what you can and cannot access.

Either way, many people is definitely not equal to everyone in this case, so blocking all traffic originating from Chinese IP addresses basically solved the problem in one fell swoop.

Plus, most of the traffic was probably originating from a link in a Chinese web site(s) that pointed directly to the Winstep web site.

The problem was that way too much traffic came from China, causing a massive problem with the Winstep site's bandwidth. Hence, Jorge blocked all Chinese IP addresses at his server. Anybody in China using a VPN (or TOR) would be able to access the Winstep site as the VPN or TOR server IP would not be in China but Singapore, Japan, or anywhere else. This was aimed really at the odd few users of paid Winstep apps in China.

It would be more than unlikely that all the thousands of people who caused the problem would suddenly decide to use TOR or a VPN. It would be next to impossible for all of them to end up with an identical IP location.

As soon as the block came into effect bandwidth use dropped dramatically to a more normal level as can be seen in the graph that Jorge posted in this thread. And we now also see more normal guest numbers in the forums.

Hope this helps clear things up for you Windy.

Yes, it does. Thanks.

My pleasure. Glad it did.

Well, we're still seeing somewhat unusually high guest numbers at times so I'd guess that's a bit more perhaps than paid up users in China? Maybe more people coming in via VPN saw a link to this post somewhere? Who knows. Worth keeping an eye on me thinketh.

What really matters is that bandwidth and GeoIP requests are no longer being sucked dry by the whole population of China.

Bandwidth usage since Chinese IPs were blocked has remained stable.

LOLOL! Imagine if the literally whole population of China logged into just a dozen sites - the whole internet would collapse! (And wouldn't that be great actually. )

Anyway, good to know. As long as things remain nice and stable.

All the same though, it still seems somewhat disturbing almost that we suddenly regularly still see 200+ guests instead of the more usual max. of about 60-70, rarely 80. Just doesn't 'feel' right somehow.